| mordred | clarkb: yah | 00:00 |
|---|---|---|
| jroll | harlowja_: it's interesting to me that oslo.concurrency doesn't provide anything for actual distributed locking between two computers | 00:01 |
| jroll | harlowja_: where it seems like that's tooz's primary goal | 00:01 |
| *** emagana has quit IRC | 00:01 | |
| *** camunoz has quit IRC | 00:02 | |
| harlowja_ | jroll yup, tooz was made to be that | 00:02 |
| harlowja_ | *or part of it was | 00:02 |
| harlowja_ | again goes back to the current idea, make generic libraries, if openstack needs special config or whatever, make a oslo.* library for just that stuff | 00:03 |
| jroll | harlowja_: right, so maybe I'm misunderstanding the problem here, but it doesn't seem like those really overlap | 00:03 |
| harlowja_ | i think the discussion is just about 'oslo.* does what, wait a second it doesn't do much but wrap something else, whatttttt' | 00:03 |
| harlowja_ | and that devolves into well than what is oslo... | 00:04 |
| harlowja_ | *which i can only offer opinions on :-P | 00:04 |
| *** annegentle has quit IRC | 00:04 | |
| clarkb | harlowja_: and specifically should we be pushing these "special" things upstream | 00:04 |
| harlowja_ | others may (and afaik do) have different opinions | 00:04 |
| harlowja_ | depends on how special, lol | 00:05 |
| harlowja_ | not everyone wants to see your ugly stepchild | 00:05 |
| harlowja_ | *even if they are special | 00:05 |
| * harlowja_ oh man, bad analogy, lol | 00:05 | |
| clarkb | harlowja_: but thats an indication that it doesn't belong in any lib | 00:05 |
| clarkb | harlowja_: eg if its ugly then all of openstack should not be using it | 00:05 |
| harlowja_ | u saying my ugly stepchild isn't pretty | 00:06 |
| harlowja_ | how dare u | 00:06 |
| harlowja_ | lol | 00:06 |
| harlowja_ | clarkb all valid points | 00:06 |
| harlowja_ | if i could of done it all differently, i might have :-P | 00:06 |
| harlowja_ | time travel not working | 00:06 |
| harlowja_ | ha | 00:06 |
| fungi | gah, it already merged | 00:07 |
| clarkb | fungi: its ok we can follow up with cleanup | 00:07 |
| openstackgerrit | Jeremy Stanley proposed openstack-infra/system-config: Simplify nodepool diskimage logging https://review.openstack.org/185788 | 00:08 |
| fungi | clarkb: that's ^ what we want? ^ | 00:08 |
| clarkb | fungi: ya I think so | 00:08 |
| fungi | i was hoping to use the pending change rebased on that to prove it didn't add new sections | 00:08 |
| fungi | that explains why gerrit was giving me fits trying to submit the rebased version | 00:08 |
| clarkb | oh :( | 00:08 |
| fungi | i wasn't going crazy after all ;) | 00:08 |
| clarkb | but the removals from the logging conf look correct to me | 00:09 |
| *** annegentle has joined #openstack-infra | 00:09 | |
| harlowja_ | bnemec probably has some ideas/comments on all this to | 00:09 |
| harlowja_ | *if he's around | 00:09 |
| harlowja_ | we chatted in #openstack-oslo this morning (my time) about that same question ;) | 00:09 |
| harlowja_ | mordred where's that mini-cloud-init repo u have | 00:10 |
| *** ZZelle_ has quit IRC | 00:10 | |
| clarkb | harlowja_: glean | 00:11 |
| harlowja_ | thx | 00:11 |
| *** nelsnelson has quit IRC | 00:11 | |
| clarkb | mordred: I think you can make a new shade now | 00:11 |
| fungi | harlowja_: http://git.openstack.org/cgit/openstack-infra/glean/tree/ | 00:11 |
| harlowja_ | yuppers, thx, mordred have u and jroll been in contact about the network format changes (that i think u mentioned) | 00:12 |
| harlowja_ | i think jroll was (or knows the person) who was proposing one (maybe the same one?) | 00:12 |
| jroll | harlowja_: what about it | 00:12 |
| jroll | harlowja_: yeah, that's my team working on it, both nova and cloud-init side | 00:12 |
| clarkb | changes atop what already is used? | 00:12 |
| jroll | primarily JoshNang | 00:12 |
| clarkb | or changes being changes to nova | 00:12 |
| harlowja_ | jroll k, the same one mordred was talking about (?) | 00:12 |
| jroll | clarkb: it *still* isn't in nova | 00:12 |
| harlowja_ | *if u know | 00:12 |
| jroll | yes | 00:12 |
| harlowja_ | cools | 00:12 |
| mordred | format change? | 00:13 |
| mordred | like, other than what's in the nova tree? | 00:13 |
| jroll | harlowja_: assuming that's the stuff running in rax cloud | 00:13 |
| harlowja_ | mordred something something, network config json | 00:13 |
| jroll | I'm not aware of schema changes in quite a while | 00:13 |
| mordred | yah | 00:13 |
| jroll | but yes, that | 00:13 |
| mordred | I test for network-config and also for vendor.json | 00:13 |
| mordred | btw - I have functional tests for all known forms of this | 00:13 |
| harlowja_ | right, k | 00:13 |
| mordred | http://git.openstack.org/cgit/openstack-infra/glean/tree/glean/tests/fixtures | 00:13 |
| harlowja_ | cool | 00:13 |
| mordred | http://git.openstack.org/cgit/openstack-infra/glean/tree/glean/tests/fixtures/liberty/mnt/config/openstack/latest/network_info.json | 00:14 |
| mordred | for instance | 00:14 |
| harlowja_ | cool beans | 00:14 |
| mordred | and http://git.openstack.org/cgit/openstack-infra/glean/tree/glean/tests/fixtures/rax/mnt/config/openstack/latest/vendor_data.json | 00:14 |
| mordred | so - PLEASE feel free to steal/use them in cloud-init v2 | 00:14 |
| jroll | harlowja_: we have patches for cloud-init v1 if that's useful to you | 00:14 |
| openstackgerrit | Colleen Murphy proposed openstack-infra/project-config: Install zlib1g-dev for puppet beaker jobs https://review.openstack.org/185791 | 00:14 |
| harlowja_ | so the next question (applicable to glean, cloud-init) is nova (or neutron) going to update that :-/ | 00:15 |
| harlowja_ | *when hotplugs happen... | 00:15 |
| *** camunoz has joined #openstack-infra | 00:15 | |
| clarkb | harlowja_: config drive is never updated aiui so no? | 00:15 |
| mordred | ah - well, glean does not support hotplug/config-drive changes | 00:15 |
| jroll | harlowja_: config drive is versioned | 00:15 |
| jroll | harlowja_: so if we change it, it will be a new configdrive version | 00:15 |
| harlowja_ | mordred right, thats one of the major reasons why scott wants new cloud-init i think | 00:15 |
| *** markvoelker has joined #openstack-infra | 00:16 | |
| harlowja_ | for that damn hotplug stuff (which apparently works on amazon?) | 00:16 |
| mordred | ah | 00:16 |
| clarkb | harlowja_: with config drive? | 00:16 |
| mordred | well, I do not want to support it myself | 00:16 |
| mordred | I am of the opinion that if you are adding devices to your cloud instance | 00:16 |
| harlowja_ | ya, except its the cloud, things are dynamic, annnd magic! | 00:16 |
| jroll | I'm not sure how configdrive would support hotplug, but hey | 00:16 |
| mordred | then you are doing it from config management | 00:16 |
| clarkb | jroll: right I don't think it can | 00:16 |
| mordred | and you should be able to reach-in and do it | 00:16 |
| fungi | i would be surprised to see a nodepool-booted diskimage-builder instance survive long enough to see a hotplug event, nor would we be too concerned if it went toes-up | 00:16 |
| clarkb | so this would be a metadata only use case | 00:16 |
| jroll | you'd have to unmount, rebuild, and remount | 00:17 |
| clarkb | jroll: which is not something that nova does | 00:17 |
| mordred | this is why I do NOTHING in glean other than get a node ssh-able | 00:17 |
| *** sigmavirus24_awa is now known as sigmavirus24 | 00:17 | |
| clarkb | jroll: and I think config drive is a nova thing? | 00:17 |
| mordred | anything past that can be managed by config management | 00:17 |
| mordred | and by config management, I mean ansible :) | 00:17 |
| clarkb | mordred: can has shade release? | 00:17 |
| mordred | clarkb: yes - one sec | 00:17 |
| jroll | clarkb: harlowja_ though I assume all this metadata stuff has the same schema in config drive and the metadata service, so maybe it's relevant | 00:17 |
| clarkb | jroll: aiui no | 00:17 |
| clarkb | jroll: and thats part of the prblem with what rax is doing on config drive? | 00:17 |
| jroll | x.x | 00:18 |
| openstackgerrit | James E. Blair proposed stackforge/gertty: Advance cursor on change list toggle https://review.openstack.org/185698 | 00:18 |
| openstackgerrit | James E. Blair proposed stackforge/gertty: Allow reviewing one change in change list https://review.openstack.org/185697 | 00:18 |
| openstackgerrit | James E. Blair proposed stackforge/gertty: Fix updating flags on threaded changes https://review.openstack.org/185696 | 00:18 |
| openstackgerrit | James E. Blair proposed stackforge/gertty: Add ability to review multiple changes at once https://review.openstack.org/185695 | 00:18 |
| jroll | clarkb: no clue if that's a problem | 00:18 |
| clarkb | mordred would know but I thought this was a completely new schema | 00:18 |
| mordred | it's a new schema in liberty | 00:19 |
| mordred | in rax it's in vendor-data (appropriately) | 00:19 |
| clarkb | and since metadata server is an ec2 thing its not really something you can just add to/change and make happ | 00:19 |
| jroll | mordred: *hopefully* in liberty :P | 00:19 |
| clarkb | mordred: ah ok | 00:19 |
| mordred | jroll: if it's not in liberty I will personally kill every single member of the nova core team with my tongue | 00:19 |
| clarkb | mordred: and metadata service supports it in liberty? | 00:19 |
| jroll | clarkb: aha, I see | 00:19 |
| mordred | clarkb: you CAN add to ec2 metadata service | 00:19 |
| jroll | mordred: right, I'd like to keep them around so I hope it's in liberty :P | 00:19 |
| mordred | clarkb: is the same as config drive - it has openstack stanzas in the path | 00:20 |
| clarkb | mordred: gotcha | 00:20 |
| mordred | clarkb: however, I know little about it - because it's terrible | 00:20 |
| *** david-lyle has quit IRC | 00:20 | |
| openstackgerrit | James E. Blair proposed openstack-infra/project-config: Update infra-specs voting rules https://review.openstack.org/185792 | 00:20 |
| clarkb | mordred: ya me too, I know more about how it works than the schema provided to VMs | 00:20 |
| clarkb | its basically lots of NAT | 00:20 |
| clarkb | and crying | 00:20 |
| fungi | or drinking | 00:20 |
| mordred | yah | 00:20 |
| mordred | it's a great example of why "because Amazon" is a terrible idea | 00:21 |
| mordred | it's a great example of why "because Amazon" is a terrible reason | 00:21 |
| *** annegentle has quit IRC | 00:21 | |
| harlowja_ | openstack also has a metadata service that isn't the ec2 one (that has a similar format as the layout on the config drive) | 00:21 |
| fungi | jeblair: 185792 is neat... i didn't realize project-specific review labels were possible | 00:22 |
| jeblair | fungi: yep, also 185785 | 00:22 |
| *** Sukhdev has joined #openstack-infra | 00:22 | |
| *** sdake has quit IRC | 00:22 | |
| clarkb | fungi: its new in gerrit 2.8 iirc | 00:22 |
| jeblair | fungi: check out https://review-dev.openstack.org//#/c/5308/1,publish to see it in action | 00:22 |
| fungi | neat-o! | 00:22 |
| clarkb | so part of our migration to 2.8 was update all-projects with our review labels instead of relying on the DB for it | 00:22 |
| harlowja_ | btw, would u guys mind getting in https://review.openstack.org/#/c/185095/ (thats been blocking tooz for a while) | 00:23 |
| *** davideagnello has quit IRC | 00:23 | |
| *** markvoelker has quit IRC | 00:23 | |
| harlowja_ | i'm getting kazoo to release (After its current travis weirdness is resolved) | 00:23 |
| *** markvoelker has joined #openstack-infra | 00:23 | |
| fungi | harlowja_: i take it the global requirements reviewers are on hiatus? | 00:23 |
| jeblair | you can also turn Code-Review off completely, but i chose to re-use the category in both those changes, just with altered definitions (it seemed close enough) | 00:23 |
| harlowja_ | fungi probably in the mountains around vancover or something | 00:24 |
| jeblair | i smoked some pork shoulder yesterday. i'm going to eat some more of it now. | 00:24 |
| clarkb | jeblair: nice | 00:25 |
| *** salv-orlando has joined #openstack-infra | 00:25 | |
| *** puranamr has quit IRC | 00:25 | |
| *** sarob has quit IRC | 00:25 | |
| *** davideagnello has joined #openstack-infra | 00:25 | |
| clarkb | ok shade upgraded on nodepool.o.o I am going to restart nodepool now | 00:26 |
| *** sputnik13 has quit IRC | 00:26 | |
| *** puranamr has joined #openstack-infra | 00:26 | |
| clarkb | thats done, we should be ready for uploading images tomorrow | 00:26 |
| clarkb | fungi: how is that pvmove? | 00:26 |
| fungi | clarkb: right at the 2/3 mark now | 00:27 |
| clarkb | awesome so that should be done prior too | 00:27 |
| fungi | probably another hour and it should wrap up | 00:27 |
| * clarkb is happy to have gotten this done today | 00:27 | |
| *** annegentle has joined #openstack-infra | 00:27 | |
| clarkb | jhesketh: you around? before I run off wanted to talk about the os loganalyze pass through now that shade/nodepool are sorted | 00:28 |
| fungi | i almost feel like looking into nodepool problems has gotten me slightly closer to testing the bindep-related things i wanted to test today | 00:28 |
| clarkb | jhesketh: the idea I had was to have apache call htmlify/foo on things that need htmlification, then have a passthrough/foo for things that should be passed through as is | 00:28 |
| clarkb | jhesketh: I think it would be very simple to get that change in and since its just an apache rewrite rule we can test it with logs-dev prior to updating logs.o.o | 00:29 |
| jhesketh | clarkb: I'm around but in a videoconf, I'll ping back in 15min? | 00:29 |
| *** mohitsharma has joined #openstack-infra | 00:29 | |
| *** davideagnello has quit IRC | 00:29 | |
| clarkb | jhesketh: sure | 00:29 |
| *** puranamr has quit IRC | 00:30 | |
| *** davideagnello has joined #openstack-infra | 00:33 | |
| *** ddutta has joined #openstack-infra | 00:34 | |
| *** mohitsharma has quit IRC | 00:34 | |
| fungi | jeblair: on 185792 how is the normal requirement for at least one code-review +2 satisfied? | 00:35 |
| fungi | jeblair: or is the requirement just that there's at least one vote of the maximum defined value for every label? | 00:36 |
| *** annegentle has quit IRC | 00:36 | |
| clarkb | fungi: thats what it is, the gerrit "function" is at least one maxmimum for every label and no mins | 00:36 |
| fungi | aha, got it | 00:37 |
| fungi | so by not definining the code-review label's vote range beyond +1, a +1 will satisfy | 00:37 |
| clarkb | yup | 00:37 |
| jeblair | fungi, clarkb: actually i'm setting code-review's function to 'noop' | 00:37 |
| jeblair | which means it's effectively ignored by gerrit | 00:37 |
| clarkb | ah ok | 00:38 |
| jeblair | the normal function is "MaxWithBlock" | 00:38 |
| clarkb | I think there are 3 built in functions, ^, Max, and noop | 00:38 |
| fungi | not i see the fnords | 00:38 |
| fungi | er, now | 00:38 |
| jeblair | https://review.openstack.org/Documentation/config-labels.html#label_function | 00:38 |
| *** otter768 has joined #openstack-infra | 00:39 | |
| clarkb | ok rax is giving us 413s for image uploads | 00:40 |
| *** btully has quit IRC | 00:40 | |
| clarkb | HTTPOverLimit retry at: footime | 00:40 |
| jroll | :| | 00:41 |
| clarkb | mordred: do you know if we need quota bumps? | 00:41 |
| clarkb | or is that purely a rate limit action? | 00:41 |
| jroll | clarkb: rate limit bumps are a thing, too | 00:41 |
| clarkb | jroll: well we are very "idle" right now so any rate limiting that is happening is very weird | 00:41 |
| clarkb | OH! its using glance v1 | 00:42 |
| clarkb | so the error messge is terrible because we are doing the wrong thing completely | 00:42 |
| jroll | :| | 00:42 |
| clarkb | I think the fix is to set the cloud: value in nodepool to rackspace | 00:42 |
| fungi | if rackspace: tasks(badnez) | 00:43 |
| *** otter768 has quit IRC | 00:44 | |
| clarkb | I am writing a nodepool.yaml patch now | 00:44 |
| clarkb | its a really weird error to give when an endpoint is not useable though | 00:44 |
| *** yamahata has quit IRC | 00:44 | |
| fungi | makes me wonder how they "disabled" the v1 api (max rate of 0.0/s?) | 00:45 |
| *** nelsnelson has joined #openstack-infra | 00:46 | |
| clarkb | I am trying to figure out where shade decides to do the glance v2 + tasks thing | 00:46 |
| clarkb | to make sure that my fix is sufficient | 00:46 |
| crinkle | nibalizer: clarkb I commented on https://review.openstack.org/#/c/185754/ , I don't think adding a new hypervisor is the way to go but I think I have workarounds for getting latest beaker to work | 00:46 |
| fungi | any time we talk about interesting openstack service design choices, i now see mordred's worstcat slide in my mind | 00:47 |
| *** tiswanso has joined #openstack-infra | 00:47 | |
| openstackgerrit | Clark Boylan proposed openstack-infra/system-config: Tell nodepool/shade that rax clouds are rax https://review.openstack.org/185795 | 00:49 |
| clarkb | now that I think about it there is probably already a change to set ^ | 00:49 |
| clarkb | https://review.openstack.org/#/c/168602/4/modules/openstack_project/templates/nodepool/nodepool.yaml.erb there is but I think the cloud value used is wrong | 00:50 |
| clarkb | mordred: ^ reading os-c-c and nodepool's cloud.yaml I believe the cloud name should be rackspace not rax. Can you confirm? | 00:50 |
| *** mwhagedorn has joined #openstack-infra | 00:51 | |
| jhesketh | clarkb: pong | 00:54 |
| clarkb | jhesketh: o/ | 00:54 |
| clarkb | jhesketh: just trying to think about how we might be able to speed up the swift data pass through feature to os loganalyze | 00:55 |
| jhesketh | so re your os-log note, yes, I agree | 00:55 |
| clarkb | jhesketh: and one idea I had was to use logs-dev.o.o which we can do if we make it controllable via the apache vhost | 00:55 |
| jhesketh | clarkb: right.. so the issue was that the refactor I did broke something so it was reverted | 00:56 |
| jhesketh | I never got the logs to look into what went wrong exactly | 00:57 |
| clarkb | jhesketh: ya neither did I | 00:57 |
| jhesketh | but the plan was to make better testing to catch the issue | 00:57 |
| clarkb | aiui that all happened when you and I were sleeping | 00:57 |
| jhesketh | then re-do the refactor | 00:57 |
| jhesketh | to be clear, the revert was the right move | 00:57 |
| clarkb | jhesketh: and I think having better testing is fine, my concern is that every patch we make has gotten -1 we should go redo this other complicated and long term thing | 00:57 |
| clarkb | jhesketh: which worries me because the goal here is a simple passthrough which we can test effectively without all that | 00:57 |
| clarkb | all we need is to have code that is dormant in prod because htmlify/ is used then test a passthrough/ root on logs-dev then add that to logs.o.o | 00:58 |
| fungi | clarkb: are you sure re rax vs rackspace? it looked like nodepool was matching its cloud parameter to keys in the "clouds" list, not in the "cloud" subkeys of the associative arrays within them | 00:58 |
| *** ddutta has quit IRC | 00:58 | |
| clarkb | fungi: I am not sure | 00:59 |
| fungi | i'm trying to renavigate that now to confirm | 00:59 |
| clarkb | fungi: but the value in cloud.yaml seems to be rackspace and the value at https://git.openstack.org/cgit/openstack/os-client-config/tree/os_client_config/vendors.py#n26 is rackspace too | 00:59 |
| *** ddieterly has joined #openstack-infra | 00:59 | |
| clarkb | fungi: mordred's change to do similar uses rax and I think that is wrong based on ^ | 01:00 |
| jhesketh | clarkb: so I'd rather not throw the work out that we did to refactor os-loganalyze, but I'm not sure how we merge it without the better testing | 01:00 |
| clarkb | jhesketh: have it switch based on the root url path, and have a unittest that covers the handling of the siwtching | 01:01 |
| clarkb | jhesketh: that allows us to test via unittests the selection method, then test the actual passthrough with logs-dev.openstack.org | 01:01 |
| clarkb | so the code paths for htmlify/ should be changing much at all (iirc the refactor doesn't touch those much) | 01:02 |
| clarkb | and the refactor is reasonably well tested via the unittests | 01:02 |
| *** alexsyip has quit IRC | 01:02 | |
| *** salv-orlando has quit IRC | 01:03 | |
| jhesketh | clarkb: the patch that failed didn't do any of the pass through stuff. It was a refactor as prereq to it https://review.openstack.org/#/c/172711/4 | 01:04 |
| jhesketh | so it should have been tested by the unit tests, but we didn't catch something | 01:04 |
| *** asselin_ has joined #openstack-infra | 01:04 | |
| jhesketh | apparently related to the wsgi app | 01:04 |
| fungi | clarkb: based on my reading of os-client-config, the function shade is calling is returning a dict of the keys/values in the "clouds" associative array from clouds.yaml, so "rax" and "hpcloud" not "rackspace" and "hp" | 01:04 |
| jhesketh | if we want that refactor to go in we need to find what was wrong and (hopefully) test it | 01:05 |
| fungi | clarkb: but i agree this is confusing enough that i'm not 100% sure | 01:05 |
| clarkb | fungi: clouds.yaml is rackspace too iirc | 01:05 |
| openstackgerrit | Merged openstack/requirements: Kazoo 2.1 has a bug which is breaking tooz gate https://review.openstack.org/185095 | 01:05 |
| fungi | clarkb: http://git.openstack.org/cgit/openstack-infra/system-config/tree/modules/openstack_project/templates/nodepool/clouds.yaml.erb says "rax" | 01:06 |
| fungi | as one of the keys in the "clouds" array | 01:06 |
| clarkb | fungi: oh I see what you are saying I would be very very very sad if the cloud: foo in cloud.yaml isn't meant to match cloud:foo in nodepool.yaml | 01:06 |
| fungi | then its "cloud" key maps to "rackspace" | 01:06 |
| clarkb | so much so that I think next step is to fix that | 01:06 |
| fungi | clarkb: we have cloud being used at two levels, so conflating the meaning of each | 01:07 |
| clarkb | how to be confusing, use the same key twice with two different semantic meanings :/ | 01:07 |
| clarkb | fungi: yup | 01:07 |
| fungi | yeah | 01:07 |
| fungi | once is meant to be "type of cloud/provider environment" and one is meant to be "some specific credential set" | 01:07 |
| clarkb | jhesketh: so your preference would be to keep trying to make the devstack integration test work? | 01:07 |
| clarkb | fungi: ok I can abandon my change, https://review.openstack.org/#/c/168602/4/modules/openstack_project/templates/nodepool/nodepool.yaml.erb is the one to review | 01:08 |
| clarkb | can worry about unconfusing the meaning of cloud later | 01:09 |
| fungi | yep, agreed. i already looked over it which is what was spurring me on to try to figure this out by relearning the codebases involved | 01:09 |
| ianw | greghaynes: has anyone ever considered some sort of conditional matching on element-deps, such that different elements are chosen based on the release? | 01:09 |
| clarkb | fungi: I am rechecking it, the nodepool test should pass now as cloud is a valid config parameter | 01:09 |
| ianw | i am considering the best way to handle differences with 'dnf' on f22 | 01:10 |
| clarkb | fungi: then I think gerrit will want a rebase | 01:10 |
| clarkb | because it depends on an outdated change | 01:10 |
| fungi | clarkb: also 168602 does say "hpcloud" rather than "hp" so this seems to agree with my interpretation of how this is translating in the chain of nodepool->shade->os-client-config | 01:10 |
| jhesketh | clarkb: I'm not sure. I want to be pragmatic about it too. I think perhaps we either need the devstack integration testing or otherwise find the bug locally, fix it, and be satisfied it's going to work somehow (which possibly means untested). Otherwise we could drop the refactor and just look at getting passthrough working | 01:10 |
| clarkb | fungi: ya I think you are correct | 01:10 |
| clarkb | jhesketh: I was fine with the integration testing until it too fell victim ot "hey lets massively rewrite this other thing" | 01:11 |
| fungi | clarkb: it's not going to merge without a rebase, so rebasing it now | 01:11 |
| *** dims has quit IRC | 01:12 | |
| fungi | merge conflict on nodepool.yaml.erb | 01:12 |
| clarkb | jhesketh: though mtreinish thought that it was possibly already something devstack supported so I may be overreacting to that (sdagues comment implied that we would need to add a new feature to devstack) | 01:13 |
| *** spzala has quit IRC | 01:13 | |
| jhesketh | clarkb: well this here should add support for having os-loganalyze installed in devstack: https://review.openstack.org/#/c/180553/ | 01:13 |
| fungi | tripped up by the api-timeout additions | 01:14 |
| jhesketh | clarkb: so now we just need a test to hit the server and check the file is there... that should confirm that apache/wsgi/os-loganlayze are playing nicely togethe | 01:14 |
| jhesketh | r | 01:14 |
| *** patrickeast has quit IRC | 01:14 | |
| jhesketh | clarkb: so it may not be that much more work | 01:14 |
| clarkb | jhesketh: no, the adding magic to requirements got shot down with a suggetion that we make devstack support that for devstack plugsin | 01:14 |
| jhesketh | but I don't know yet because I haven't done it | 01:14 |
| jhesketh | oh, I see now :-s | 01:15 |
| clarkb | https://review.openstack.org/#/c/180552/ | 01:15 |
| clarkb | which is a dep | 01:15 |
| openstackgerrit | Jeremy Stanley proposed openstack-infra/system-config: Turn on persistent dbm-based caching in shade https://review.openstack.org/168632 | 01:15 |
| clarkb | and why I am frustrated :/ | 01:15 |
| openstackgerrit | Jeremy Stanley proposed openstack-infra/system-config: Switch to consuming cloud auth from clouds.yaml https://review.openstack.org/168602 | 01:15 |
| fungi | clarkb: ^ updated the remainder of that stack with a rebase to master tip | 01:15 |
| clarkb | fungi: thanks | 01:15 |
| *** btully has joined #openstack-infra | 01:15 | |
| greghaynes | ianw yea, we explicitly try not to use element deps for element selection, only for actual dependencies | 01:16 |
| jhesketh | clarkb: so we could work on fixing the refactor without tests, get enough buy in that it's good, merge it and when requirements/d-g/whatever figure out how to install magic then we add the test | 01:16 |
| jhesketh | clarkb: it's not ideal, but it's a way forward | 01:16 |
| clarkb | fungi: thoughts on approving 168602 now? it will either just work or it will completely break nodepool :) | 01:16 |
| *** mohitsharma has joined #openstack-infra | 01:17 | |
| jhesketh | clarkb: otherwise the most direct route is to drop the refactor and just do the minimal amount for a passthrough | 01:17 |
| fungi | clarkb: wouldn't be the first time today? | 01:17 |
| clarkb | fungi: nah nodepool is mostly working today, just image uploads are broken :) | 01:17 |
| clarkb | fungi: but I guess I am around so can babysit | 01:17 |
| greghaynes | ianw: The idea is the composition is supposed to be as independent as possible. There are some cases where meta elements abd stuff crop up though | 01:17 |
| ianw | greghaynes: what i mean is that DIB_RELEASE=22 for fedora should probably select a dnf element, rather than yum | 01:18 |
| clarkb | jhesketh: ya so I was trying to figure a way we could sanely test and I rmembered logs-dev exists | 01:18 |
| ianw | greghaynes: although, i think KISS at this point and a few branches will do us ... still testing | 01:18 |
| clarkb | jhesketh: probably iwth that is it shares the same install so its mostly only good for checking vhost changes | 01:18 |
| *** SumitNaiksatam has quit IRC | 01:18 | |
| clarkb | jhesketh: we could chnage it to install to a venv and run out of there separately then use it | 01:18 |
| jhesketh | clarkb: not sure how that'll help test it unless we manually patch logs-dev? | 01:19 |
| clarkb | jhesketh: ya we would either manually update the venv or have puppet install a non merged version of os-loganalyze | 01:19 |
| *** btully has quit IRC | 01:20 | |
| *** tiswanso has quit IRC | 01:20 | |
| clarkb | (ftr I don't think we should throw out the work that has been done, testing is good, I just don't want to wait for devstack to figure out how plugins can/should affect requirements) | 01:21 |
| *** tiswanso has joined #openstack-infra | 01:21 | |
| jhesketh | clarkb: okay, so maybe we do the plan where we find the bug, (confirm it works in logs-dev) merge it without tests then once devstack stuff is figured out add them | 01:22 |
| openstackgerrit | Jeremy Stanley proposed openstack-infra/system-config: Turn on persistent dbm-based caching in shade https://review.openstack.org/168632 | 01:22 |
| fungi | fixed a typo i found in that ^ one | 01:22 |
| *** jamielennox|away is now known as jamielennox | 01:23 | |
| clarkb | jhesketh: sounds good to me | 01:24 |
| *** spzala has joined #openstack-infra | 01:24 | |
| *** sigmavirus24 is now known as sigmavirus24_awa | 01:24 | |
| clarkb | I need to pop out now for dinner (will keep an eye on nodepool since I approved that change), but I think next step for that is deploy the reverted chagne to logs-dev via a virtualenv, there is a puppet module to manage venvs so shouldn't be too bad | 01:25 |
| clarkb | jhesketh: I can take a stab at ^ tomorrow likely | 01:25 |
| jhesketh | clarkb: okay thanks, sounds good | 01:25 |
| greghaynes | ianw: ah, yea thats an annoying case. I cant remember what all the yum element does but I wonder if we cant make it just DTRT depending on distro version? | 01:25 |
| jhesketh | clarkb: I'll push up the revert revert and see if I can replicate locally | 01:26 |
| clarkb | jhesketh: kk | 01:26 |
| fungi | okay, pvmove just completed. i'll clean up the sata pv now that it's all running from ssd | 01:26 |
| greghaynes | I'm on my way to airport so only half here btw | 01:26 |
| *** mohitsharma has quit IRC | 01:26 | |
| openstackgerrit | Joshua Hesketh proposed openstack-infra/os-loganalyze: Add support for setting filters and views per file https://review.openstack.org/173329 | 01:27 |
| openstackgerrit | Joshua Hesketh proposed openstack-infra/os-loganalyze: Allow the config to set the filter and view https://review.openstack.org/173328 | 01:27 |
| openstackgerrit | Joshua Hesketh proposed openstack-infra/os-loganalyze: Reapply tidy up generators into contained objects https://review.openstack.org/185804 | 01:27 |
| ianw | greghaynes: OK, no worries, just didn't want to go doing things that have been rejected before. i'll start out by keeping the yum element & just trying some s/yum/dnf/ | 01:27 |
| *** annegentle has joined #openstack-infra | 01:27 | |
| *** annegentle has quit IRC | 01:28 | |
| ianw | if it gets ridiculous, which i think over time it will, even if not right now, separate elements is probably the way to go | 01:28 |
| greghaynes | ianw: yea, were generally fearful of adding complexity to our dependencies, so id do that as the last resort | 01:28 |
| clarkb | we could just use zypper everywhere and call it good | 01:29 |
| greghaynes | Lets do it | 01:29 |
| fungi | okay, nodepool.o.o now has only one cinder volume and it's ssd. the old one is vgreduced, pvremoved, nova volume-detached and cinder deleted | 01:30 |
| *** ayoung has joined #openstack-infra | 01:33 | |
| *** spzala has quit IRC | 01:34 | |
| greghaynes | vroom vroom | 01:34 |
| *** david-lyle has joined #openstack-infra | 01:35 | |
| fungi | also "we moved our server's data from sata to ssd and never took it offline" (openstack: surprising useful!) | 01:35 |
| *** dahoo has joined #openstack-infra | 01:35 | |
| fungi | though really that was all thanks to the magic of lvm2, except the bits where we hotplugged virtual disks into the domu. but cinder++ | 01:36 |
| *** dahoo is now known as jerryz | 01:37 | |
| *** puranamr has joined #openstack-infra | 01:38 | |
| openstackgerrit | Merged openstack-infra/system-config: Switch to consuming cloud auth from clouds.yaml https://review.openstack.org/168602 | 01:39 |
| *** zhiwei has joined #openstack-infra | 01:40 | |
| fungi | let the broken commence | 01:41 |
| mordred | woot | 01:41 |
| *** koolhead17 has quit IRC | 01:42 | |
| *** puranamr has quit IRC | 01:42 | |
| openstackgerrit | Tony Breeds proposed openstack-infra/irc-meetings: Update Chef cookbooks meeting https://review.openstack.org/182926 | 01:46 |
| openstackgerrit | Tony Breeds proposed openstack-infra/irc-meetings: Change refstack Chair https://review.openstack.org/182925 | 01:46 |
| openstackgerrit | Tony Breeds proposed openstack-infra/irc-meetings: Remove Storyboard Meeting. https://review.openstack.org/185808 | 01:46 |
| openstackgerrit | Tony Breeds proposed openstack-infra/irc-meetings: Update Docs team meeting https://review.openstack.org/185809 | 01:46 |
| openstackgerrit | Tony Breeds proposed openstack-infra/irc-meetings: Update heat meeting https://review.openstack.org/185810 | 01:46 |
| openstackgerrit | Tony Breeds proposed openstack-infra/irc-meetings: Add Ironic/Neutron meeting https://review.openstack.org/185811 | 01:46 |
| *** zz_dimtruck is now known as dimtruck | 01:47 | |
| fungi | mordred: on 185701 would returning "" instead of "None" have worked too? at least the former is a string which also evaluates false | 01:48 |
| fungi | or does it also need to be nonempty? | 01:48 |
| *** annegentle has joined #openstack-infra | 01:49 | |
| *** Sukhdev has quit IRC | 01:51 | |
| *** jyuso has quit IRC | 01:51 | |
| *** yamahata has joined #openstack-infra | 01:53 | |
| *** koolhead17 has joined #openstack-infra | 02:02 | |
| greghaynes | Didbwe verify that keystone token timeouts was the nodepool issue? | 02:03 |
| *** ivar-lazzaro has quit IRC | 02:03 | |
| *** salv-orlando has joined #openstack-infra | 02:04 | |
| greghaynes | seems scary to pile the clouds.yaml change on top before weve gotten the other issue fix... | 02:05 |
| fungi | on the gerrit 2.10 front, looks like the jgit patch we probably want (d906214) is merged and covered by today's v4.0.0.201505260635-rc2 tag | 02:05 |
| *** mwhagedorn has quit IRC | 02:05 | |
| greghaynes | ah, yes, that does look like the patch we want | 02:07 |
| clarkb | greghaynes we didnt test but know glanceclient wont get new token | 02:08 |
| greghaynes | yea | 02:08 |
| clarkb | and we need cloud change for rax | 02:08 |
| clarkb | doesnt work otherwise. uses glance v1 | 02:08 |
| clarkb | and get rate limited which is lol | 02:08 |
| greghaynes | aye | 02:09 |
| *** salv-orlando has quit IRC | 02:09 | |
| greghaynes | :/ | 02:09 |
| greghaynes | theres way too many layers of bugs were working through | 02:09 |
| greghaynes | although, the rax thing isbt an issue until we start usibg dib there yes? | 02:10 |
| *** unicell has quit IRC | 02:11 | |
| *** yamamoto has quit IRC | 02:12 | |
| fungi | which is trying to be nowish | 02:13 |
| greghaynes | oh, wow | 02:13 |
| fungi | because... why not? | 02:13 |
| greghaynes | heh | 02:13 |
| greghaynes | well, we were beibg all conservative about the first patch | 02:14 |
| fungi | it's at least for label types we're not actively running jobs on yet | 02:14 |
| greghaynes | I certainly dont mind just going for it though | 02:14 |
| fungi | so if they don't build/upload, meh | 02:14 |
| clarkb | greghaynes: the first patch was the scary one | 02:15 |
| clarkb | greghaynes: but now that we are using shade its all much less scary | 02:15 |
| fungi | (for some definitions of scary) | 02:15 |
| clarkb | and using cloud does appear to have broken the world | 02:16 |
| clarkb | OpenStackCloudException: Error getting compute endpoint: Error constructing auth plugin: password __init__() got an unexpected keyword argument 'project-id' | 02:16 |
| greghaynes | that tends to happen when you use clouds | 02:16 |
| clarkb | that looks like a "you gave me a kwarg I didn't want so here have an exception" exception | 02:17 |
| greghaynes | hah, I wonder if project -> tenant rename not happening? | 02:17 |
| greghaynes | we map between those two properties in a few places | 02:17 |
| clarkb | well looking at where the exception comes from we don't pass a project id so I have no idea | 02:18 |
| clarkb | we don't pass a tenant either | 02:18 |
| clarkb | but cloud.yaml has project-id | 02:18 |
| clarkb | greghaynes: any idea what it should be? tenant? | 02:19 |
| * clarkb can ninja that change in | 02:19 | |
| clarkb | and see if it all gets happy | 02:19 |
| clarkb | maybe its project_id? | 02:20 |
| *** salv-orlando has joined #openstack-infra | 02:21 | |
| *** taron has quit IRC | 02:21 | |
| clarkb | and of course it read that config once and now we have to restart nodepool to get it to reread | 02:21 |
| clarkb | mordred: ^ that is a regression btw | 02:21 |
| greghaynes | clarkb: looks like it can be any of project_id, tenant_id, tenant_name, project_name | 02:22 |
| *** taron has joined #openstack-infra | 02:22 | |
| clarkb | ya | 02:23 |
| clarkb | so the - was the problem, I think rh1 has similar problems with auth-url | 02:23 |
| *** ddieterl_ has joined #openstack-infra | 02:23 | |
| greghaynes | looks like also project_name is what we should really be calling it | 02:23 |
| clarkb | and we need to make it so that its reread | 02:24 |
| clarkb | but I think thats a shade thing so not sure how to influence from nodepoo | 02:25 |
| clarkb | (we can probably check file mod time) | 02:25 |
| greghaynes | hrm | 02:25 |
| *** salv-orlando has quit IRC | 02:25 | |
| clarkb | that is honestly probably the biggest problem here | 02:25 |
| greghaynes | seems like a good test | 02:26 |
| *** ddieterly has quit IRC | 02:27 | |
| clarkb | greghaynes: project_name is the least performant option | 02:27 |
| greghaynes | hah | 02:27 |
| clarkb | tenant_id the most | 02:27 |
| clarkb | (its a small number of iterations so meh, just pointing that out) | 02:27 |
| greghaynes | so the os-c-c code makes it read like they are all just backwards compat options for project_name | 02:28 |
| clarkb | I will just s/-/_/ | 02:28 |
| greghaynes | yea | 02:28 |
| greghaynes | The amount it matters asymptotically approaches 0 | 02:28 |
| clarkb | ok I have puppet stopped so it won't break my manual edits and puppet change in a moment | 02:28 |
| openstackgerrit | Clark Boylan proposed openstack-infra/system-config: Use _ instead of - in clouds.yaml https://review.openstack.org/185824 | 02:29 |
| clarkb | fungi: mordred greghaynes ^ tahts manually in place and seems to make nodepool happy | 02:29 |
| greghaynes | clarkb: you checked auth_url has same issue? | 02:30 |
| asselin_ | greghaynes, does this stacktrace mean anything to you? http://paste.openstack.org/show/238941/ | 02:30 |
| clarkb | greghaynes: yup, rh1 was failing until I made that change (and restarted nodepool again, grumble grumble) | 02:31 |
| clarkb | actually thinking about this more I wonder if we should just make the nodepool cloud: value mean what oscc value rather than the clouds.yaml value | 02:31 |
| greghaynes | asselin_: We just ran into that error and pushed up a patch to fix | 02:31 |
| clarkb | if we do that then we don't have any of this restarting nonsense | 02:31 |
| asselin_ | greghaynes, ok great! | 02:32 |
| greghaynes | asselin_: youll need to grab latest shade | 02:32 |
| clarkb | asselin_: ya you want to upgrade to latest shade | 02:32 |
| asselin_ | ok thanks | 02:32 |
| greghaynes | clarkb: huh? | 02:32 |
| *** david-lyle has quit IRC | 02:33 | |
| greghaynes | clarkb: os-c-c just pulls out of clouds.yaml | 02:33 |
| clarkb | greghaynes: so I have discovered (and fungi helped me grok) that the oscc and nodepool and shade config is all needless confusing and needs a serious overhaul imo | 02:33 |
| clarkb | greghaynes: nodepool.yaml has a cloud: foo config option | 02:33 |
| greghaynes | yea, that maps to an os-c-c cloud | 02:33 |
| clarkb | greghaynes: no it doesn't :) | 02:33 |
| nikeshm | hi | 02:33 |
| greghaynes | ok, so the terms here suck | 02:33 |
| clarkb | greghaynes: it instead maps to a key in your clouds.yaml | 02:33 |
| greghaynes | that maps to a clouds.yaml cloud | 02:33 |
| greghaynes | yea | 02:34 |
| clarkb | greghaynes: that key then maps to a oscc cloud | 02:34 |
| clarkb | greghaynes: and this is the problem | 02:34 |
| nikeshm | i am getting this issue on nodepool image-upload http://paste.openstack.org/show/238941/ | 02:34 |
| clarkb | greghaynes: if instead of relying on clouds.yaml at all which fails to reload our config properly when changes are made we kept everything but the static oscc cloud data in nodepool.yaml we could make cloud point to oscc clouds properly | 02:34 |
| nikeshm | this is my yaml for nodepool http://paste.openstack.org/show/238951/ | 02:34 |
| clarkb | greghaynes: then we stop being confusing and have the benefit of loading our config properly | 02:34 |
| *** SumitNaiksatam has joined #openstack-infra | 02:35 | |
| greghaynes | os-c-c just does config overlay, one of those steps overlays the clouds.yaml config, which cloud in clouds.yaml is indexed by that clouds parameter | 02:35 |
| clarkb | greghaynes: well its done poorly then | 02:35 |
| *** dkranz has quit IRC | 02:35 | |
| clarkb | greghaynes: it should be transparent based on name | 02:35 |
| clarkb | nikeshm: yes see the response to asselin_ , you need to upgrade to latest shade | 02:35 |
| clarkb | greghaynes: because as is it is incredibly confusing and fails to handle configs properly in nodepool | 02:35 |
| *** gyee has quit IRC | 02:36 | |
| greghaynes | I think a lot of this comes down to the use of the word 'cloud' being the worst choice possible | 02:36 |
| greghaynes | there literally cannot be a more overloaded term here | 02:36 |
| clarkb | greghaynes: yes, but also the configs are not reloaded properly by shade/oscc | 02:36 |
| clarkb | greghaynes: I can fix both if we make a single definition of cloud: in nodepool.yaml | 02:36 |
| greghaynes | yea, thats a solvable thing which can be transparent to the user | 02:36 |
| greghaynes | I think I need to see a config or something to grok what youre suggesting | 02:37 |
| clarkb | which is load static cloud config data that oscc knows about, then continue to load the non static info from nodepool.yaml | 02:37 |
| greghaynes | what are the non-static bits? | 02:37 |
| clarkb | greghaynes: literlaly https://review.openstack.org/#/c/185795/1/modules/openstack_project/templates/nodepool/nodepool.yaml.erb | 02:37 |
| asselin_ | nikeshm, sent you the details in -cinder | 02:37 |
| clarkb | greghaynes: auth url, tenant, password etc | 02:37 |
| clarkb | greghaynes: those can and do change | 02:37 |
| clarkb | greghaynes: so put them in the config that is checked and reread and reprocessed if necessary, only read the static cloud: foo bits from oscc | 02:38 |
| clarkb | greghaynes: which are things like what version of glance do I need t ouse | 02:38 |
| greghaynes | clarkb: yea, so you can do that - it should overlay what you specify in nodepool.yaml ontop of the os-c-c config (which will be backed by clouds.yaml if you specify cloud:) | 02:38 |
| greghaynes | I think theres a test for that functionality too | 02:38 |
| *** annegentle has quit IRC | 02:38 | |
| clarkb | greghaynes: not really because the way nodepool uses cloud today is to look at cloud.yaml | 02:39 |
| clarkb | greghaynes: my suggestion is to kill cluod.yaml and delete it | 02:39 |
| clarkb | its redundant and adds no value | 02:39 |
| clarkb | all it did was confuse both fungi and I and has regressed our config update handling | 02:39 |
| greghaynes | clarkb: http://git.openstack.org/cgit/openstack-infra/nodepool/tree/nodepool/tests/test_shade_integration.py#n42 is what youre describing I think | 02:40 |
| clarkb | greghaynes: specifically what I am saying is load cloud foo directly from oscc | 02:40 |
| *** otter768 has joined #openstack-infra | 02:40 | |
| *** annegentle has joined #openstack-infra | 02:40 | |
| greghaynes | clarkb: that is overlaying ontop of the config specified in a clouds.yaml | 02:40 |
| clarkb | greghaynes: do not involve clouds.yaml at all | 02:40 |
| clarkb | greghaynes: right but I am telling you clouds.aml is the problem | 02:40 |
| clarkb | greghaynes: because it does not know how to be reread AND it overloads "cloud" to the point of confusion | 02:40 |
| clarkb | if we just stop using it, then there is no config update problem, and there is no confusion about what "cloud" means | 02:41 |
| clarkb | and it will just work | 02:41 |
| *** claudiub has quit IRC | 02:41 | |
| greghaynes | for nodepool I really am ambivilent on whether it uses clouds.yaml... I dont think it buys us much | 02:42 |
| greghaynes | I do like the clouds.yaml thing in general though | 02:42 |
| greghaynes | so meh | 02:42 |
| greghaynes | Youall already have the composition logic for secrets via puppet | 02:42 |
| greghaynes | so its just not an issue for you | 02:42 |
| greghaynes | Also, I would like to propose we rename the os-c-c term 'cloud' to be 'profile' | 02:43 |
| clarkb | that seems sane | 02:43 |
| *** otter768 has quit IRC | 02:44 | |
| * morganfainberg looks at the clouds.yaml closely. | 02:48 | |
| morganfainberg | greghaynes: there is some talk of client config moving. If it does, a rename/update might be easier than doing it in-place. | 02:49 |
| morganfainberg | *shrugs* | 02:49 |
| greghaynes | morganfainberg: yea, I talked about thatvwith jamie a bunch at summit | 02:50 |
| morganfainberg | Yep. | 02:50 |
| greghaynes | moving os c c that is | 02:50 |
| morganfainberg | I support it moving btw. | 02:50 |
| greghaynes | w00t | 02:50 |
| morganfainberg | But I am not the final say. | 02:51 |
| *** fedexo has joined #openstack-infra | 02:51 | |
| * greghaynes plugs ears and goes lalalala | 02:51 | |
| * morganfainberg snerks. | 02:52 | |
| *** gokrokve has joined #openstack-infra | 02:52 | |
| clarkb | but if clouds.yaml is to stay we have to figure out updates to it in shade/oscc | 02:52 |
| greghaynes | yep | 02:52 |
| greghaynes | I dont think itll be too hard of a problem to fix | 02:53 |
| *** nelsnelson has quit IRC | 02:53 | |
| *** asselin_ has quit IRC | 02:54 | |
| *** nelsnelson has joined #openstack-infra | 02:54 | |
| mordred | what's the problem? | 02:55 |
| mordred | and not it's not redundant | 02:55 |
| clarkb | mordred we dont reload clouds.yaml | 02:55 |
| greghaynes | mordred: changes to clouds.yaml arent reflected in nodepool until restart | 02:55 |
| clarkb | which is a regression | 02:56 |
| clarkb | its redundant for nodepool | 02:56 |
| clarkb | not for other things | 02:56 |
| mordred | ok. we can fix the file reload thing | 02:57 |
| mordred | I dont' think it's redundant because it allows us to use other tools from the same config | 02:57 |
| mordred | but reverting for now is fine | 02:57 |
| clarkb | wad going to let it run | 02:57 |
| clarkb | alreadu restarted 3 times | 02:57 |
| mordred | kk | 02:58 |
| *** spzala has joined #openstack-infra | 02:58 | |
| mordred | btw - renaing things in clouds.yaml is a non-starter without a deprecation path | 02:58 |
| clarkb | see 185824 | 02:58 |
| mordred | so renaming cloud to profile would need to have completely working both things - which obviously isn't that hard | 02:59 |
| greghaynes | mordred: just support both namings | 02:59 |
| clarkb | mordred just have an alias like tenant and project | 02:59 |
| mordred | but it's out in the wild | 02:59 |
| mordred | sure | 02:59 |
| mordred | just saying | 02:59 |
| greghaynes | yes, deprecate for sure | 02:59 |
| clarkb | sure, but it will help parsing significantly | 02:59 |
| mordred | it's production now, which means we must be very very careful | 02:59 |
| mordred | wanna here a funny? | 03:03 |
| mordred | hear | 03:03 |
| mordred | so - my roku decided to unpair with showtimeanytime tonight | 03:03 |
| mordred | which means you have to go back to a browser somewhere and activate | 03:03 |
| mordred | BUT | 03:03 |
| mordred | showtime has deployed an invalid SSL cert, so firefox is complaining | 03:03 |
| mordred | and one let you do the thing | 03:03 |
| mordred | won't | 03:03 |
| *** btully has joined #openstack-infra | 03:04 | |
| clarkb | sounds fun | 03:04 |
| mordred | solution: spin up Windows VM and try with IE - which, as expected, doesn't know how security works | 03:04 |
| mordred | no warning from IE AT ALL about the broken cert | 03:04 |
| clarkb | mordred: I think the confusion is clouds.yaml has a list of clouds under clouds: | 03:04 |
| *** fedexo has quit IRC | 03:05 | |
| clarkb | mordred: then inside each cloud we specify a cloud | 03:05 |
| greghaynes | I heard you like clouds | 03:05 |
| clarkb | and that cloud is different than the cloud we are describing under clouds: and it just gets the brain confused | 03:05 |
| mordred | agree - I never could come up with a better name for that than "vendor" or something | 03:05 |
| mordred | except for people using private clouds | 03:05 |
| mordred | it might not be a vendor - it might just be a configured cloud | 03:05 |
| StevenK | provider ? | 03:05 |
| mordred | so "profile" works fine | 03:05 |
| greghaynes | They are config profiles from the clients perspective | 03:05 |
| mordred | or provider | 03:05 |
| mordred | yah | 03:05 |
| StevenK | That might an overloaded term, though | 03:05 |
| mordred | I like profile | 03:05 |
| clarkb | I think one of my nodepool restarts may have made rax ubuntu-trusty uploads unhappy, going to try deleting one of them to see if it starts again happy | 03:06 |
| mordred | also - I agree - we should s/-/_/ everywhere | 03:06 |
| *** gokrokve has quit IRC | 03:06 | |
| mordred | I _thought_ we already were | 03:06 |
| mordred | maybe we're not doing that transform on the auth dit? | 03:06 |
| mordred | dict? | 03:06 |
| clarkb | mordred: definitely aren't with latest sahde and latest oscc | 03:06 |
| mordred | kk | 03:07 |
| *** gokrokve has joined #openstack-infra | 03:07 | |
| clarkb | mordred: I can get tracebacks from nodepool logs if you want them | 03:07 |
| mordred | we can fix that | 03:07 |
| mordred | nah | 03:07 |
| mordred | if we're not - it's certainly not hard to :) | 03:07 |
| mordred | we process all the dicts and whatnot - so doing a key transform is cake | 03:07 |
| mordred | btw - for hpcloud it shold be project_name not project_id | 03:08 |
| *** btully has quit IRC | 03:08 | |
| mordred | and looks like it for tripleo as well | 03:08 |
| *** SumitNaiksatam has quit IRC | 03:08 | |
| clarkb | mordred: those are aliases | 03:08 |
| clarkb | mordred: so it doesn't matter and project_name is the least best option according to oscc code | 03:09 |
| *** SumitNaiksatam has joined #openstack-infra | 03:09 | |
| *** dpaterson has quit IRC | 03:09 | |
| clarkb | (its listed last so you have to check all of them before you get to that one) | 03:09 |
| mordred | it should be the most preferred | 03:09 |
| mordred | really? | 03:09 |
| clarkb | yes I can get a link to the line, one sec | 03:09 |
| mordred | hrm. well, ANYTHING with tenant is broken and evil and shoudl die in a fire | 03:09 |
| greghaynes | eh, all the otger options are listed as backwards compat options | 03:09 |
| greghaynes | but yes, you will loop over them first | 03:10 |
| mordred | between project_id and project_name, I believe you can pass a project_id to the project_name param but you cannot pass a name to the id param | 03:10 |
| mordred | they will all wind up in project_name at the end of the day | 03:10 |
| clarkb | https://git.openstack.org/cgit/openstack/os-client-config/tree/os_client_config/config.py#n235 | 03:10 |
| mordred | morganfainberg: ^^ remind me to punch someone about this, btw | 03:10 |
| mordred | "so that project_name is the ultimate winner" | 03:10 |
| clarkb | mordred: so oscc should change the order if project_name is preferred with tenant* being least preferred | 03:10 |
| mordred | it is | 03:11 |
| mordred | it doesn't bail on a match | 03:11 |
| * morganfainberg reads backscroll | 03:11 | |
| mordred | it keeps looping | 03:11 |
| mordred | and will overwrite anything it finds | 03:11 |
| greghaynes | I think clarkb is uber optimizing | 03:11 |
| clarkb | mordred: wat | 03:11 |
| *** gokrokve has quit IRC | 03:11 | |
| mordred | so project_name being checked last makes it most preferred | 03:11 |
| mordred | we do it that way | 03:11 |
| clarkb | thats but haivng >1 doesn't make sense | 03:11 |
| *** david-lyle has joined #openstack-infra | 03:11 | |
| clarkb | you can only have one | 03:11 |
| mordred | rather than searching and bailing | 03:11 |
| mordred | because we want to delete anything that is less preferred from the dict so that we only have one answer | 03:12 |
| *** spzala has quit IRC | 03:12 | |
| mordred | the job of this library isn't to yell at you - it's to do its best to work around the fact that openstack's story on project and tenant has been broken for years | 03:12 |
| clarkb | this is the sort of thing we should keep a link to for whenever people want to change rest apis | 03:12 |
| morganfainberg | Oh god. The name vs ID thing? | 03:12 |
| mordred | and it's ENTIRELY expectable that someone might have both set in an env var | 03:12 |
| clarkb | morganfainberg: and tenant vs project | 03:12 |
| morganfainberg | That has annoyed me a ton. | 03:13 |
| clarkb | so we have broken this on two axies | 03:13 |
| clarkb | which is O_O | 03:13 |
| morganfainberg | clarkb: if I could make project never had happened. I would have. | 03:13 |
| mordred | so, since someone could now set OS_PROJECT_NAME and OS_PROJECT_ID in env vars means we have to do the crazy above | 03:13 |
| mordred | I do apologize | 03:13 |
| morganfainberg | Tenant was a much more descriptive / better term. | 03:13 |
| mordred | it COULD do with a better comment | 03:13 |
| mordred | morganfainberg: tenant makes no sense to me, btw | 03:13 |
| clarkb | mordred: and OS_TENANT_ID and OS_TENANT_NAME | 03:13 |
| mordred | morganfainberg: I think NEITHER make any sense, tbh | 03:13 |
| mordred | clarkb: yah | 03:13 |
| mordred | clarkb: a person could set all four and expect the world to work | 03:13 |
| morganfainberg | mordred: better because of less overload. But domain should have been realm. Anyway. | 03:14 |
| greghaynes | you know what makes the least sense though - 4 options | 03:14 |
| mordred | greghaynes: yah | 03:14 |
| morganfainberg | The id to name transform is just weird. | 03:14 |
| mordred | morganfainberg: TELL ME ABOUT IT | 03:14 |
| clarkb | anyways, I don't want to change id to name because that will require another nodepool restart | 03:14 |
| mordred | yah. sure | 03:14 |
| clarkb | so my vote is stick with simple s/-/_/ | 03:14 |
| mordred | it's not a problem - at _some_ point we may want to do that just to clean up our file so that it's better correct documentation for other people | 03:15 |
| mordred | totally not urgent | 03:15 |
| morganfainberg | clarkb: please stick with simple! :) | 03:15 |
| clarkb | mordred: it might be good for oscc to document the "correct" thing and then "this is acceptable for raisins" | 03:15 |
| mordred | ++ | 03:15 |
| mordred | clarkb: I'm taking notes on a few different things we need better docs on | 03:15 |
| mordred | this, and also the dogpile key creation | 03:15 |
| *** tiswanso has quit IRC | 03:16 | |
| * morganfainberg renames | 03:16 | |
| morganfainberg | Projects to "cloudythings" | 03:16 |
| mordred | clarkb: for noticing the restarts ... that probably wants to be a nodepool thing, but perhaps getting info about what file to watch for changes from shade? | 03:16 |
| clarkb | we should just use uuids for everything then ship a lookup table book | 03:16 |
| mordred | clarkb: don't say that outloud | 03:17 |
| mordred | clarkb: someone will think you're serious | 03:17 |
| mordred | clarkb: youknow, because cloud | 03:17 |
| morganfainberg | clarkb: can we use a pdf format for that. That is all in comic sans. | 03:17 |
| clarkb | mordred: but then we can print and bind some really awesome uuids mapping books | 03:17 |
| mordred | clarkb: with LaTeX | 03:17 |
| clarkb | morganfainberg: no I want hardcover textbook size book | 03:17 |
| mordred | I know someone who does book binding :) | 03:17 |
| clarkb | so that you can't just copy paste uuids | 03:17 |
| clarkb | you have to type them in by hand or figure out ocr | 03:17 |
| morganfainberg | Oh crap. Wait. Did irccloud just change "comic sans" in line on my phone. *snicker* | 03:18 |
| mordred | clarkb: ok - I have notes on the two bugs and the two docs update we need | 03:18 |
| morganfainberg | yes it did. Ahahhahahahah. Someone has a sense of humor. | 03:18 |
| mordred | clarkb: any other issues I might have missed in scrollback to tackle in the morning? | 03:18 |
| clarkb | mordred: we will have to wait and see how the rax-ord ubuntu-trusty upload goes | 03:18 |
| clarkb | mordred: its the one I restarted for fear that a nodepool restart got the initial uploads into a funny state | 03:19 |
| mordred | k | 03:19 |
| clarkb | it thinks its still "building" | 03:19 |
| mordred | but so far nothing vomited like the -id did | 03:19 |
| clarkb | nope its running without tracebacks as far as I can see | 03:19 |
| mordred | ossum | 03:19 |
| *** rossella_s has quit IRC | 03:19 | |
| mordred | well, the other 4 are easy to do in the morning | 03:19 |
| mordred | well, 3 of them are | 03:19 |
| mordred | I'm going to need a combination of SpamapS and morganfainberg to write the docstrings about why we're doing the dogpile key stuff | 03:19 |
| clarkb | mordred: for reloads I think oscc/shade need to expose a flag that says "my config is updated" then the application consuming the libs can make a decision on whether or not to handle that | 03:19 |
| *** rossella_s has joined #openstack-infra | 03:20 | |
| mordred | clarkb: awesome. I will do that | 03:20 |
| clarkb | mordred: then you do if shade.config_updated: nodepool handle application logic here | 03:20 |
| *** kaisers has quit IRC | 03:20 | |
| mordred | yup | 03:20 |
| clarkb | mordred: probably need to make it a little smarter than that so we don't reconfigure the whole world | 03:20 |
| morganfainberg | mordred: yes. I will help doc that stuff. | 03:20 |
| clarkb | mordred: but thats the general idea of what I think makes this better for nodepool | 03:20 |
| mordred | morganfainberg: tl;dr - clarkb looked at the code and had a WTF why are we doing all of this? earlier today | 03:20 |
| *** salv-orlando has joined #openstack-infra | 03:21 | |
| mordred | clarkb: I'm going to start with dumb "my config changed" | 03:21 |
| *** panda has quit IRC | 03:21 | |
| clarkb | mordred: speaking of, I did eventually go digging into the dogpile docs and the default make a key function is undocumented :/ | 03:21 |
| mordred | clarkb: yah - this is why SpamapS and morganfainberg have to write the shade doc on the subject | 03:21 |
| clarkb | mordred: so not only does it apparently not do the right thing, I have no idea what it actually does | 03:21 |
| mordred | clarkb: only morganfainberg and SpamapS do | 03:21 |
| mordred | (understand what upstream does) | 03:21 |
| *** panda has joined #openstack-infra | 03:21 | |
| morganfainberg | mordred: and zzzeeekk | 03:21 |
| *** samueldmq has quit IRC | 03:21 | |
| mordred | yah. him too | 03:21 |
| morganfainberg | If I was less ptl-ing I have like 10+ backlogs fixes for dogpile. | 03:22 |
| morganfainberg | This being related to one of them. | 03:22 |
| *** fawadkhaliq has joined #openstack-infra | 03:22 | |
| mordred | ok. I'm going back to the TVs | 03:22 |
| morganfainberg | I will help document at the very least *plans to corner SpamapS to work on doc things soon* | 03:23 |
| clarkb | but I have learned a lot about dogpile (that was fun) and a lot about oscc (which is basically shade lite) and hopefully won't be learning everything there is to know about rax's glance task image upload process tomorrow | 03:23 |
| clarkb | mordred: rax-ord upload just failed, trying to find a traceback now | 03:24 |
| clarkb | mordred: ClientException: Object POST failed returned a 504 Gateway Time-out | 03:27 |
| clarkb | when uploading to swift | 03:27 |
| *** yamamoto has joined #openstack-infra | 03:27 | |
| greghaynes | clarkb: nice! | 03:27 |
| *** salv-orl_ has joined #openstack-infra | 03:28 | |
| clarkb | so thats progress I guess | 03:28 |
| clarkb | now time for me to play these new wiiu games | 03:29 |
| clarkb | (nodepool should be fine with that being the only issue on an unused image) | 03:29 |
| * greghaynes -> airborne | 03:29 | |
| *** salv-orlando has quit IRC | 03:29 | |
| *** salv-orl_ has quit IRC | 03:32 | |
| *** annegentle has quit IRC | 03:32 | |
| *** changbl has joined #openstack-infra | 03:32 | |
| *** jerryz has quit IRC | 03:32 | |
| *** bhunter71 has quit IRC | 03:32 | |
| *** koolhead17 has quit IRC | 03:33 | |
| *** stevemar has joined #openstack-infra | 03:33 | |
| nikeshm | clarkb: hi i tried as rasselin said,but i am not getting any /opt/shade | 03:35 |
| nikeshm | and still same error | 03:35 |
| *** salv-orlando has joined #openstack-infra | 03:36 | |
| cody-somerville | "Currently ansible only supports HTTP proxies" <-- seriously? | 03:37 |
| *** annegentle has joined #openstack-infra | 03:37 | |
| *** dalgaaf has quit IRC | 03:38 | |
| clarkb | nikeshm pip install shade==0.6.3 should work | 03:40 |
| clarkb | ansible only supports http proxies for what particular functionality? | 03:40 |
| clarkb | ansible istelf uses ssh/fireball | 03:41 |
| *** salv-orlando has quit IRC | 03:41 | |
| *** salv-orl_ has joined #openstack-infra | 03:43 | |
| *** salv-orl_ has quit IRC | 03:43 | |
| nikeshm | clarkb: shade==0.6.1 is already there,is 0.6.3 is required | 03:44 |
| clarkb | yes you need the latest version | 03:46 |
| nikeshm | so again do we have to run install_master,sh | 03:47 |
| *** sarob has joined #openstack-infra | 03:49 | |
| *** camunoz has quit IRC | 03:49 | |
| jhesketh | clarkb: hmm, so I can't reproduce the problem with os-loganalyze... I've deployed 185804 on an instance and it serves up content just fine from disk and swift | 03:50 |
| jhesketh | well the simple console.log I was testing | 03:50 |
| jhesketh | but I don't have much more to go on in terms of what broke | 03:50 |
| clarkb | nikeshm no just run the pip install command I gave you | 03:51 |
| *** emagana has joined #openstack-infra | 03:51 | |
| *** sarob has quit IRC | 03:53 | |
| *** sarob has joined #openstack-infra | 03:53 | |
| *** armax has quit IRC | 03:54 | |
| *** Sukhdev has joined #openstack-infra | 03:58 | |
| *** yfried has joined #openstack-infra | 04:00 | |
| *** jerryz has joined #openstack-infra | 04:01 | |
| *** achanda has joined #openstack-infra | 04:02 | |
| *** krtaylor has joined #openstack-infra | 04:03 | |
| *** gnuoy has quit IRC | 04:05 | |
| *** jamespage has quit IRC | 04:05 | |
| *** gnuoy has joined #openstack-infra | 04:06 | |
| *** jamespage has joined #openstack-infra | 04:06 | |
| *** dbyte has joined #openstack-infra | 04:10 | |
| *** dimtruck is now known as zz_dimtruck | 04:15 | |
| *** links has joined #openstack-infra | 04:15 | |
| nikeshm | clarkb: thanks, is this nodepool yaml file ok http://paste.openstack.org/show/238951/ i want to use same image on two cloud-provider | 04:17 |
| *** ddieterl_ has quit IRC | 04:18 | |
| nikeshm | clarkb: one cloud provider i will use for FC drivers tempest and another for iSCSI driver tempest | 04:18 |
| nikeshm | job | 04:18 |
| *** xyang1 has quit IRC | 04:20 | |
| mordred | cody-somerville: what other sorts of proxies do you need? | 04:22 |
| mordred | cody-somerville: also - ansible knows how to handle ssh jump hosts too | 04:22 |
| mordred | clarkb: just on ord? or on iad and dfw too? I think I've seen issues with ORD swift in the past too | 04:23 |
| *** annegentle has quit IRC | 04:26 | |
| *** maishsk_afk has joined #openstack-infra | 04:31 | |
| *** maishsk_afk is now known as maishsk | 04:31 | |
| *** markvoelker has quit IRC | 04:31 | |
| *** maishsk has quit IRC | 04:33 | |
| *** sarob has quit IRC | 04:34 | |
| *** yamamoto has quit IRC | 04:34 | |
| *** unicell has joined #openstack-infra | 04:34 | |
| *** puranamr has joined #openstack-infra | 04:34 | |
| *** amotoki_ has joined #openstack-infra | 04:35 | |
| *** sks has joined #openstack-infra | 04:39 | |
| *** yfried has quit IRC | 04:40 | |
| *** otter768 has joined #openstack-infra | 04:41 | |
| *** jerryz has quit IRC | 04:44 | |
| *** otter768 has quit IRC | 04:45 | |
| *** abregman has joined #openstack-infra | 04:48 | |
| nikeshm | clarkb: thanks | 04:49 |
| nikeshm | thats issue solved | 04:49 |
| *** btully has joined #openstack-infra | 04:52 | |
| *** sputnik13 has joined #openstack-infra | 04:55 | |
| *** gokrokve has joined #openstack-infra | 04:56 | |
| *** harlowja_at_home has joined #openstack-infra | 04:57 | |
| *** btully has quit IRC | 04:57 | |
| *** sputnik13 has quit IRC | 04:58 | |
| *** sputnik13 has joined #openstack-infra | 04:59 | |
| *** sputnik13 has quit IRC | 04:59 | |
| *** ParsectiX has quit IRC | 05:13 | |
| *** gokrokve has quit IRC | 05:14 | |
| *** btully has joined #openstack-infra | 05:15 | |
| *** salv-orlando has joined #openstack-infra | 05:16 | |
| *** sdake has joined #openstack-infra | 05:18 | |
| *** leopoldj has joined #openstack-infra | 05:18 | |
| *** ddieterly has joined #openstack-infra | 05:18 | |
| *** achanda has quit IRC | 05:19 | |
| *** salv-orlando has quit IRC | 05:21 | |
| *** ddieterly has quit IRC | 05:23 | |
| *** harlowja_at_home has quit IRC | 05:24 | |
| *** deepakcs has joined #openstack-infra | 05:28 | |
| *** cloudm2 has quit IRC | 05:32 | |
| *** Sukhdev has quit IRC | 05:34 | |
| *** yamamoto has joined #openstack-infra | 05:34 | |
| *** fawadk has joined #openstack-infra | 05:37 | |
| *** fawadk has joined #openstack-infra | 05:37 | |
| *** fawadk is now known as fawadkhaliq_ | 05:37 | |
| *** fawadkhaliq has quit IRC | 05:39 | |
| *** sdake has quit IRC | 05:40 | |
| *** salv-orlando has joined #openstack-infra | 05:44 | |
| *** salv-orl_ has joined #openstack-infra | 05:47 | |
| *** yamamoto has quit IRC | 05:47 | |
| *** salv-orlando has quit IRC | 05:48 | |
| *** Ala has joined #openstack-infra | 05:51 | |
| *** achanda has joined #openstack-infra | 05:52 | |
| *** puranamr has quit IRC | 05:52 | |
| *** rlandy has joined #openstack-infra | 05:54 | |
| *** salv-or__ has joined #openstack-infra | 05:55 | |
| *** salv-orl_ has quit IRC | 05:55 | |
| *** salv-orl_ has joined #openstack-infra | 05:57 | |
| *** achanda_ has joined #openstack-infra | 05:58 | |
| *** achanda has quit IRC | 05:59 | |
| *** salv-or__ has quit IRC | 06:00 | |
| *** emagana has quit IRC | 06:00 | |
| *** koolhead17 has joined #openstack-infra | 06:01 | |
| *** emagana has joined #openstack-infra | 06:01 | |
| *** koolhead17 has quit IRC | 06:01 | |
| *** koolhead17 has joined #openstack-infra | 06:01 | |
| *** kaisers has joined #openstack-infra | 06:01 | |
| *** e0ne has joined #openstack-infra | 06:01 | |
| *** salv-orlando has joined #openstack-infra | 06:01 | |
| *** salv-orl_ has quit IRC | 06:02 | |
| *** kaisers has quit IRC | 06:02 | |
| *** yfried has joined #openstack-infra | 06:03 | |
| *** fawadkhaliq has joined #openstack-infra | 06:03 | |
| *** flashgordon is now known as jogo | 06:04 | |
| *** yamamoto has joined #openstack-infra | 06:04 | |
| *** jerryz has joined #openstack-infra | 06:05 | |
| *** emagana has quit IRC | 06:05 | |
| *** fawadkhaliq_ has quit IRC | 06:05 | |
| *** salv-orlando has quit IRC | 06:06 | |
| *** fawadk has joined #openstack-infra | 06:06 | |
| *** fijal has quit IRC | 06:06 | |
| *** kaisers has joined #openstack-infra | 06:06 | |
| *** salv-orlando has joined #openstack-infra | 06:06 | |
| *** BharatK has joined #openstack-infra | 06:08 | |
| *** fawadkhaliq has quit IRC | 06:08 | |
| *** unicell has quit IRC | 06:09 | |
| *** unicell has joined #openstack-infra | 06:09 | |
| *** e0ne has quit IRC | 06:11 | |
| *** fijal has joined #openstack-infra | 06:12 | |
| *** yamamoto has quit IRC | 06:14 | |
| *** unicell1 has joined #openstack-infra | 06:14 | |
| *** achanda_ has quit IRC | 06:15 | |
| *** unicell has quit IRC | 06:16 | |
| *** stevebaker has quit IRC | 06:16 | |
| *** stevebaker has joined #openstack-infra | 06:16 | |
| *** ildikov has quit IRC | 06:18 | |
| *** SumitNaiksatam has quit IRC | 06:19 | |
| *** hdd has joined #openstack-infra | 06:19 | |
| *** ddieterly has joined #openstack-infra | 06:19 | |
| *** SumitNaiksatam has joined #openstack-infra | 06:20 | |
| *** salv-orlando has quit IRC | 06:21 | |
| *** ayoung has quit IRC | 06:22 | |
| *** ddieterly has quit IRC | 06:24 | |
| *** salv-orl_ has joined #openstack-infra | 06:28 | |
| *** stevebaker has quit IRC | 06:30 | |
| *** kaisers has quit IRC | 06:30 | |
| *** AzherKhan has joined #openstack-infra | 06:30 | |
| *** salv-orlando has joined #openstack-infra | 06:31 | |
| *** stevebaker has joined #openstack-infra | 06:33 | |
| *** salv-or__ has joined #openstack-infra | 06:33 | |
| *** salv-orl_ has quit IRC | 06:33 | |
| *** shardy has joined #openstack-infra | 06:35 | |
| *** salv-orlando has quit IRC | 06:35 | |
| flip214 | Hi. I've got a devstack plugin, but check-tempest-dsvm-full-drbd-devstack says "Cloning not allowed in this configuration". what am I doing wrong? | 06:36 |
| *** hdd has quit IRC | 06:36 | |
| *** salv-or__ has quit IRC | 06:37 | |
| *** ZZelle_ has joined #openstack-infra | 06:39 | |
| *** maishsk_afk has joined #openstack-infra | 06:39 | |
| *** oomichi has joined #openstack-infra | 06:41 | |
| *** otter768 has joined #openstack-infra | 06:42 | |
| *** AzherKhan has quit IRC | 06:42 | |
| *** soren_ has joined #openstack-infra | 06:44 | |
| *** AzherKhan has joined #openstack-infra | 06:44 | |
| *** otter768 has quit IRC | 06:46 | |
| *** yfried is now known as yfried|afk | 06:48 | |
| *** salv-orlando has joined #openstack-infra | 06:49 | |
| *** yamamoto has joined #openstack-infra | 06:50 | |
| *** fawadkhaliq has joined #openstack-infra | 06:51 | |
| *** salv-orl_ has joined #openstack-infra | 06:53 | |
| *** oomichi has quit IRC | 06:53 | |
| *** salv-orlando has quit IRC | 06:53 | |
| *** fawadk has quit IRC | 06:54 | |
| *** yfried|afk is now known as yfried | 06:56 | |
| *** koolhead17 has quit IRC | 06:56 | |
| *** salv-orl_ has quit IRC | 06:58 | |
| *** salv-orlando has joined #openstack-infra | 06:58 | |
| *** mpaolino has joined #openstack-infra | 06:59 | |
| *** salv-orl_ has joined #openstack-infra | 07:00 | |
| *** achanda has joined #openstack-infra | 07:02 | |
| *** salv-orlando has quit IRC | 07:03 | |
| *** mpaolino has quit IRC | 07:04 | |
| *** claudiub has joined #openstack-infra | 07:04 | |
| *** salv-orl_ has quit IRC | 07:05 | |
| *** dbyte has quit IRC | 07:06 | |
| *** dbyte has joined #openstack-infra | 07:07 | |
| *** dbyte has quit IRC | 07:07 | |
| *** dbite has quit IRC | 07:07 | |
| *** dguitarbite has quit IRC | 07:07 | |
| *** dguitarbite has joined #openstack-infra | 07:07 | |
| *** yolanda has quit IRC | 07:10 | |
| *** mpavone has joined #openstack-infra | 07:11 | |
| *** fijal_ has joined #openstack-infra | 07:11 | |
| *** k4n0 has joined #openstack-infra | 07:11 | |
| *** fijal has quit IRC | 07:14 | |
| *** AzherKhan has quit IRC | 07:15 | |
| *** markvoelker has joined #openstack-infra | 07:17 | |
| openstackgerrit | Sergey Kraynev proposed openstack-infra/project-config: Enable Heat services for functional tests https://review.openstack.org/182974 | 07:18 |
| *** stevemar has quit IRC | 07:20 | |
| *** ddieterly has joined #openstack-infra | 07:20 | |
| *** panda has quit IRC | 07:21 | |
| *** panda has joined #openstack-infra | 07:22 | |
| *** fijal_ is now known as fijal | 07:22 | |
| *** jtomasek has quit IRC | 07:23 | |
| *** salv-orlando has joined #openstack-infra | 07:23 | |
| *** AzherKhan has joined #openstack-infra | 07:23 | |
| *** stevebaker has quit IRC | 07:24 | |
| *** stevebaker has joined #openstack-infra | 07:24 | |
| *** ddieterly has quit IRC | 07:25 | |
| *** emagana has joined #openstack-infra | 07:25 | |
| *** annegentle has joined #openstack-infra | 07:27 | |
| *** salv-orlando has quit IRC | 07:27 | |
| *** salv-orlando has joined #openstack-infra | 07:27 | |
| *** emagana has quit IRC | 07:30 | |
| *** salv-orlando has quit IRC | 07:31 | |
| *** sshnaidm has joined #openstack-infra | 07:31 | |
| *** annegentle has quit IRC | 07:32 | |
| *** salv-orlando has joined #openstack-infra | 07:32 | |
| *** markus_z has joined #openstack-infra | 07:32 | |
| *** arxcruz has joined #openstack-infra | 07:33 | |
| *** pasquier-s has joined #openstack-infra | 07:34 | |
| *** salv-orl_ has joined #openstack-infra | 07:35 | |
| *** salv-orlando has quit IRC | 07:36 | |
| *** achanda has quit IRC | 07:36 | |
| *** Hal has joined #openstack-infra | 07:37 | |
| *** dtantsur|afk is now known as dtantsur | 07:37 | |
| *** Hal is now known as Guest36182 | 07:38 | |
| *** achanda has joined #openstack-infra | 07:38 | |
| *** dims has joined #openstack-infra | 07:38 | |
| *** dguerri`away is now known as dguerri | 07:40 | |
| *** jlanoux has joined #openstack-infra | 07:42 | |
| *** dims has quit IRC | 07:44 | |
| *** chlong has quit IRC | 07:45 | |
| *** achanda has quit IRC | 07:47 | |
| *** dalgaaf has joined #openstack-infra | 07:48 | |
| *** dizquierdo has joined #openstack-infra | 07:49 | |
| *** fawadkhaliq has quit IRC | 07:50 | |
| lxsli | https://review.openstack.org/185620 https://review.openstack.org/185632 | 07:51 |
| lxsli | ^-- couple of elastic recheck queries | 07:51 |
| *** jistr has joined #openstack-infra | 07:52 | |
| *** devvesa has joined #openstack-infra | 07:56 | |
| *** achanda has joined #openstack-infra | 08:00 | |
| openstackgerrit | Fabien Boucher proposed openstack-infra/puppet-zuul: Add an acceptance test with beaker for puppet-zuul https://review.openstack.org/180214 | 08:03 |
| *** emagana has joined #openstack-infra | 08:04 | |
| *** matbu_afk has quit IRC | 08:05 | |
| *** AzherKhan has quit IRC | 08:06 | |
| *** emagana has quit IRC | 08:09 | |
| *** matbu has joined #openstack-infra | 08:10 | |
| *** HeOS has quit IRC | 08:11 | |
| *** matbu has quit IRC | 08:11 | |
| *** matbu has joined #openstack-infra | 08:14 | |
| *** e0ne has joined #openstack-infra | 08:15 | |
| *** fijal has quit IRC | 08:15 | |
| BobBall | I have a very stupid question... How is the pypi mirror setup for nodepool nodes? I see the configure_mirror script but that only seems to be called by multinode_setup which I'm definitely not calling... Yet my pypi mirrors are trying to use the openstack.org ones which is wrong | 08:15 |
| *** e0ne has quit IRC | 08:19 | |
| *** ddieterly has joined #openstack-infra | 08:21 | |
| openstackgerrit | Fabien Boucher proposed openstack-infra/puppet-zuul: Add an acceptance test with beaker for puppet-zuul https://review.openstack.org/180214 | 08:22 |
| *** ildikov has joined #openstack-infra | 08:22 | |
| *** btully has quit IRC | 08:25 | |
| *** ddieterly has quit IRC | 08:25 | |
| *** ZZelle has quit IRC | 08:27 | |
| *** ZZelle has joined #openstack-infra | 08:27 | |
| *** mpavone has quit IRC | 08:31 | |
| maishsk_afk | Q: is it possible to get a list of all etherpads under https://etherpad.openstack.org ? | 08:34 |
| *** maishsk_afk is now known as maishsk | 08:34 | |
| *** heyongli has joined #openstack-infra | 08:34 | |
| *** fijal has joined #openstack-infra | 08:35 | |
| *** mpavone has joined #openstack-infra | 08:37 | |
| *** markus_z has quit IRC | 08:38 | |
| *** otter768 has joined #openstack-infra | 08:43 | |
| *** jerryz has quit IRC | 08:44 | |
| *** otter768 has quit IRC | 08:47 | |
| *** markus_z has joined #openstack-infra | 08:51 | |
| openstackgerrit | Fabien Boucher proposed openstack-infra/puppet-zuul: Fix wrong indent reported by recent puppet-lint https://review.openstack.org/185889 | 08:52 |
| *** fhubik has joined #openstack-infra | 08:53 | |
| *** mrmartin has joined #openstack-infra | 08:53 | |
| openstackgerrit | Fabien Boucher proposed openstack-infra/puppet-zuul: Add an acceptance test with beaker for puppet-zuul https://review.openstack.org/180214 | 08:54 |
| *** u_glide has joined #openstack-infra | 08:54 | |
| *** shardy_ has joined #openstack-infra | 08:55 | |
| *** shardy has quit IRC | 08:57 | |
| *** achanda has quit IRC | 08:58 | |
| *** ihrachyshka has joined #openstack-infra | 08:58 | |
| *** emagana has joined #openstack-infra | 08:59 | |
| GheRivero | morning all | 08:59 |
| *** lifeless_ has joined #openstack-infra | 09:00 | |
| *** lifeless has quit IRC | 09:00 | |
| *** shardy_ has quit IRC | 09:01 | |
| *** dizquierdo has quit IRC | 09:01 | |
| *** shardy has joined #openstack-infra | 09:01 | |
| *** emagana has quit IRC | 09:03 | |
| *** mpavone has quit IRC | 09:04 | |
| openstackgerrit | Joshua Hesketh proposed stackforge/gertty: Only enqueue a task if its not already there https://review.openstack.org/185890 | 09:06 |
| openstackgerrit | Joshua Hesketh proposed stackforge/gertty: Perform multiple sync tasks asynchronously https://review.openstack.org/185891 | 09:06 |
| *** freerunner has joined #openstack-infra | 09:07 | |
| *** teran has quit IRC | 09:09 | |
| *** fhubik is now known as fhubik_afk | 09:12 | |
| *** HeOS has joined #openstack-infra | 09:14 | |
| electrical | Morning GheRivero | 09:14 |
| *** amotoki_ has quit IRC | 09:14 | |
| *** freerunner has quit IRC | 09:16 | |
| *** ekarlso has quit IRC | 09:16 | |
| *** rossella_s has quit IRC | 09:19 | |
| *** rossella_s has joined #openstack-infra | 09:19 | |
| *** ddieterly has joined #openstack-infra | 09:22 | |
| *** ekarlso has joined #openstack-infra | 09:22 | |
| *** e0ne has joined #openstack-infra | 09:23 | |
| *** amotoki_ has joined #openstack-infra | 09:23 | |
| *** e0ne is now known as e0ne_ | 09:23 | |
| *** ddieterly has quit IRC | 09:26 | |
| *** jtomasek has joined #openstack-infra | 09:27 | |
| *** shardy_ has joined #openstack-infra | 09:28 | |
| *** hichihara has quit IRC | 09:28 | |
| *** shardy has quit IRC | 09:30 | |
| *** e0ne_ is now known as e0ne | 09:31 | |
| *** fhubik_afk is now known as fhubik | 09:33 | |
| *** mrmartin has quit IRC | 09:33 | |
| *** shardy_ has quit IRC | 09:34 | |
| *** shardy has joined #openstack-infra | 09:34 | |
| openstackgerrit | David TARDIVEL proposed openstack-infra/project-config: Create new stackforge project python-watcherclient https://review.openstack.org/185006 | 09:34 |
| *** dosaboy_ has quit IRC | 09:36 | |
| *** dosaboy has joined #openstack-infra | 09:36 | |
| *** derekh has joined #openstack-infra | 09:37 | |
| *** amotoki_ has quit IRC | 09:39 | |
| *** kaisers has joined #openstack-infra | 09:39 | |
| *** mattfarina has joined #openstack-infra | 09:41 | |
| *** mattfarina has quit IRC | 09:42 | |
| *** mpavone has joined #openstack-infra | 09:42 | |
| u_glide | Hello all! Could anyone help me with strange ECONNREFUSED error in gate? | 09:43 |
| openstackgerrit | Merged openstack-infra/irc-meetings: updated swift meeting time https://review.openstack.org/185447 | 09:43 |
| u_glide | #link http://logs.openstack.org/78/184778/7/check/gate-manila-tempest-dsvm-neutron/d7cea82/logs/screen-m-shr.txt.gz | 09:44 |
| *** dizquierdo has joined #openstack-infra | 09:44 | |
| u_glide | looks like apache refuses connection to keystone | 09:45 |
| u_glide | oh, I figured out problem :) | 09:47 |
| *** nikeshm has quit IRC | 09:48 | |
| *** Guest36182 has quit IRC | 09:51 | |
| *** emagana has joined #openstack-infra | 09:53 | |
| *** Hal has joined #openstack-infra | 09:56 | |
| *** Hal is now known as Guest59858 | 09:56 | |
| *** emagana has quit IRC | 09:57 | |
| openstackgerrit | Merged openstack-infra/irc-meetings: Remove Storyboard Meeting. https://review.openstack.org/185808 | 09:58 |
| openstackgerrit | Merged openstack-infra/irc-meetings: Update Docs team meeting https://review.openstack.org/185809 | 09:58 |
| openstackgerrit | Merged openstack-infra/irc-meetings: Update Chef cookbooks meeting https://review.openstack.org/182926 | 09:59 |
| *** maishsk has quit IRC | 10:00 | |
| openstackgerrit | Thierry Carrez proposed openstack-infra/irc-meetings: Change refstack Chair https://review.openstack.org/182925 | 10:03 |
| *** maishsk_afk has joined #openstack-infra | 10:04 | |
| openstackgerrit | Thierry Carrez proposed openstack-infra/irc-meetings: Update heat meeting https://review.openstack.org/185810 | 10:04 |
| *** freerunner has joined #openstack-infra | 10:04 | |
| openstackgerrit | Thierry Carrez proposed openstack-infra/irc-meetings: Add Ironic/Neutron meeting https://review.openstack.org/185811 | 10:05 |
| openstackgerrit | Merged openstack-infra/irc-meetings: Change refstack Chair https://review.openstack.org/182925 | 10:07 |
| openstackgerrit | Merged openstack-infra/irc-meetings: Update heat meeting https://review.openstack.org/185810 | 10:07 |
| *** teran has joined #openstack-infra | 10:08 | |
| openstackgerrit | Merged openstack-infra/irc-meetings: Add Ironic/Neutron meeting https://review.openstack.org/185811 | 10:09 |
| *** dims has joined #openstack-infra | 10:09 | |
| electrical | Some one is having a merge party :p | 10:09 |
| *** achanda has joined #openstack-infra | 10:09 | |
| *** mfmcdonagh has joined #openstack-infra | 10:09 | |
| *** yamamoto has quit IRC | 10:10 | |
| *** teran_ has joined #openstack-infra | 10:11 | |
| *** teran has quit IRC | 10:13 | |
| *** fhubik has quit IRC | 10:14 | |
| *** pblaho has quit IRC | 10:14 | |
| *** fhubik has joined #openstack-infra | 10:14 | |
| *** pblaho has joined #openstack-infra | 10:15 | |
| *** linuxmaniac has quit IRC | 10:21 | |
| *** ddieterly has joined #openstack-infra | 10:23 | |
| *** ddieterly has quit IRC | 10:27 | |
| *** _nadya_ has joined #openstack-infra | 10:33 | |
| *** linuxmaniac has joined #openstack-infra | 10:36 | |
| *** matbu has quit IRC | 10:38 | |
| *** matbu has joined #openstack-infra | 10:39 | |
| *** fhubik is now known as fhubik_afk | 10:42 | |
| *** akrivoka has joined #openstack-infra | 10:43 | |
| *** otter768 has joined #openstack-infra | 10:44 | |
| *** yfried is now known as yfried|afk | 10:44 | |
| *** ParsectiX has joined #openstack-infra | 10:45 | |
| *** zhiwei has quit IRC | 10:45 | |
| *** yfried|afk is now known as yfried | 10:45 | |
| *** emagana has joined #openstack-infra | 10:47 | |
| *** emagana has quit IRC | 10:48 | |
| *** otter768 has quit IRC | 10:48 | |
| *** emagana has joined #openstack-infra | 10:48 | |
| *** pc_m has joined #openstack-infra | 10:50 | |
| akrivoka | we seem to have a problem in with stable/kilo horizon patches | 10:53 |
| *** emagana has quit IRC | 10:53 | |
| akrivoka | https://review.openstack.org/#/q/status:open+project:openstack/horizon+branch:stable/kilo,n,z | 10:53 |
| akrivoka | grenade is failing for all of them in the same way, unrelated to the code in the patches | 10:54 |
| akrivoka | I think it is a package version conflict http://logs.openstack.org/56/183656/4/check/check-grenade-dsvm/3acba73/logs/grenade.sh.txt.gz#_2015-05-27_07_15_12_726 | 10:55 |
| akrivoka | but I don't know how to fix that issue - any help would be appreciated | 10:55 |
| *** samueldmq has joined #openstack-infra | 11:00 | |
| samueldmq | morning | 11:00 |
| *** ildikov has quit IRC | 11:01 | |
| jklare | hi, does anybody here know if a project renaming / move has to happen if one is currently under stackforge/.. and has become part of the big tent ? | 11:01 |
| *** weshay has joined #openstack-infra | 11:01 | |
| *** mrmartin has joined #openstack-infra | 11:04 | |
| *** freerunner has quit IRC | 11:10 | |
| *** freerunner has joined #openstack-infra | 11:12 | |
| *** freerunner has quit IRC | 11:15 | |
| *** ildikov has joined #openstack-infra | 11:16 | |
| *** dguitarbite has quit IRC | 11:18 | |
| *** achanda has quit IRC | 11:20 | |
| *** ddieterly has joined #openstack-infra | 11:23 | |
| *** achanda has joined #openstack-infra | 11:25 | |
| *** e0ne is now known as e0ne_ | 11:26 | |
| *** ddieterly has quit IRC | 11:29 | |
| *** openstack has joined #openstack-infra | 11:39 | |
| *** sshnaidm has quit IRC | 11:39 | |
| *** emagana has joined #openstack-infra | 11:43 | |
| mancdaz | hi peeps - any chance we could get some eyes on this review: https://review.openstack.org/#/c/153057/ | 11:44 |
| mancdaz | it's been kicking around for quite a while | 11:44 |
| *** emagana has quit IRC | 11:48 | |
| *** dprince has joined #openstack-infra | 11:52 | |
| *** sshnaidm has joined #openstack-infra | 11:55 | |
| *** yfried is now known as yfried|afk | 11:56 | |
| *** deepakcs has quit IRC | 11:56 | |
| *** yfried|afk is now known as yfried | 11:58 | |
| *** yamamoto has joined #openstack-infra | 11:59 | |
| *** baoli has joined #openstack-infra | 12:01 | |
| *** BharatK has quit IRC | 12:01 | |
| openstackgerrit | Fabien Boucher proposed openstack-infra/puppet-zuul: Add an acceptance test with beaker for puppet-zuul https://review.openstack.org/180214 | 12:03 |
| *** yamamoto has quit IRC | 12:04 | |
| *** markvoelker has quit IRC | 12:04 | |
| *** markvoelker has joined #openstack-infra | 12:04 | |
| *** freerunner has joined #openstack-infra | 12:05 | |
| *** ildikov has quit IRC | 12:07 | |
| *** e0ne is now known as e0ne_ | 12:09 | |
| *** e0ne_ is now known as e0ne | 12:14 | |
| *** aysyanne has joined #openstack-infra | 12:14 | |
| *** BharatK has joined #openstack-infra | 12:15 | |
| openstackgerrit | Fabien Boucher proposed openstack-infra/project-config: Add beaker-rspec jobs to puppet-zuul repo https://review.openstack.org/185952 | 12:16 |
| *** bswartz has quit IRC | 12:17 | |
| *** amotoki_ has joined #openstack-infra | 12:19 | |
| *** BharatK has quit IRC | 12:19 | |
| *** achanda has quit IRC | 12:20 | |
| *** ildikov has joined #openstack-infra | 12:21 | |
| *** ddieterly has joined #openstack-infra | 12:24 | |
| *** ddieterly has quit IRC | 12:29 | |
| *** yolanda has joined #openstack-infra | 12:29 | |
| yolanda | hi, i have an user that created a project with a non-accessible branch, and now he cannot land any change due to Remote %s did not have any references | 12:30 |
| yolanda | how can i fix that? (not deleting the project) | 12:30 |
| *** mwagner_lap has joined #openstack-infra | 12:31 | |
| *** dims has quit IRC | 12:31 | |
| *** mwagner_lap is now known as mwagner | 12:31 | |
| *** miqui has joined #openstack-infra | 12:32 | |
| *** olaph has joined #openstack-infra | 12:32 | |
| *** dims has joined #openstack-infra | 12:32 | |
| *** ParsectiX has quit IRC | 12:33 | |
| *** rbradfor_ has quit IRC | 12:34 | |
| *** rbradfor_ has joined #openstack-infra | 12:35 | |
| *** emagana has joined #openstack-infra | 12:37 | |
| yolanda | fungi, you know? ^ | 12:39 |
| *** akshai has joined #openstack-infra | 12:40 | |
| *** eharney has quit IRC | 12:40 | |
| electrical | yolanda: is that in gerrit? can't change the branch name? | 12:40 |
| yolanda | electrical, yes , it's happening on gerrit | 12:40 |
| *** mrmartin has quit IRC | 12:40 | |
| yolanda | master branch | 12:41 |
| yolanda | just any change can land due to that, for that project | 12:41 |
| *** emagana has quit IRC | 12:41 | |
| electrical | yolanda: hm okay. and you can't change it in gerrit any more then after creating the project? | 12:41 |
| *** annegentle has joined #openstack-infra | 12:41 | |
| yolanda | if i create a new branch, it tries to point to HEAD and shows invalid revision | 12:42 |
| yolanda | something went really wrong in the creation of that project, i'm just thinking in removing | 12:42 |
| electrical | Ah okay, i see. hmm. good question. | 12:42 |
| *** mrunge has joined #openstack-infra | 12:44 | |
| *** mrunge_ has joined #openstack-infra | 12:44 | |
| *** otter768 has joined #openstack-infra | 12:44 | |
| electrical | so the master branch doesn't exist anymore right? i think it means that in the .git/config file there is a reference to the master branch. you could edit that file and change it. but not sure if that's possible with gerrit :-( | 12:44 |
| *** mrunge has quit IRC | 12:45 | |
| *** mrunge_ has quit IRC | 12:45 | |
| electrical | must admit i only used gerrit so far as a user ( for reviews ) | 12:45 |
| *** kaisers has quit IRC | 12:45 | |
| *** mrunge has joined #openstack-infra | 12:45 | |
| TheJulia | yolanda: It kind of sounds like their git repo is in a very bad state | 12:46 |
| yolanda | so yes, very bad state | 12:47 |
| *** fijal has quit IRC | 12:47 | |
| yolanda | i removed the project and running manage-projects again | 12:48 |
| *** julim has joined #openstack-infra | 12:48 | |
| *** otter768 has quit IRC | 12:49 | |
| *** prad has joined #openstack-infra | 12:49 | |
| *** fijal has joined #openstack-infra | 12:50 | |
| *** marcusvrn has joined #openstack-infra | 12:52 | |
| *** ekarlso has quit IRC | 12:53 | |
| *** ekarlso has joined #openstack-infra | 12:53 | |
| *** xyang1 has joined #openstack-infra | 12:56 | |
| *** sks has quit IRC | 12:56 | |
| *** gordc has joined #openstack-infra | 12:58 | |
| electrical | yolanda: stupid computers ;-) | 12:59 |
| *** fawadkhaliq has joined #openstack-infra | 13:00 | |
| yolanda | electrical , there are these days when nothing wants to work... | 13:01 |
| electrical | know that feeling yeah | 13:01 |
| openstackgerrit | Monty Taylor proposed openstack-infra/nodepool: Consume os-client-config directly https://review.openstack.org/185972 | 13:02 |
| *** gaelL has quit IRC | 13:02 | |
| *** mwhagedorn has joined #openstack-infra | 13:02 | |
| mordred | fungi, clarkb: ^^ that should take care of the "notice when the config has changed without a restart" thing | 13:04 |
| *** ayoung has joined #openstack-infra | 13:04 | |
| *** dboik has joined #openstack-infra | 13:05 | |
| *** annegentle has quit IRC | 13:06 | |
| *** bknudson has joined #openstack-infra | 13:06 | |
| *** hashar has joined #openstack-infra | 13:06 | |
| *** macjack has joined #openstack-infra | 13:07 | |
| *** sks has joined #openstack-infra | 13:10 | |
| *** ddieterly has joined #openstack-infra | 13:11 | |
| waynr | electrical: yo | 13:12 |
| electrical | waynr: hiya bud | 13:12 |
| waynr | you mind if i take a crack at the bug you reported yesterday? not sure if you wanted to work on it yourself | 13:12 |
| *** radez_g0n3 is now known as radez | 13:13 | |
| openstackgerrit | Monty Taylor proposed openstack-infra/shade: Turn server object to dict so we can pop things https://review.openstack.org/185976 | 13:13 |
| *** ddieterly has quit IRC | 13:13 | |
| *** shardy_ has joined #openstack-infra | 13:14 | |
| *** shardy has quit IRC | 13:15 | |
| openstackgerrit | Christian Berendt proposed openstack-infra/system-config: Use the correct name for the "Repository Creator's Guide" https://review.openstack.org/185978 | 13:15 |
| *** salv-orl_ has quit IRC | 13:17 | |
| *** mriedem_away is now known as mriedem | 13:17 | |
| *** pradk has joined #openstack-infra | 13:18 | |
| *** shardy_ has quit IRC | 13:19 | |
| *** shardy has joined #openstack-infra | 13:20 | |
| *** tiswanso has joined #openstack-infra | 13:22 | |
| *** dpaterson has joined #openstack-infra | 13:22 | |
| fungi | mordred: i spy a typo in your 185972 change | 13:23 |
| *** doug-fish has joined #openstack-infra | 13:23 | |
| fungi | yolanda: replacing the backend git repo sounds like the right solution, if it only got as far as importing/creating it (or perhaps not even that far?) | 13:25 |
| fungi | either replacing it by hand or letting manage-projects reclone it | 13:25 |
| yolanda | fungi, i removed the project, just deleting from gerrit, and executed manage-projects | 13:25 |
| yolanda | but it is is buggy anyway | 13:25 |
| yolanda | it creates a project with branch=HEAD, revision=master | 13:26 |
| yolanda | and cannot create new branches | 13:26 |
| fungi | did it import a repo lacking a "master" branch? | 13:26 |
| *** dustins has joined #openstack-infra | 13:26 | |
| openstackgerrit | Monty Taylor proposed openstack-infra/nodepool: Consume os-client-config directly https://review.openstack.org/185972 | 13:26 |
| *** macjack has left #openstack-infra | 13:26 | |
| fungi | i thought we had safeguards in m-p against that (we'd previously just said we didn't support importing projects with no master branch) | 13:26 |
| yolanda | yes, it did | 13:26 |
| yolanda | maybe we have an outdated version of jeepyb, but i woudn't say | 13:27 |
| fungi | so, anyway, the solution is to fix the repo being imported (rename their default branch to master or something) | 13:27 |
| fungi | and then delete it from gerrit and replace it or let manage-projects replace it | 13:27 |
| yolanda | fungi, they just removed the upstream ref from that project | 13:27 |
| yolanda | as it was pointing to a pass-protected repo, that was the root cause | 13:28 |
| yolanda | but seems i cannot get it back to life | 13:28 |
| *** annegentle has joined #openstack-infra | 13:28 | |
| openstackgerrit | Monty Taylor proposed openstack-infra/shade: Turn server object to dict so we can pop things https://review.openstack.org/185976 | 13:29 |
| *** _nadya_ has quit IRC | 13:29 | |
| *** eharney has joined #openstack-infra | 13:30 | |
| *** annegentle has quit IRC | 13:30 | |
| *** abregman has quit IRC | 13:30 | |
| *** _nadya_ has joined #openstack-infra | 13:31 | |
| *** _nadya_ has quit IRC | 13:31 | |
| *** oomichi has joined #openstack-infra | 13:31 | |
| *** emagana has joined #openstack-infra | 13:31 | |
| *** sdake_ has joined #openstack-infra | 13:31 | |
| *** amitgandhinz has joined #openstack-infra | 13:33 | |
| *** _nadya_ has joined #openstack-infra | 13:33 | |
| *** annegentle has joined #openstack-infra | 13:33 | |
| electrical | waynr: sorry, had a meeting. please go ahead. I've been looking at it my self but no idea where to start tbh | 13:33 |
| *** yfried is now known as yfried|afk | 13:33 | |
| waynr | electrical: this code is seriously tangled, i don't think the bug is actually in applyDefaults | 13:33 |
| *** _nadya_ has quit IRC | 13:33 | |
| *** jtomasek has quit IRC | 13:34 | |
| electrical | waynr: i tend to agree. it picks up most things from like namespace 'puppet' and if its not there it picks up global name space correctly | 13:34 |
| electrical | so i think its the sequence of the other parts thats wrong | 13:34 |
| waynr | i believe the bug may be somewhere in expandYaml or expandYamlForTemplatejob | 13:34 |
| *** _nadya_ has joined #openstack-infra | 13:34 | |
| *** emagana has quit IRC | 13:35 | |
| waynr | right now i am just pprinting stuff that gets passed into applyDefaults and i can see at least one instance where the "data" parameter appears to be a "project" dictionary | 13:36 |
| *** jsavak has joined #openstack-infra | 13:36 | |
| waynr | wait scratch that i am just confused | 13:37 |
| electrical | waynr: i did some extra logging as well when running my tests and it gets very confusing at some point. I was digging into the expandYamlForTemplateJob code di | 13:38 |
| *** zz_jgrimm is now known as jgrimm | 13:38 | |
| *** peristeri has joined #openstack-infra | 13:39 | |
| waynr | okay yeah it looks like the project named "project" in concat_defaults004 is being passed as the "data" to applyDefaults | 13:39 |
| electrical | waynr: the template var in there holds the correct vars but it gets lost somewhere at some point. | 13:39 |
| *** dkranz has joined #openstack-infra | 13:41 | |
| *** links has quit IRC | 13:41 | |
| openstackgerrit | Thierry Carrez proposed openstack-infra/irc-meetings: Remove redundant/stale scheduling info from desc https://review.openstack.org/185988 | 13:42 |
| *** cloudm2 has joined #openstack-infra | 13:42 | |
| openstackgerrit | John Griffith proposed openstack-infra/project-config: Revert "Stop running non-voting docker job on nova" https://review.openstack.org/185992 | 13:43 |
| electrical | waynr: 004 barfs at expanded = deep_format(template, params, allow_empty_variables) | 13:44 |
| electrical | Where params: {'name': 'project', 'description': 'my project description'} | 13:45 |
| *** abregman has joined #openstack-infra | 13:45 | |
| electrical | and template; {'confversion': 'v0.42', 'fooversion': 'v0.10', 'name': 'job-template', 'builders': [OrderedDict([('shell', "echo 'confVersion: {confversion} & fooVersion: {fooversion}'")])], 'defaults': 'myproject', 'description': 'my project description'} | 13:45 |
| waynr | fg | 13:46 |
| waynr | derp | 13:46 |
| *** sdake has joined #openstack-infra | 13:47 | |
| *** _nadya_ has quit IRC | 13:48 | |
| *** jtomasek has joined #openstack-infra | 13:50 | |
| *** _nadya_ has joined #openstack-infra | 13:50 | |
| *** sdake_ has quit IRC | 13:51 | |
| *** annegentle has quit IRC | 13:52 | |
| *** annegentle has joined #openstack-infra | 13:53 | |
| *** amotoki_ has quit IRC | 13:54 | |
| *** dustins_ has joined #openstack-infra | 13:56 | |
| *** pradk has quit IRC | 13:56 | |
| *** sks has quit IRC | 13:58 | |
| *** abregman has quit IRC | 13:59 | |
| openstackgerrit | Thierry Carrez proposed openstack-infra/irc-meetings: Small improvements in template https://review.openstack.org/186004 | 13:59 |
| electrical | waynr: getJobTemplate does get all the right defaults as far as i can see. | 13:59 |
| *** dustins has quit IRC | 14:00 | |
| waynr | yeah | 14:00 |
| *** matbu has quit IRC | 14:00 | |
| waynr | i'm gonna try reverting this allow_empty_variables patch just out of curiosity | 14:00 |
| *** pradk has joined #openstack-infra | 14:01 | |
| waynr | d'oh...that came before the code reorganization | 14:02 |
| *** yfried|afk is now known as yfried | 14:02 | |
| *** k4n0 has quit IRC | 14:02 | |
| *** zz_dimtruck is now known as dimtruck | 14:03 | |
| electrical | waynr: i could for fun apply https://review.openstack.org/#/c/179459/ and see what happens? | 14:04 |
| electrical | or the part with your comment | 14:04 |
| *** chlong has joined #openstack-infra | 14:05 | |
| *** oomichi_ has joined #openstack-infra | 14:05 | |
| waynr | go for it | 14:06 |
| *** sigmavirus24_awa is now known as sigmavirus24 | 14:06 | |
| *** matbu has joined #openstack-infra | 14:07 | |
| electrical | still fails | 14:07 |
| waynr | yeah all that patch does is reorder parameter overrides | 14:07 |
| electrical | it has to do with the deep_merge thing. it doesn't seem to merge it correctly | 14:08 |
| *** dalgaaf has quit IRC | 14:08 | |
| electrical | sorry, deep_format | 14:08 |
| *** tonytan4ever has joined #openstack-infra | 14:09 | |
| *** oomichi has quit IRC | 14:09 | |
| waynr | i played around with the concat_defaults004.yaml example and it ceases to faile when setting the "myproject" defaults on either the project or the job-group level | 14:09 |
| electrical | so it works on job-template? | 14:09 |
| waynr | yeah, the defaults are being applied correctly on job-template | 14:10 |
| electrical | you have made a change then i assume? because with current master it fails | 14:10 |
| *** dustins_ is now known as dustins | 14:10 | |
| waynr | yeah what i am saying is that i modify the testcase concat_defaults004.yaml | 14:10 |
| *** dustins is now known as dustins_away | 14:10 | |
| *** bswartz has joined #openstack-infra | 14:10 | |
| waynr | just to try different situations to get a feel for what might be going wrong | 14:10 |
| *** abregman has joined #openstack-infra | 14:11 | |
| waynr | whenever i set "defaults: myproject" on either the project JJB object or the job-group JJB object the example passes | 14:11 |
| *** koolhead17 has joined #openstack-infra | 14:11 | |
| electrical | ah okay. 004, 005 and 006 do the same but only with a moved defaults location ( job template, job-group and project ) and they all fail with current master | 14:11 |
| clarkb | mordred: looks like all 3 regions have gateway timeouts so not just an ord thing | 14:11 |
| waynr | the problem isn't in the deep_format function itself bu somewhere in expandYamlForTemplateJob | 14:12 |
| waynr | hmm i just deleted all the other examples in your patch other than 004 | 14:12 |
| electrical | ah okay | 14:12 |
| openstackgerrit | Flavio Percoco proposed openstack-infra/project-config: Disable py26 jos for glance_store https://review.openstack.org/186011 | 14:13 |
| electrical | 002 and 003 are a clone of 001 but with the defaults part moved.. 004, 005 and 006 tests are the same but also with moved default location and no defaults set in the global namespace | 14:13 |
| waynr | electrical: i think 005 and 006 fail because no defaults are set on the job-template itself which is a differnt bug | 14:13 |
| electrical | yeah. the job-group and project don't pass the defaults correctly further down. | 14:14 |
| openstackgerrit | Merged openstack-infra/devstack-gate: Add manila-ui to cloned projects https://review.openstack.org/184788 | 14:14 |
| electrical | which is an other bug indeed. | 14:14 |
| openstackgerrit | Jeremy Stanley proposed openstack-infra/project-config: Alter governance repo voting rules https://review.openstack.org/185785 | 14:14 |
| electrical | 004 is sort of a clone of 001 ( which passes ) but without the vars in the global name space | 14:14 |
| waynr | i think the real bug is all the lack of intentional design put into scoping of various JJB constructs throughout the YamlParser | 14:15 |
| electrical | hmm yeah.. | 14:15 |
| *** btully has joined #openstack-infra | 14:15 | |
| *** oomichi has joined #openstack-infra | 14:16 | |
| waynr | it'd be really nice to just to come up with a well-defined spec that describes the behavior of JJB Yaml and re-build the implementation to match that spec from the ground up | 14:16 |
| electrical | yeah, but that's not something that can be done overnight :-) | 14:16 |
| waynr | every time somone adds a new feature or fixes an old bug there is some corner case that no one considered that can lead to error states | 14:16 |
| waynr | s/can/will/ | 14:17 |
| *** annegentle has quit IRC | 14:17 | |
| electrical | from what i noticed with this bug at least is that it tries to merge myproject defaults with global defaults.. and if it can't find the key in the global defaults it will just ignore it ( or something like this ) | 14:17 |
| waynr | we could at least put a moratorium on new yamlparser features until we have a chance to rewrite the current implementation | 14:17 |
| *** leopoldj has quit IRC | 14:17 | |
| *** oomichi_ has quit IRC | 14:18 | |
| openstackgerrit | Andreas Jaeger proposed openstack-infra/project-config: Add project-team-guide https://review.openstack.org/185746 | 14:18 |
| openstackgerrit | Jeremy Stanley proposed openstack-infra/project-config: Update infra-specs voting rules https://review.openstack.org/185792 | 14:19 |
| clarkb | mordred: my hunch is this is related to image size and upload time vs whatever proxy is sitting infront of swift | 14:19 |
| clarkb | but considering that we got rate limit 413s asthe error when using glance v1 that is probably completely wrong | 14:19 |
| mordred | clarkb: AWESOME | 14:21 |
| electrical | waynr: i agree.. your https://review.openstack.org/#/c/182542/ should spark the rewrite part i guess. | 14:21 |
| mordred | I'm going to go bug the swift people | 14:21 |
| *** timcline has joined #openstack-infra | 14:21 | |
| fungi | clarkb: any attempt to reason about openstack error codes is likely doomed to failure and madness | 14:22 |
| waynr | electrical: well that's mostly targeted at solidifying the programmatic API, i have no intention of rewriting the YamlParser in relation to that spec | 14:22 |
| electrical | waynr: ah okay, i see. well, perhaps we should include that part in there for a 2.0 release :-) | 14:23 |
| waynr | well once the API is solidified we should be able to rewrite the yamlparser more easily without worrying about how its implementation affects the rest of the program | 14:23 |
| clarkb | mordred: as an aside, if we get image uploads to dfw working then get swift object passthrough on logs.o.o working we will have solved the image publishing problem for the most part | 14:23 |
| mordred | ya | 14:23 |
| mordred | I'll be super happy about that | 14:24 |
| electrical | waynr: ah i see, okay | 14:24 |
| mordred | clarkb: btw - I have patches up for all of the things | 14:24 |
| clarkb | mordred: http://logs.openstack.org/images/ will host our images, but we won't do index generation for it so indexing is the last bit to figure out | 14:24 |
| waynr | so we can release 2.0.0 then begin working on yamlparser rewrite for a minor release of 2.x series that is backwards-compatible with 2.0.0 | 14:24 |
| clarkb | mordred: ya, the nodepool one failed tests pretty hard, still digging through oscc before getting to that one | 14:24 |
| *** jokke_ has quit IRC | 14:24 | |
| *** jokke_ has joined #openstack-infra | 14:24 | |
| mordred | clarkb: I did not do your suggestion on the config refresh because of how nodepool wants to think about it | 14:24 |
| mordred | clarkb: although I did also start to add the plumbing to do the way you suggested | 14:25 |
| clarkb | mordred: you actually did my original suggestion | 14:25 |
| mordred | I did? | 14:25 |
| mordred | neat | 14:25 |
| waynr | although to be honest once the API is cleaned up i am hoping to begin working on a pure-python approach to configuring jenkins jobs | 14:25 |
| clarkb | mordred: which was that nodepool should talk to oscc directly | 14:25 |
| waynr | which will remove the need for me to worry about the yamlparser | 14:25 |
| mordred | ah. I did not catch that as a suggestion. and yes - I completely agree having dug into that bit of code | 14:25 |
| electrical | waynr: how do you mean pure python approach ? | 14:25 |
| openstackgerrit | Yanis Guenane proposed openstack-infra/tripleo-ci: Pin puppetlabs-mysql https://review.openstack.org/186016 | 14:25 |
| clarkb | mordred: it was part of my why do we have a clouds.yaml anyways. Though I did think we should do it that way so that nodepool's cloud key could point to what may soon be called profile | 14:26 |
| waynr | i mean like "job = JenkinsJob(some_dict)" where some_dict contains the basic key-value pair necessary to generate XML for a job | 14:26 |
| ttx | mordred: ok, I think with those two things we are ready to go live on irc-meetings. preparing launch email. Already updated wiki page | 14:27 |
| *** rbradfor_ is now known as rbradfor | 14:27 | |
| ttx | I'll wait for the new template to be picked up | 14:27 |
| openstackgerrit | Merged openstack-infra/irc-meetings: Remove redundant/stale scheduling info from desc https://review.openstack.org/185988 | 14:27 |
| waynr | this JenkinsJob object (subclass of OrderedDict) can then be passed to the upcoming XmlBuilder class to actually generate the correct XML | 14:27 |
| openstackgerrit | Merged openstack-infra/irc-meetings: Small improvements in template https://review.openstack.org/186004 | 14:27 |
| mordred | clarkb: oh - so - the other reason we want clouds.yaml and not just direct config in nodepool.yaml is so that we can handle cloud providers that do more advanced keystone things without having to teach nodepool all of that logic | 14:28 |
| *** ddieterly has joined #openstack-infra | 14:28 | |
| electrical | waynr: ah, i think i understand what you mean. | 14:28 |
| waynr | I haven't put a great deal of thought into the programmatic interface I want to use for these python classes but they should be arbitrarily extensible using the full power of the Python language | 14:28 |
| mordred | clarkb: for instance, unitedstack is a keystone v3 cloud, so you have to pass user_domain_id and project_domain_id as well | 14:28 |
| mordred | but you only need to do that for them | 14:28 |
| *** wenlock has joined #openstack-infra | 14:29 | |
| mordred | or if we grow a cloud provider that decides to use a different keystone auth plugin than password | 14:29 |
| waynr | it will also allow for arbitrary abstractions to represent relationships between jobs such as a SerialParameterizedPipeline class which could be a sublcass of "list" | 14:29 |
| mordred | (we could decide to use kerberos for infra cloud :) ) | 14:30 |
| clarkb | mordred: except we are teaching nodepool about all that in your change aren't we? | 14:30 |
| mordred | we are not | 14:30 |
| clarkb | mordred: its the way we figure out if we need a new provider manager | 14:30 |
| fungi | mordred: kerberos is keystone v4 from what i hear | 14:30 |
| waynr | rather than trying to squeeze all kinds of strange and competing logic in the YamlParser | 14:30 |
| *** bnemec has quit IRC | 14:30 | |
| mordred | clarkb: we compare cloud configs in my change | 14:30 |
| openstackgerrit | Merged openstack-infra/project-config: Add barbican-dogtag flag https://review.openstack.org/185183 | 14:30 |
| electrical | waynr: ah okay. well, i think we can agree that in the current state of the code it will be near impossible to fix this defaults merging behaviour right? | 14:30 |
| waynr | oh i don't know about that | 14:30 |
| mordred | the direct username/password comparison is in the old code | 14:30 |
| *** claudiub has quit IRC | 14:31 | |
| ttx | fungi: I don't remember if we wanted to CNAME eavesdrop before we announce... | 14:31 |
| waynr | i think we can probably fix it but it's just really confusing | 14:31 |
| waynr | well | 14:31 |
| *** liuxg has joined #openstack-infra | 14:31 | |
| electrical | waynr: hehe i already gave up after a few hours of debugging this morning. | 14:31 |
| fungi | ttx: we don't have any different vhost for it right now anyway | 14:31 |
| mordred | clarkb: in my change we pass the username/password/auth-url to the client config constructor, and then compare old and new client configs - that means ultimately the only thing the infra production case will be testing will be client configs | 14:31 |
| clarkb | mordred: I see so you are saying the only way we would support non passwd auth is via oscc clouds.yaml? (your change does directly handle passwords) | 14:31 |
| waynr | i have an idea or two i wanna try, will know in 20 minutes or so if it works | 14:31 |
| fungi | ttx: we can always run a url redirect at the old name if we decide to change it later | 14:31 |
| mordred | clarkb: yes | 14:31 |
| electrical | waynr: cool :-) | 14:31 |
| waynr | but i'm only gonna try to fix the 004 example | 14:31 |
| waynr | well maybe 005 and 006 | 14:32 |
| electrical | yeah indeed | 14:32 |
| mordred | clarkb: or, we'd have to teach nodepool all about the ins and outs of all the different ways to configure different keystone auth types | 14:32 |
| ttx | fungi: so I should just point to eavesdrop for the moment ? | 14:32 |
| mordred | which I promise we do not want to do | 14:32 |
| electrical | if you would solve 005/006 then you would also fix 002/003 since they represent the same issue ( location of the defaults definition ) | 14:32 |
| clarkb | mordred: ok, so we keep what it knows today for backward compat and simplicity, but moving forward extra fancyness stays in oscc | 14:32 |
| mordred | yah | 14:32 |
| *** annegentle has joined #openstack-infra | 14:32 | |
| fungi | ttx: yeah, http://eavesdrop.openstack.org/ seems fine for now. if people really want a different name, i can do whatever's necessary to make that happen later once the bikeshedding has subsided | 14:32 |
| waynr | oh | 14:32 |
| electrical | 004 with the defaults in the job-template would be easiest them i guess. | 14:32 |
| ttx | ack | 14:33 |
| mordred | clarkb: I'll try to capture that in a "why" comment somewhere | 14:33 |
| *** esker has joined #openstack-infra | 14:33 | |
| *** mestery has quit IRC | 14:33 | |
| *** mtanino has joined #openstack-infra | 14:33 | |
| clarkb | mordred: fwiw I think pep8 test may have caught the reason the tests failed | 14:34 |
| mordred | clarkb: oh - and wow, yeah. nodepool tests failed HARD. fixing now | 14:34 |
| mordred | clarkb: well, they failed because I suck | 14:34 |
| *** mestery has joined #openstack-infra | 14:34 | |
| clarkb | well specifically because an undefined name was used :) | 14:34 |
| fungi | ttx: who added the timeanddate.com hyperlinks? that's pretty nifty | 14:34 |
| clarkb | mordred: btw mysql in docker for nodepool tests is pretty handy | 14:34 |
| mordred | clarkb: ++ | 14:35 |
| *** bnemec has joined #openstack-infra | 14:35 | |
| mordred | clarkb: I'm a big fan of the idea of docker containers with our database configs in them | 14:35 |
| ttx | fungi: they were there on the original template I wrote | 14:35 |
| *** dims has quit IRC | 14:35 | |
| mordred | clarkb: and kinda want to publish docker containers with the mysql and postgres databases with openstackci user added to them | 14:35 |
| fungi | this is probably the first time i've looked at it while not braindead | 14:35 |
| yolanda | hi, we need to have this setting on place : VIRTUALENV_NO_PIP=1 | 14:36 |
| mordred | clarkb: I mean, if there were a simple way like that to spin one up for unittests - I'm pretty sure we could kill sqlite with fire | 14:36 |
| yolanda | but seems as tox overwrites it | 14:36 |
| clarkb | mordred: if you show me how to make the db just start at boot without me making custom init changes I would be so happy, but its been such a low priority I have never bothered to figure it out | 14:36 |
| mordred | clarkb: but - I don't have time to think about the full stry yet there | 14:36 |
| mordred | clarkb: yup. same here | 14:36 |
| yolanda | do you know how to mak tox to don't install pip, but use the one from the system? | 14:37 |
| waynr | electrical: fixed 004 | 14:37 |
| electrical | waynr: that's quick :-) | 14:37 |
| clarkb | yolanda: if using tox >= 2.0 it cleans the env prior to testing. You can whitelist the env var if you need to | 14:37 |
| fungi | yolanda: did this start breaking for you with pip 2.0? if so, you probably need a passenv added to your tox.ini | 14:37 |
| clarkb | yolanda: its also possible that tox supports that functionality as a config in tox.ini directly | 14:37 |
| yolanda | fungi, clarkb, we see that virtuaenv installs an old pip version 1.5 | 14:38 |
| yolanda | and we need newer ones | 14:38 |
| yolanda | pip 6 or pip 7 | 14:38 |
| electrical | waynr: diff? would love to try it out | 14:38 |
| *** nelsnelson has quit IRC | 14:38 | |
| mordred | yolanda: that soudns like very old virtualenv | 14:38 |
| fungi | yolanda: virtualenv provides its own bundled pip version | 14:38 |
| clarkb | ya, you can just install a newer virtualenv | 14:38 |
| fungi | pip 1.5 would mean you have virtualenv 11 | 14:38 |
| yolanda | let me check, last time i looked it brought a very old version | 14:39 |
| fungi | if memory serves | 14:39 |
| mordred | yolanda: current virtualenv has current pipi | 14:39 |
| mordred | yolanda: they release together | 14:39 |
| fungi | "very old" in this case means _months_ old | 14:39 |
| yolanda | ok, let me install latest | 14:39 |
| yolanda | and check | 14:39 |
| *** achanda has joined #openstack-infra | 14:40 | |
| *** emagana has joined #openstack-infra | 14:40 | |
| waynr | electrical: sure, one moment | 14:41 |
| yolanda | fungi, i did a pip install --upgrade for virtualenv and it only raises to 1.11.6 | 14:41 |
| yolanda | how do you install them, and which version do you have upstream? | 14:41 |
| *** dangers_away is now known as dangers | 14:41 | |
| mordred | yolanda: mordred@camelot:~$ virtualenv --version | 14:42 |
| mordred | 13.0.0 | 14:42 |
| *** eharney has quit IRC | 14:42 | |
| fungi | yolanda: 13.0.1 is the latest on pypi | 14:42 |
| mordred | yolanda: are you pointing at a stale mirror? | 14:42 |
| fungi | yolanda: are you using an outdated mirror | 14:42 |
| fungi | ? | 14:42 |
| yolanda | looks like | 14:42 |
| yolanda | yep, our mirror | 14:42 |
| *** kgiusti has joined #openstack-infra | 14:43 | |
| waynr | electrical: http://paste.openstack.org/show/240096/ | 14:43 |
| yolanda | mordred, fungi, you rock | 14:43 |
| fungi | yolanda: i take it you don't use bandersnatch to generate your mirror | 14:44 |
| waynr | i left in some debug print statements | 14:44 |
| yolanda | fungi, you are right | 14:44 |
| *** salv-orlando has joined #openstack-infra | 14:45 | |
| *** otter768 has joined #openstack-infra | 14:45 | |
| openstackgerrit | Andreas Jaeger proposed openstack-infra/project-config: Sort PO files to avoid random reorder https://review.openstack.org/186026 | 14:46 |
| clarkb | mordred: https://review.openstack.org/#/c/185824/1 any chance we can get that in so I can reenable puppet on nodepool.o.o | 14:49 |
| *** yfried is now known as yfried|afk | 14:49 | |
| *** otter768 has quit IRC | 14:50 | |
| *** hemnafk is now known as hemna | 14:50 | |
| mordred | clarkb: done | 14:51 |
| *** patrickeast has joined #openstack-infra | 14:51 | |
| *** achanda has quit IRC | 14:52 | |
| *** zz_ja is now known as zz_zz_ja | 14:52 | |
| clarkb | danke | 14:53 |
| *** nelsnelson has joined #openstack-infra | 14:53 | |
| *** fijal has quit IRC | 14:53 | |
| yolanda | fungi, mordred, looking at system-config, it says desired_virtualenv = 12.1.1 | 14:53 |
| yolanda | how are you on 13 ? | 14:54 |
| clarkb | yolanda: we are not | 14:54 |
| *** fijal has joined #openstack-infra | 14:54 | |
| yolanda | 12.1.1 then? | 14:54 |
| electrical | waynr: looks like you solved it indeed :-) | 14:54 |
| clarkb | ya that is what we are using | 14:54 |
| yolanda | mordred, why did you told me 13 ?! is that your local node? | 14:54 |
| clarkb | yolanda: yes camelot is mordred's laptop | 14:54 |
| yolanda | ok | 14:54 |
| mordred | oh - sorry - yeah, that's what I have locally | 14:55 |
| yolanda | i'll stick on what upstream has | 14:55 |
| yolanda | mordred, you are lying to me :) | 14:55 |
| *** Kennan has quit IRC | 14:55 | |
| *** amotoki_ has joined #openstack-infra | 14:55 | |
| *** dims_ has joined #openstack-infra | 14:55 | |
| mordred | yolanda: I do that :) | 14:55 |
| ttx | Alright we are LIVE. Thx everyone who helped making this happen | 14:56 |
| *** Kennan has joined #openstack-infra | 14:56 | |
| yolanda | clarkb, which pip brings 12.1.1? | 14:56 |
| fungi | 12.1.1 was the latest when we updated our puppet to install latest virtualenv. ages ago. like a couple of weeks already | 14:56 |
| mordred | ttx: w00t! | 14:56 |
| clarkb | mordred: fungi: any thoughts on further debugging of swift image uploads? if not I guess next step is a rackspace support ticket | 14:56 |
| mordred | clarkb: I asked in the swift channel | 14:56 |
| mordred | no response yet | 14:56 |
| clarkb | yolanda: I think its latest pip 6.x | 14:56 |
| fungi | yolanda: pip 6.1.1 | 14:56 |
| yolanda | ok | 14:56 |
| yolanda | thx | 14:56 |
| *** pblaho has quit IRC | 14:57 | |
| fungi | yolanda: virtualenv_version - 6 = pip_version | 14:57 |
| waynr | electrical: i am close to solving the others | 14:57 |
| clarkb | mordred: oh cool I can wait for a response there | 14:57 |
| *** pblaho has joined #openstack-infra | 14:57 | |
| electrical | waynr: i wonder where the actual bug is then why it doesn't merge both ways. so confusing :-) | 14:57 |
| fungi | (starting with virtualenv 12 and pip 6 anyway) | 14:57 |
| fungi | yolanda: so virtualenv 13.0.1 vendors a copy of pip 7.0.1 | 14:57 |
| yolanda | fungi, that's more risky | 14:58 |
| yolanda | are you going to update? | 14:58 |
| fungi | yolanda: definitely. just no idea when | 14:58 |
| fungi | clarkb: what is the current error we're getting now that we're using glance v2? | 14:58 |
| clarkb | fungi: likely after we get image builds fixed | 14:58 |
| clarkb | fungi: ClientException: Object POST failed: 504 Gateway Time-out is the edited error message | 14:59 |
| *** maishsk_afk has quit IRC | 14:59 | |
| fungi | 504. thanks | 14:59 |
| waynr | electrical: the bug for 002 and 003? | 14:59 |
| clarkb | if the dib builds for hpcloud successfully upload today then I think we can worry about latest virtualenv | 14:59 |
| *** gokrokve has joined #openstack-infra | 14:59 | |
| electrical | waynr: the bug for 004 | 14:59 |
| waynr | oh | 14:59 |
| *** freerunner has quit IRC | 14:59 | |
| *** amotoki_ has quit IRC | 15:00 | |
| electrical | i do see what you are doing but i wonder where its going wrong. or its something thats not being handled at all anywhere in the code | 15:00 |
| *** claudiub has joined #openstack-infra | 15:00 | |
| fungi | clarkb: are you able to reproduce that error when posting the same file to swift manually? | 15:00 |
| clarkb | fungi: no I haven't attempted a manual upload yet, I will be honest. It is because I have no idea what that entails as its "complicated" | 15:00 |
| *** salv-orlando has quit IRC | 15:00 | |
| clarkb | maybe mordred or greghaynes can tl;dr it for me and I can give it a go | 15:00 |
| *** ddieterly has quit IRC | 15:01 | |
| waynr | well the problem is that there is no way for the "project" or "job-group" to receive parameters set at the job-template level | 15:01 |
| clarkb | fungi: you basically have to do completely undocumented things (because its not upstreamed yet) with swift uploads and glance tasks | 15:01 |
| *** timcline has quit IRC | 15:01 | |
| waynr | it has to happen on a job-template by job-template basis as in the patch i linked you to | 15:01 |
| clarkb | fungi: I could probably figure it out by reading shade code | 15:01 |
| *** armax has joined #openstack-infra | 15:01 | |
| openstackgerrit | Merged openstack-infra/system-config: Use _ instead of - in clouds.yaml https://review.openstack.org/185824 | 15:01 |
| mordred | I have an ansible playbook | 15:01 |
| *** timcline has joined #openstack-infra | 15:02 | |
| mordred | that will do it | 15:02 |
| mordred | although it relies on ansible code that is not yet upstreamed :) | 15:02 |
| waynr | that "params" dict is being used to populate template strings on the merged job-template/job-group/project | 15:02 |
| fungi | clarkb: oh. it's unfortunate that rackspace isn't using openstack. again | 15:02 |
| *** jamespage_ has joined #openstack-infra | 15:02 | |
| waynr | the "params" is the merged set of job-group+project parameters | 15:02 |
| mordred | oh - do you mean testing doing it without shade? I do have shell scripts you can use to do that | 15:02 |
| clarkb | mordred: yes, we should probably try it without shade entirely | 15:03 |
| clarkb | but as I said thats "complicated" | 15:03 |
| fungi | mordred: just wondering if we have distilled a reproducible use case that doesn't start out "install and run this code we wrote" | 15:03 |
| mordred | clarkb: https://review.openstack.org/#/c/140840/13/tools/HOW-TO-TEST.rst | 15:03 |
| mordred | if you grab 140840 you'll get some new scripts in the tools dir | 15:03 |
| mordred | they can be used to upload to rackspace using just normal command line clients | 15:04 |
| mordred | the instructions on how are in that file | 15:04 |
| mordred | the swift command line for just doing the swift upload is in tools/upload-rax.sh | 15:04 |
| mordred | I think just trying that first is likely a good first test | 15:04 |
| *** asselin_ has joined #openstack-infra | 15:04 | |
| *** Ala has quit IRC | 15:04 | |
| mordred | swift upload --object-name ${IMAGE}.vhd images ${IMAGE}.vhd | 15:05 |
| electrical | waynr: feels like there is a bug then in params = self.applyDefaults(params, template) handling. template should be the override dict in the applyDefaults function which does the same as you done. | 15:06 |
| clarkb | OH! I think we are trying to upload qcows still :) | 15:06 |
| clarkb | der | 15:06 |
| mordred | really? that's no fun | 15:06 |
| mordred | but it was a swift gateway timeout that we got, no? | 15:06 |
| clarkb | I mean it doesn't matter because swift fails | 15:06 |
| mordred | yah | 15:06 |
| fungi | clarkb: aha. missing the convert step? or just passing the wrong filename to upload? | 15:06 |
| clarkb | fungi: I think both | 15:07 |
| mordred | you want I should try the swift upload? | 15:07 |
| *** dprince has quit IRC | 15:07 | |
| clarkb | mordred: ya if you want to at least see if swift upload can be made to work with the qcow2 that would be good | 15:07 |
| electrical | waynr: only difference is that you do a if key not in params: and applyDefaults does a if key in params: | 15:07 |
| *** maishsk_afk has joined #openstack-infra | 15:08 | |
| openstackgerrit | Christian Berendt proposed openstack-infra/irc-meetings: Whitelist mkdir as external in the tox.ini file https://review.openstack.org/186032 | 15:08 |
| *** ddieterly has joined #openstack-infra | 15:08 | |
| *** freerunner has joined #openstack-infra | 15:09 | |
| mordred | clarkb: trying now | 15:09 |
| clarkb | I will write a change for the file format issue | 15:09 |
| electrical | waynr: seems i was wrong. broke more things then i fixed with it :p | 15:09 |
| katyafervent | Hi everyone! please help us to make stable/juno working https://review.openstack.org/#/c/184809/ | 15:09 |
| *** bswartz has quit IRC | 15:10 | |
| electrical | waynr: i would have assumed that https://git.openstack.org/cgit/openstack-infra/jenkins-job-builder/tree/jenkins_jobs/parser.py#n137 would take care of the part you just fixed | 15:11 |
| electrical | very confusing. | 15:11 |
| waynr | electrical: agreed, very confusing | 15:12 |
| *** bswartz has joined #openstack-infra | 15:12 | |
| waynr | but that code block addresses a different issue related to parameter interpolation | 15:12 |
| openstackgerrit | Clark Boylan proposed openstack-infra/system-config: Set image-type to vhd for rax nodepool providers https://review.openstack.org/186034 | 15:13 |
| electrical | waynr: yeah indeed. | 15:13 |
| *** maishsk_afk_ has joined #openstack-infra | 15:13 | |
| *** e0ne is now known as e0ne_ | 15:13 | |
| clarkb | mordred: ^ I think we may also need to be telling rax to trea these at pvhvm somehow, but that can happen after we have working images | 15:13 |
| *** maishsk_afk has quit IRC | 15:13 | |
| *** maishsk_afk_ is now known as maishsk_afk | 15:13 | |
| electrical | waynr: in any case a good rewrite is mandatory for the project to move forward :-) | 15:14 |
| waynr | agreed! | 15:15 |
| mordred | clarkb: yes, that's right - we need the properties to be set | 15:15 |
| mordred | clarkb: I think that support is already added to nodepool? | 15:15 |
| waynr | i wonder how we should go about socializing the need for a moratorium on new JJB yamlparser features | 15:15 |
| clarkb | mordred: ya I think we can set it on the image, but would have to go double check | 15:15 |
| clarkb | mordred: its definitely less urgent than make swift go | 15:15 |
| clarkb | no pun intended | 15:15 |
| mordred | clarkb: ok. so - uploading via swift command line client worked | 15:16 |
| mordred | clarkb: so there is a different problem | 15:16 |
| mordred | clarkb: I'll start digging in | 15:16 |
| electrical | waynr: a large banner in the top of the file and be done? ;-) | 15:16 |
| waynr | yeah, i guess i'll keep commenting on new features indicating that we should wait for a rewrite | 15:17 |
| waynr | (i personally don't ready comments at the top of files very often) | 15:17 |
| *** hashar has quit IRC | 15:17 | |
| electrical | waynr: hehe very true :-) | 15:17 |
| electrical | at least the testing framework is easy to use, that we don't have to change i hope :p | 15:18 |
| openstackgerrit | James E. Blair proposed openstack-infra/project-config: Add project-team-guide https://review.openstack.org/185746 | 15:18 |
| waynr | yeah | 15:19 |
| *** spzala has joined #openstack-infra | 15:19 | |
| *** _nadya_ has quit IRC | 15:19 | |
| electrical | waynr: with 004 solved at least i won't have to duplicate entries from other name spaces into global, so that will be much cleaner | 15:20 |
| *** rlandy has quit IRC | 15:20 | |
| *** sslypushenko has joined #openstack-infra | 15:20 | |
| jeblair | mordred, SergeyLukjanov: if you have a minute to leave a vote on https://review.openstack.org/182811 that would be great | 15:21 |
| *** panda has quit IRC | 15:21 | |
| *** u_glide has left #openstack-infra | 15:21 | |
| *** cody-somerville has quit IRC | 15:21 | |
| *** acruz_ has quit IRC | 15:22 | |
| *** yfried|afk is now known as yfried | 15:22 | |
| jeblair | clarkb: revised gertty migration work okay? | 15:22 |
| mordred | jeblair: done | 15:22 |
| clarkb | jeblair: yup, appears to be up and happy | 15:22 |
| *** e0ne_ is now known as e0ne | 15:22 | |
| *** panda has joined #openstack-infra | 15:22 | |
| openstackgerrit | Yanis Guenane proposed openstack-infra/tripleo-ci: Pin puppetlabs-mysql https://review.openstack.org/186016 | 15:22 |
| *** koolhead17 has quit IRC | 15:22 | |
| *** claudiub has quit IRC | 15:23 | |
| liuxg | hi, my CI failed with test case " test_minimum_basic_scenario" for establish authenticated ssh connection to server failed | 15:23 |
| *** dtantsur is now known as dtantsur|afk | 15:23 | |
| liuxg | who can give me a help | 15:23 |
| *** koolhead17 has joined #openstack-infra | 15:24 | |
| openstackgerrit | Merged openstack-infra/infra-specs: Add priority efforts section https://review.openstack.org/183337 | 15:24 |
| *** nelsnelson has quit IRC | 15:24 | |
| waynr | electrical: hmm looking more closely at 002, i don't think this is an example we should support before the yamlparser rewrite | 15:24 |
| *** rbrooker_ has joined #openstack-infra | 15:25 | |
| waynr | that is to say, we shouldn't expect non-global defaults to merge into global defaults | 15:25 |
| openstackgerrit | Davanum Srinivas (dims) proposed openstack-infra/irc-meetings: Switch oslo chair https://review.openstack.org/186038 | 15:25 |
| asselin_ | liuxg, you can join us in -meeting4 | 15:25 |
| *** dprince has joined #openstack-infra | 15:25 | |
| waynr | the way defaults are designed is as an "either-or" arrangement | 15:25 |
| waynr | either you use global defaults, or you use an alternate defaults | 15:25 |
| liuxg | asselin: thanks, I will join the meeting | 15:26 |
| waynr | although in general i think in the future JJB should support something like 002/005 and 003/006 | 15:26 |
| waynr | so you were correct earlier ;) | 15:27 |
| *** gokrokve has quit IRC | 15:27 | |
| waynr | (correct that we leave the fix at targeting 004 for now) | 15:27 |
| *** gokrokve has joined #openstack-infra | 15:28 | |
| *** bhunter71 has joined #openstack-infra | 15:28 | |
| *** devvesa has quit IRC | 15:28 | |
| waynr | electrical: you want to update the patch to remove 002,003,005, and 006 and add the fix for 004? | 15:28 |
| *** maishsk_afk has quit IRC | 15:28 | |
| clarkb | mordred: is this possibly related to how that account has a different project_name for swift than it does for nova? | 15:29 |
| *** koolhead17 has quit IRC | 15:29 | |
| mordred | clarkb: it shoudl not - I'm using the same keystone creds in each case | 15:29 |
| clarkb | mordred: the url in the errors seem to have the correct mosso files path so I don'y think that is it | 15:29 |
| mordred | clarkb: I've made a small python script to do just the upload using shade and those creds since the swift command line worked | 15:30 |
| *** maishsk_afk has joined #openstack-infra | 15:30 | |
| mordred | clarkb: and I'm running that with debug on | 15:30 |
| mordred | we'll see what, if anything, that shows | 15:30 |
| *** marun has joined #openstack-infra | 15:31 | |
| *** eharney has joined #openstack-infra | 15:32 | |
| openstackgerrit | James E. Blair proposed stackforge/gertty: Fix searching for message https://review.openstack.org/172667 | 15:32 |
| openstackgerrit | James E. Blair proposed stackforge/gertty: Advance cursor on change list toggle https://review.openstack.org/185698 | 15:32 |
| openstackgerrit | James E. Blair proposed stackforge/gertty: Allow reviewing one change in change list https://review.openstack.org/185697 | 15:32 |
| openstackgerrit | James E. Blair proposed stackforge/gertty: Fix updating flags on threaded changes https://review.openstack.org/185696 | 15:32 |
| openstackgerrit | James E. Blair proposed stackforge/gertty: Make "limit" a noop in queries https://review.openstack.org/172670 | 15:32 |
| openstackgerrit | James E. Blair proposed stackforge/gertty: Fix searching by reviewer account id https://review.openstack.org/172669 | 15:32 |
| openstackgerrit | James E. Blair proposed stackforge/gertty: Batch sync change by commit tasks https://review.openstack.org/172691 | 15:32 |
| openstackgerrit | James E. Blair proposed stackforge/gertty: Fix age searching https://review.openstack.org/172668 | 15:32 |
| openstackgerrit | James E. Blair proposed stackforge/gertty: Fix refresh on project and change lists https://review.openstack.org/183887 | 15:32 |
| openstackgerrit | James E. Blair proposed stackforge/gertty: Fix diff display of deleted empty files https://review.openstack.org/180546 | 15:32 |
| openstackgerrit | James E. Blair proposed stackforge/gertty: Add database pruning https://review.openstack.org/179981 | 15:32 |
| openstackgerrit | James E. Blair proposed stackforge/gertty: Attach comments to files https://review.openstack.org/179980 | 15:32 |
| openstackgerrit | James E. Blair proposed stackforge/gertty: Add files table https://review.openstack.org/179979 | 15:32 |
| openstackgerrit | James E. Blair proposed stackforge/gertty: Add ability to review multiple changes at once https://review.openstack.org/185695 | 15:32 |
| *** gokrokve has quit IRC | 15:32 | |
| SergeyLukjanov | jeblair, looking | 15:32 |
| mordred | it's the jeblair patch bomb! | 15:32 |
| *** liuxg has left #openstack-infra | 15:32 | |
| jeblair | it's going to be followed by a merge bomb :) | 15:33 |
| zigo | StevenK: May I help to get the packages that you need into Debian? | 15:33 |
| *** mrmartin has joined #openstack-infra | 15:33 | |
| zigo | Zuul maybe? | 15:33 |
| zigo | And nodepool? | 15:33 |
| clarkb | I think hashar has both of those in debian already | 15:33 |
| zigo | Oh ok. | 15:33 |
| zigo | clarkb: https://packages.debian.org/search?keywords=zuul&searchon=names&suite=all§ion=all | 15:34 |
| zigo | No zuul in Debian... | 15:34 |
| zigo | No nodepool either. | 15:34 |
| clarkb | huh, I was pretty sure hashar had done packaging work for both | 15:34 |
| *** BharatK has joined #openstack-infra | 15:34 | |
| openstackgerrit | Merged stackforge/gertty: Add support for SQLAlchemy 1.0.4 https://review.openstack.org/185761 | 15:34 |
| zigo | clarkb: Problably, but it was never uploaded. | 15:35 |
| jeblair | maybe just wikimedia-local packaging? | 15:35 |
| fungi | zigo: also pretty sure StevenK is a dd and can upload on his own, though not sure if you were offering packaging help | 15:35 |
| zigo | fungi: I can help anyone who ask for such things. | 15:35 |
| mordred | we're popular | 15:35 |
| rcarrillocruz | asselin: hiya! | 15:36 |
| rcarrillocruz | you around? | 15:36 |
| fungi | zigo: ahh, yep, he's a.k.a. stevenk@debian.org | 15:36 |
| zigo | fungi: I just found out on db.debian.org! :) | 15:36 |
| *** claudiub has joined #openstack-infra | 15:37 | |
| zigo | I wonder why zuul and nodepool were never uploaded then. | 15:37 |
| asselin_ | rcarrillocruz, yes....in 2 meetings... | 15:37 |
| asselin_ | rcarrillocruz, hi :) | 15:37 |
| rcarrillocruz | duh, sorry about that :/ | 15:37 |
| *** ParsectiX has joined #openstack-infra | 15:37 | |
| rcarrillocruz | so, we are prepping our Gerrit upgrade | 15:37 |
| rcarrillocruz | and turns out our internal manifest won't work on Trusty systems | 15:37 |
| rcarrillocruz | we really want to do upgrade to 2.8 on Trusty | 15:38 |
| SergeyLukjanov | jeblair, +2 on the new team structure - very clear and reasonable! thx | 15:38 |
| rcarrillocruz | i'll take a stab at puppet-gerrit module | 15:38 |
| rcarrillocruz | do I need to add myself to the summit etherpad? | 15:38 |
| *** pradk has quit IRC | 15:38 | |
| rcarrillocruz | or is it put on read-only ? | 15:38 |
| openstackgerrit | Christian Berendt proposed openstack-infra/irc-meetings: Use valid HTML5 syntax in the meetingindex.jinja template https://review.openstack.org/186044 | 15:38 |
| clarkb | rcarrillocruz: fwiw we did 2.4 -> 2.8 on precise, then upgraded to trusty | 15:38 |
| clarkb | rcarrillocruz: so that is known to work | 15:38 |
| jeblair | SergeyLukjanov: cool, thanks! | 15:38 |
| *** koolhead17 has joined #openstack-infra | 15:39 | |
| *** harlowja_at_home has joined #openstack-infra | 15:40 | |
| rbradfor | I have documented my limited understanding of the Fedora image used by Magnum at https://blueprints.launchpad.net/magnum/+spec/cache-required-fedora-image I have requested input from the Magnum team and would value any Infra knowledge that can help achieve the objective | 15:40 |
| rcarrillocruz | clarkb: did you do an upgrade on the same instance? or you did a spin of a new instance, import of mysqldump and gerrit repos and then did the upgrade on new instance? | 15:40 |
| clarkb | rcarrillocruz: we used a new instance that we switched dns to point to | 15:40 |
| *** tim_o has joined #openstack-infra | 15:41 | |
| rcarrillocruz | so you did the mysql export/import then... | 15:41 |
| *** gokrokve has joined #openstack-infra | 15:42 | |
| jeblair | rbradfor: ack, i'll try to take a look at it in a bit | 15:42 |
| mordred | clarkb: found the swift issue and have learned somethign about swift | 15:43 |
| mordred | clarkb: we were actually failing in the post-upload metadata update | 15:43 |
| openstackgerrit | Matthew Treinish proposed openstack-infra/subunit2sql: Cleanup std dev backfill in run_time graph https://review.openstack.org/186046 | 15:43 |
| mordred | clarkb: turns out that's much more expensive than one might expect | 15:43 |
| clarkb | fungi: devstack-trusty image build was about 30 minutes today | 15:43 |
| mordred | clarkb: so I will now work up a patch to set the metadata at object creation time | 15:43 |
| clarkb | mordred: oh? due to hashing? | 15:43 |
| *** maishsk_afk has quit IRC | 15:44 | |
| fungi | clarkb: and was an hour yesterday? | 15:44 |
| notmyname | in many swift deployments (including rackspace), updating metadata is a whole object copy operation (server-side). therefore it's expensive | 15:44 |
| clarkb | fungi: ya working to confirm exact numbers | 15:44 |
| clarkb | fungi: its all logged, just have to get the correct log | 15:45 |
| *** yfried has quit IRC | 15:45 | |
| fungi | clarkb: so cocktail napkin analysis says that sata is 25% as fast as ephemeral and ssd is 50% as fast as ephemeral? | 15:45 |
| *** nelsnelson has joined #openstack-infra | 15:46 | |
| clarkb | fungi: old build was ~49 minutes, new build ~32 minutes | 15:46 |
| clarkb | and ya ephemeral was ~15 | 15:46 |
| *** wznoinsk_ has joined #openstack-infra | 15:47 | |
| openstackgerrit | James E. Blair proposed stackforge/gertty: Advance cursor on change list toggle https://review.openstack.org/185698 | 15:47 |
| openstackgerrit | James E. Blair proposed stackforge/gertty: Allow reviewing one change in change list https://review.openstack.org/185697 | 15:47 |
| openstackgerrit | James E. Blair proposed stackforge/gertty: Fix updating flags on threaded changes https://review.openstack.org/185696 | 15:47 |
| openstackgerrit | James E. Blair proposed stackforge/gertty: Add ability to review multiple changes at once https://review.openstack.org/185695 | 15:47 |
| *** annegentle has quit IRC | 15:47 | |
| jeblair | found a bug in the mass-review feature | 15:48 |
| *** wznoinsk_ is now known as wznoinsk | 15:48 | |
| *** Guest59858 has quit IRC | 15:48 | |
| openstackgerrit | Monty Taylor proposed openstack-infra/shade: Set metadata headers on object create https://review.openstack.org/186048 | 15:49 |
| pabelanger | So, I'm not part of the openstack-infra-packagers but need to get some patches merged. Could a core please approve the following? https://review.openstack.org/#/c/59590/ | 15:49 |
| pabelanger | same with https://review.openstack.org/#/c/59591/ | 15:49 |
| electrical | waynr: yeah will do. thanks | 15:49 |
| pabelanger | this is for the zuul-packaging repo | 15:49 |
| *** Sukhdev has joined #openstack-infra | 15:50 | |
| *** ihrachyshka has quit IRC | 15:51 | |
| openstackgerrit | Merged stackforge/gertty: Support regexes in search https://review.openstack.org/172649 | 15:51 |
| openstackgerrit | Merged stackforge/gertty: Do not enqueue duplicate tasks https://review.openstack.org/172650 | 15:51 |
| openstackgerrit | Merged stackforge/gertty: Fix searching for message https://review.openstack.org/172667 | 15:51 |
| openstackgerrit | Merged stackforge/gertty: Fix age searching https://review.openstack.org/172668 | 15:51 |
| openstackgerrit | Merged stackforge/gertty: Fix searching by reviewer account id https://review.openstack.org/172669 | 15:51 |
| openstackgerrit | Merged stackforge/gertty: Make "limit" a noop in queries https://review.openstack.org/172670 | 15:51 |
| clarkb | fungi: its a decent improvement that we can likely live with particularly as we trim down the number of images we host | 15:51 |
| *** ParsectiX has quit IRC | 15:51 | |
| openstackgerrit | Merged stackforge/gertty: Batch sync change by commit tasks https://review.openstack.org/172691 | 15:51 |
| openstackgerrit | Merged stackforge/gertty: Add files table https://review.openstack.org/179979 | 15:51 |
| openstackgerrit | Merged stackforge/gertty: Attach comments to files https://review.openstack.org/179980 | 15:52 |
| openstackgerrit | Merged stackforge/gertty: Add database pruning https://review.openstack.org/179981 | 15:52 |
| openstackgerrit | Merged stackforge/gertty: Fix diff display of deleted empty files https://review.openstack.org/180546 | 15:52 |
| openstackgerrit | Merged stackforge/gertty: Fix refresh on project and change lists https://review.openstack.org/183887 | 15:52 |
| openstackgerrit | Merged stackforge/gertty: Add ability to review multiple changes at once https://review.openstack.org/185695 | 15:52 |
| openstackgerrit | Merged stackforge/gertty: Fix updating flags on threaded changes https://review.openstack.org/185696 | 15:52 |
| openstackgerrit | Merged stackforge/gertty: Allow reviewing one change in change list https://review.openstack.org/185697 | 15:52 |
| openstackgerrit | Merged stackforge/gertty: Advance cursor on change list toggle https://review.openstack.org/185698 | 15:52 |
| jeblair | fungi, crinkle: that was mass-reviewed ^ | 15:53 |
| *** lifeless_ is now known as lifeless | 15:53 | |
| *** jamespage_ has quit IRC | 15:54 | |
| crinkle | jeblair: :D | 15:54 |
| *** jamespage_ has joined #openstack-infra | 15:54 | |
| fungi | heh | 15:54 |
| openstackgerrit | Monty Taylor proposed openstack-infra/shade: Set metadata headers on object create https://review.openstack.org/186048 | 15:55 |
| fungi | clarkb: we have a few diskimages we're no longer using still in the config, right? maybe we should go ahead and cull those now that we're moving actual production labels to dib anyway? | 15:55 |
| openstackgerrit | Richard Pijnenburg proposed openstack-infra/jenkins-job-builder: Adding tests surrounding handling of defaults values https://review.openstack.org/185661 | 15:55 |
| mordred | clarkb, fungi, notmyname: ^^ I have tested that that can upload one of our images | 15:55 |
| *** amotoki_ has joined #openstack-infra | 15:56 | |
| *** arxcruz has joined #openstack-infra | 15:56 | |
| *** arxcruz has quit IRC | 15:56 | |
| electrical | waynr: thank you for fixing the bug :-) | 15:56 |
| *** dannywilson has joined #openstack-infra | 15:56 | |
| clarkb | fungi: its possible, we may want to sync up with ianw as those are largely images he has in progress | 15:56 |
| clarkb | though I think devstack keeps breaking on fedora and centos for unrelated reasons making it more complicated | 15:56 |
| electrical | clarkb: fungi mordred https://review.openstack.org/#/c/185661/ interesting fix. would love your feedback. | 15:56 |
| *** fhubik is now known as fhubik_afk | 15:57 | |
| fungi | clarkb: oh, i meant some of the duplicate "foo-dib" labels | 15:57 |
| mordred | clarkb, fungi, notmyname: for reference, this is the test script I used: http://paste.openstack.org/show/240206/ | 15:57 |
| clarkb | fungi: ya I think I got rid of the true duplicates, the others are the ones that ianw was working on and aren't really duplicates but I maybe missing something | 15:57 |
| *** tiswanso has quit IRC | 15:58 | |
| mordred | I did the md5/sha256 params to skip the calculation of them - they are not required | 15:58 |
| mordred | (pasting that both to show what I verified, and the simple case of how to do this task from python) | 15:58 |
| mordred | Shrews: ^^ | 15:59 |
| *** jamespage_ has quit IRC | 15:59 | |
| *** deepakcs has joined #openstack-infra | 16:00 | |
| *** baoli has quit IRC | 16:00 | |
| *** emagana has quit IRC | 16:00 | |
| *** julim has quit IRC | 16:00 | |
| *** yamahata has quit IRC | 16:00 | |
| *** zz_zz_ja is now known as zz_zz_zz_ja | 16:01 | |
| *** ildikov has quit IRC | 16:01 | |
| *** amotoki_ has quit IRC | 16:01 | |
| *** yamahata has joined #openstack-infra | 16:01 | |
| openstackgerrit | Merged stackforge/gertty: Be more verbose on non-tagged versions https://review.openstack.org/179776 | 16:01 |
| openstackgerrit | Merged stackforge/gertty: Make ButtonDialog scrollable https://review.openstack.org/183809 | 16:01 |
| openstackgerrit | Merged stackforge/gertty: Explain how to install on Arch Linux https://review.openstack.org/175008 | 16:01 |
| *** baoli_ has joined #openstack-infra | 16:02 | |
| *** emagana has joined #openstack-infra | 16:02 | |
| mtreinish | ^^^ \o/ | 16:02 |
| *** tiswanso has joined #openstack-infra | 16:02 | |
| jeblair | mordred: what version of pbr should gertty be using? | 16:03 |
| openstackgerrit | Matt Riedemann proposed openstack-infra/elastic-recheck: Add query for zake 0.2.2 stable break bug 1459322 https://review.openstack.org/186055 | 16:03 |
| clarkb | jeblair: >=1.0,<2.0 iirc | 16:03 |
| openstack | bug 1459322 in tooz "stable/juno broken with zake 0.2.2 release" [Undecided,New] https://launchpad.net/bugs/1459322 | 16:03 |
| clarkb | jeblair: global requirments should say | 16:03 |
| *** julim has joined #openstack-infra | 16:03 | |
| *** dustins_away is now known as dustins | 16:03 | |
| jeblair | clarkb: was 1.0 the one that fixed the alembic thing? | 16:04 |
| clarkb | jeblair: 1.0 is the pbr release that added the semver enforcement | 16:04 |
| jeblair | global says pbr>=0.11,<2.0 | 16:04 |
| jeblair | so maybe it was 0.11 that fixed alembic? | 16:04 |
| * jeblair digs | 16:04 | |
| mordred | yes | 16:05 |
| mtreinish | jeblair: I just switched subunit2sql to >=1.0.0 but that's because I started using extras | 16:05 |
| mordred | jeblair: the fix for alembic is in 0.11 | 16:05 |
| *** ashleighfarnham has joined #openstack-infra | 16:05 | |
| jeblair | mordred: i agree, found "Don't avoid egg-info quite so often | 16:06 |
| jeblair | in that tags history | 16:06 |
| jeblair | so yeah pbr>=0.11,<2.0 seems reasonable | 16:06 |
| *** oomichi has quit IRC | 16:06 | |
| *** jistr has quit IRC | 16:07 | |
| lxsli | https://review.openstack.org/185620 https://review.openstack.org/185632 | 16:07 |
| lxsli | mtreinish: ^-- couple of elastic recheck queries | 16:07 |
| openstackgerrit | James E. Blair proposed stackforge/gertty: Update PBR requirement to >=0.11 https://review.openstack.org/186057 | 16:07 |
| jeblair | mordred: ^ | 16:07 |
| *** amitgandhinz has quit IRC | 16:08 | |
| clarkb | puppet is reenabled on nodepool.o.o | 16:08 |
| mtreinish | lxsli: nice, 98k hits on that second one | 16:08 |
| openstackgerrit | Merged openstack-infra/elastic-recheck: Add query for zake 0.2.2 stable break bug 1459322 https://review.openstack.org/186055 | 16:09 |
| openstack | bug 1459322 in tooz "stable/juno broken with zake 0.2.2 release" [Undecided,New] https://launchpad.net/bugs/1459322 | 16:09 |
| mordred | jeblair: ++ | 16:09 |
| openstackgerrit | Merged stackforge/gertty: Update PBR requirement to >=0.11 https://review.openstack.org/186057 | 16:09 |
| *** rcarrillocruz has left #openstack-infra | 16:09 | |
| pleia2 | good morning | 16:09 |
| *** mpavone has quit IRC | 16:10 | |
| *** jlanoux has quit IRC | 16:10 | |
| openstackgerrit | Travis Tripp proposed openstack-infra/irc-meetings: Add Search Team Meeting https://review.openstack.org/186060 | 16:10 |
| anteaya | morning pleia2 | 16:10 |
| *** asselin__ has joined #openstack-infra | 16:11 | |
| *** stevemar has joined #openstack-infra | 16:11 | |
| *** amitgandhinz has joined #openstack-infra | 16:11 | |
| mtreinish | ttx, devananda: it looks like python-ironicclient's lp isn't a member of the openstack group | 16:12 |
| *** asselin_ has quit IRC | 16:12 | |
| SpamapS | morganfainberg: when you have like, 20 minutes, we should actually discuss why the generic dogpile.cache key generator can't be made to work as well as the one we have in shade. | 16:12 |
| *** harlowja_at_home has quit IRC | 16:12 | |
| mtreinish | or at least e-r tests are complaining about it: http://logs.openstack.org/20/185620/2/check/gate-elastic-recheck-tox-queries/e8d8b3c/console.html#_2015-05-26_14_40_17_064 | 16:12 |
| mtreinish | lxsli: ^^^ | 16:12 |
| *** unicell1 has quit IRC | 16:12 | |
| morganfainberg | SpamapS: let me write some | 16:12 |
| morganfainberg | Emails and I have time. | 16:12 |
| morganfainberg | SpamapS: :) | 16:12 |
| ttx | mtreinish: will fix it if this page ever loads | 16:13 |
| pabelanger | ttx, With eavesdrop.o.o do you see adding the ability to have individual iCals for teams? On top of the single iCal for all teams? | 16:13 |
| morganfainberg | SpamapS: maybe pre lunch but post breakfast? | 16:13 |
| ttx | pabelanger: I see it. tonyb suggested added that capability to yaml2ical | 16:13 |
| *** freerunner has quit IRC | 16:13 | |
| mordred | jeblair: if you have a sec: https://review.openstack.org/#/c/186048/ is needed for dib-nodepool to get us past the current sad | 16:13 |
| *** jogo has quit IRC | 16:13 | |
| lxsli | mtreinish: nice spot | 16:13 |
| *** freerunner has joined #openstack-infra | 16:13 | |
| clarkb | jeblair: mordred https://review.openstack.org/#/c/186034/ too | 16:14 |
| pabelanger | ttx, Nice. Good to know | 16:14 |
| fungi | pabelanger: if you get bored, you could just submit a change to generate them, and another to link them in the template | 16:15 |
| pabelanger | fungi, so, patches welcome? ;) | 16:15 |
| clarkb | mordred: also your nodepool change is still failing pep8 due to the undefined name | 16:15 |
| *** alexsyip has joined #openstack-infra | 16:16 | |
| mtreinish | ttx: thanks | 16:16 |
| fungi | pabelanger: that is an excellent distillation of what i said, yep! ;) | 16:16 |
| pabelanger | hehe | 16:16 |
| *** abregman has quit IRC | 16:16 | |
| openstackgerrit | Matt Riedemann proposed openstack/requirements: Raise cap on tooz to <0.13 https://review.openstack.org/186065 | 16:16 |
| *** hdd has joined #openstack-infra | 16:16 | |
| *** nikeshm has joined #openstack-infra | 16:16 | |
| ttx | mtreinish: should be good now | 16:17 |
| *** jogo has joined #openstack-infra | 16:17 | |
| *** achanda has joined #openstack-infra | 16:17 | |
| *** rbrooker_ has quit IRC | 16:17 | |
| mordred | clarkb: yah - I'm looking in to that now. there's something else broken after I fixed that | 16:18 |
| pleia2 | ooh bug 1226855 is getting some attention again, hooray for the rise of containers? :) | 16:18 |
| openstack | bug 1226855 in lxc (Ubuntu) "Cannot use open-iscsi inside LXC container" [Wishlist,Confirmed] https://launchpad.net/bugs/1226855 | 16:18 |
| openstackgerrit | James E. Blair proposed openstack-infra/irc-meetings: Remove Install Guide Update Meeting https://review.openstack.org/186066 | 16:18 |
| pleia2 | seems some folks at canonical want to raise priority, so fingers crossed that serge can make time for it now | 16:18 |
| *** fhubik_afk is now known as fhubik | 16:18 | |
| *** gokrokve has quit IRC | 16:19 | |
| SpamapS | morganfainberg: just when you have some time. I think I understand, but I suspect you have it in a much more succinct package than my ramblings. | 16:19 |
| openstackgerrit | James E. Blair proposed openstack-infra/irc-meetings: Remove Networking Guide Update Meeting https://review.openstack.org/186067 | 16:19 |
| openstackgerrit | Sean M. Collins proposed openstack-infra/irc-meetings: Remove the IPv6 subteam meeting https://review.openstack.org/186068 | 16:19 |
| *** gokrokve has joined #openstack-infra | 16:19 | |
| morganfainberg | SpamapS: sure! I think I know what the issue(s) are will ping you post breakfast. | 16:19 |
| openstackgerrit | Merged openstack-infra/elastic-recheck: gate-heat-python27 #1458399 https://review.openstack.org/185632 | 16:19 |
| nikeshm | hi,i am tring to run dsvm-tempest-ful job in jenkins slaves,but its failing from yesterday night | 16:20 |
| nikeshm | http://paste.openstack.org/show/240214/ | 16:20 |
| *** vahidh has joined #openstack-infra | 16:20 | |
| openstackgerrit | Merged openstack-infra/system-config: Update team structure and add council https://review.openstack.org/182811 | 16:21 |
| *** gokrokve has quit IRC | 16:21 | |
| *** derekh_ has joined #openstack-infra | 16:21 | |
| nikeshm | asselin: hi | 16:21 |
| asselin__ | hi | 16:21 |
| asselin__ | nikeshm, look at your devstack.log file | 16:22 |
| clarkb | pleia2 I saw that, may make testing nested openstacks much simpler in the near future | 16:22 |
| pleia2 | clarkb: yep | 16:23 |
| mordred | clarkb: so - I fixed a couple of things and now am hitting a unit test fail that I think I need help with | 16:23 |
| asselin__ | line 226 of your post says to look there | 16:23 |
| openstackgerrit | Monty Taylor proposed openstack-infra/nodepool: Consume os-client-config directly https://review.openstack.org/185972 | 16:23 |
| asselin__ | nikeshm, logs in logs/devstacklog.txt.gz | 16:23 |
| openstackgerrit | Merged openstack-infra/irc-meetings: Remove Install Guide Update Meeting https://review.openstack.org/186066 | 16:23 |
| clarkb | mordred, ok I have the tea kettle on but can dig in once my tea is madr | 16:23 |
| mordred | clarkb: I'm still looking- but a second set of eyes would now be helpful | 16:23 |
| openstackgerrit | Merged openstack-infra/irc-meetings: Remove Networking Guide Update Meeting https://review.openstack.org/186067 | 16:23 |
| mordred | clarkb: awesome | 16:23 |
| *** ashleighfarnham has quit IRC | 16:23 | |
| *** claudiub has quit IRC | 16:24 | |
| mtreinish | ttx: cool, thanks | 16:24 |
| *** derekh has quit IRC | 16:25 | |
| openstackgerrit | Merged openstack-infra/elastic-recheck: gate-python-ironicclient-python34 #1454051 https://review.openstack.org/185620 | 16:26 |
| lxsli | mtreinish: thanks :) | 16:27 |
| akrivoka | anyone have any idea how to fix the jenkins grenade failure all stable/kilo patches are getting? e.g. https://review.openstack.org/#/c/183656/ | 16:28 |
| zaro | morning | 16:28 |
| Daviey | akrivoka: it is in progress | 16:28 |
| openstackgerrit | Anita Kuno proposed openstack-infra/irc-meetings: Correct an error in a meeting description. https://review.openstack.org/186075 | 16:29 |
| anteaya | morning zaro | 16:29 |
| Daviey | akrivoka: details on -dev ML and https://etherpad.openstack.org/p/stable-tracker , being discussed in #openstack-stable | 16:29 |
| akrivoka | Daviey: thanks for the info! | 16:29 |
| *** fawadkhaliq has quit IRC | 16:30 | |
| *** julim has quit IRC | 16:31 | |
| nikeshm | asselin_ : oh i was not getting this file in my logs directory http://115.119.152.41/dothill-logs/logs/$LOG_PATH/logs/ ,BTW,just curious to know if anything break in upstream,becuase i was able to run this earlier but from yesterday night its not working | 16:32 |
| *** dannywilson has quit IRC | 16:32 | |
| asselin__ | nikeshm, not sure why you have $LOG_PATH in your url...something not right. | 16:36 |
| asselin__ | nikeshm, my system's running..you'll need that log file to see why your ci is failing | 16:36 |
| pabelanger | sigh, I posted this into the wrong channel. So I apologizes for the duplication if you are in the other channel. | 16:36 |
| pabelanger | We talked a little about this yesterday, revolving around having nodepool create build environments. I was envisioning a new prepare_node_pkg.sh script (calling slave.pp and slave_pkg.pp), which would be its own package image, to allow jenkins to run the tooling. This was simple to remove some of the dependencies that prepare_node_base.sh adds. However, I'm not apposed to using prepare_node_bare.sh and adding into it build | 16:36 |
| pabelanger | dependencies, however seems to add more overhead into that image. | 16:36 |
| *** maishsk_afk has joined #openstack-infra | 16:37 | |
| anteaya | ttx: we seem to have lost the ability to link from the meeting wikipage to individual meeting wikipages with agendas | 16:37 |
| mordred | pabelanger: yeah - I think we want to avoid having multiple images - and instead just have an ubuntu-trusty image and a fedora-22 image, etc.... all of those images should have nothing installed but a maximal set of things pre-cached | 16:38 |
| openstackgerrit | Richard Pijnenburg proposed openstack-infra/jenkins-job-builder: Adding tests surrounding handling of defaults values https://review.openstack.org/185661 | 16:38 |
| clarkb | mordred: it looks like the provider manager thread is dying for some reason | 16:38 |
| clarkb | mordred: unfortunately it doesn't seem to haev logged the reason? | 16:38 |
| mordred | clarkb: yah. that's how it's looking to me too | 16:39 |
| electrical | zaro: uploaded new patch set with hopefully a better explanation of the bug :-) | 16:39 |
| clarkb | mordred: I am going to guess somethingrelated to the chnage in https://review.openstack.org/#/c/185972/3/nodepool/provider_manager.py | 16:39 |
| anteaya | ttx: I can find the information in the history of the meetings wikipage, but where should the links to meeting agendas live? | 16:39 |
| mordred | clarkb: yah. let me try something | 16:39 |
| jeblair | anteaya: it looks like you can link within the meeting descriptions... at least some of them do | 16:39 |
| anteaya | jeblair: do you see an example? I don't see one yet | 16:40 |
| jeblair | anteaya: barbican links to blueprints | 16:40 |
| pabelanger | mordred, Okay, so for the moment I am actually using Ubuntu 14.04 to do sbuild(debian) and mock(fedora) builds; for the POC. However, like you said, I could see somebody wanting a fedora base image. Getting a chroot into the base OS, is pretty easy. | 16:40 |
| mordred | pabelanger: yah. having a fedora base image - totally awesome | 16:41 |
| jeblair | clarkb, mordred: gah -- nodepool should never hit an exception that isn't logged, this is extremely troubling | 16:41 |
| *** pblaho has quit IRC | 16:41 | |
| anteaya | jeblair: so they do | 16:41 |
| clarkb | jeblair: ya not sure how iut happened | 16:41 |
| pabelanger | mordred, so, would you see a packaging-trusty and packaging-fedora image in nodepool then? | 16:41 |
| mordred | nope | 16:42 |
| mordred | pabelanger: I would see an ubuntu-trusty and fedora-22 image in nodepool | 16:42 |
| mordred | I do not think we want/need a different image for packaging | 16:42 |
| pabelanger | okay | 16:42 |
| jeblair | yeah, we really want to drive the number of images down to minimal | 16:42 |
| *** unicell has joined #openstack-infra | 16:42 | |
| flip214 | can somebody please explain me the meaning of "Cloning not allowed in this configuration" for https://review.openstack.org/#/c/178049/ (devstack plugin to install drbd)? I understand the literal meaning, but not the reason or how to fix it. | 16:43 |
| openstackgerrit | Merged openstack-infra/shade: Set metadata headers on object create https://review.openstack.org/186048 | 16:43 |
| jeblair | clarkb: why do you think it died? | 16:43 |
| *** jtomasek has quit IRC | 16:43 | |
| *** julim has joined #openstack-infra | 16:43 | |
| jeblair | flip214: you may need to add it to PROJECTS in the job configuration; there should be other examples related to devstack plugins | 16:43 |
| pabelanger | so, with that in mind, how could pre-caching of chroots happen atop of ubuntu-trusty? If I understand properly, once the jenkins slave came online, the chroots would be built then. | 16:44 |
| openstackgerrit | Anita Kuno proposed openstack-infra/irc-meetings: Correct an error in a meeting description. https://review.openstack.org/186075 | 16:44 |
| clarkb | jeblair: my hunch is an uncaught exception in the main runloop of the thread | 16:44 |
| pabelanger | unless we add packaging bits into the ubuntu-trusty dib? | 16:44 |
| flip214 | jeblair: I'll try to find that, thanks for the hint. | 16:45 |
| *** soren_ has quit IRC | 16:45 | |
| mordred | payes | 16:45 |
| mordred | gah | 16:45 |
| *** julim has quit IRC | 16:45 | |
| greghaynes | 'morning | 16:45 |
| mordred | pabelanger: yes. I think bulding the chroots in the ubuntu-trusty dib is a totally decent idae | 16:45 |
| greghaynes | clarkb: we broke nodepool? | 16:46 |
| openstackgerrit | Anita Kuno proposed openstack-infra/irc-meetings: Correct an error in a meeting description. https://review.openstack.org/186075 | 16:46 |
| jeblair | clarkb: are we running head? | 16:46 |
| pabelanger | mordred, okay. let me see how that would look | 16:46 |
| clarkb | greghaynes: not really, things that didn't work in the past continue to not work. The only real regression is the config reloading | 16:46 |
| clarkb | jeblair: yes | 16:46 |
| mordred | pabelanger: we'll want to take care to not leave the tools needed to build the chroot in the image itself, since that would pollute the minimal image | 16:46 |
| *** otter768 has joined #openstack-infra | 16:46 | |
| pabelanger | mordred, right, that is what I was thinking too | 16:46 |
| jeblair | clarkb: did you do a thread dump and see the thread was absent? | 16:46 |
| clarkb | jeblair: note its only broken on mordreds change | 16:46 |
| mordred | pabelanger: it's possible taht you may want to brainmeld with greghaynes to figure out how the dib caching mechanism might be used, and/or also how the building of the chroot might be done image-host side | 16:46 |
| clarkb | HEAD is fine | 16:47 |
| clarkb | and so is production | 16:47 |
| greghaynes | oh, dib things | 16:47 |
| jeblair | oh, you're running 185972 in local testing and seeing it break | 16:47 |
| clarkb | see https://review.openstack.org/#/c/185972/3 | 16:47 |
| mordred | jeblair: yah - this is us trying to debug why my change is breaking all the tests | 16:47 |
| jeblair | i missed that part :) | 16:47 |
| fungi | the pyyaml.org trac has been overrun by commentspam | 16:48 |
| fungi | that's unfortunate | 16:48 |
| *** radez is now known as radez_g0n3 | 16:48 | |
| clarkb | hrm the run method has a catch all Exception catcher | 16:48 |
| *** annegentle has joined #openstack-infra | 16:48 | |
| clarkb | oh wait I lied | 16:48 |
| pabelanger | mordred, one issue to keep in mind is tools like mock require user group changes, but that _could_ be pre-baked into dib. EG: adding jenkins to the mock group. | 16:49 |
| pabelanger | Q: Is there any existing job that launches puppet before actually doing a gate test? | 16:50 |
| *** timcline has quit IRC | 16:50 | |
| openstackgerrit | Clark Boylan proposed openstack-infra/nodepool: Debug dying task managers https://review.openstack.org/186079 | 16:51 |
| clarkb | mordred: ^ I think that will log the info you need | 16:51 |
| *** yamahata has quit IRC | 16:51 | |
| greghaynes | pabelanger: do sbuild and mock already support feeding them a pre-made chroot? | 16:51 |
| clarkb | pabelanger: you have sudo, you should affect that change at job run time if necessar | 16:51 |
| *** otter768 has quit IRC | 16:51 | |
| flip214 | jeblair: in the infra-project-config, do you mean in jenkins/jobs/….yaml an 'export PROJECTS="..."' ? | 16:51 |
| pabelanger | greghaynes, define feeding? building a chroot outside of mock or sbuild tooling? | 16:52 |
| flip214 | because we don't even have a job file in there yet. | 16:53 |
| greghaynes | pabelanger: im just trying to figure out, without dib how would you cache a chroot and have those tools use it | 16:53 |
| *** annegentle has quit IRC | 16:53 | |
| *** emagana has quit IRC | 16:53 | |
| pabelanger | greghaynes, right now, I do mock -r flavor init, which bootstraps my chroots | 16:53 |
| openstackgerrit | James E. Blair proposed openstack-infra/project-config: Upload ical output to log server https://review.openstack.org/186081 | 16:54 |
| Clint | sbuild can operate on a chroot that you've built | 16:54 |
| Clint | or an lvm snapshot or whatever | 16:54 |
| mordred | greghaynes, pabelanger: naive/stupid version is an element that does "yum install mock ; mock -r flavor init ; yum remove mock" | 16:54 |
| pabelanger | then, once my job launches from jenkins, I'd re-update the chroot (ensuring the latest) then do builds | 16:54 |
| greghaynes | I wonder if we should make the chroot with dib | 16:54 |
| jeblair | flip214: hrm. let me look. | 16:55 |
| pabelanger | mordred, Right, something like that | 16:55 |
| clarkb | I think it would be good to see the non optimized version of the job | 16:55 |
| clarkb | then worry about image changes | 16:55 |
| clarkb | because right now we all seem to be doing a lot of hand waving at each other | 16:56 |
| greghaynes | Yea. Sounds like it shouldnt be challenging to turn that into an element after the fact | 16:56 |
| greghaynes | which will make it happen at build time | 16:56 |
| clarkb | and having a concrete job that is running would help all involved | 16:56 |
| mordred | clarkb: AHA! | 16:57 |
| greghaynes | mordred: Re: nodepool consuming os-c-c directly - can we not make that a shade feature? | 16:57 |
| mordred | clarkb: I have foudn it I think | 16:57 |
| *** amotoki_ has joined #openstack-infra | 16:57 | |
| mordred | greghaynes: what? | 16:57 |
| clarkb | mordred: did my change help? I am waiting for jenkins to finish running the test | 16:57 |
| mordred | greghaynes: (moar words please) | 16:57 |
| mordred | clarkb: no - I found it with print statements | 16:57 |
| mordred | clarkb: I think | 16:57 |
| greghaynes | mordred: where shade will reload clients if the clouds.yaml it is using changes | 16:57 |
| pabelanger | greghaynes, Right, okay. So let me see how to add them into dib for ubuntu-trusty | 16:57 |
| pabelanger | I need to get my local nodepool running again | 16:58 |
| *** claudiub has joined #openstack-infra | 16:58 | |
| mordred | greghaynes: _possibly_ ... although we need to think about how to interface with other things like nodepool that might also be managing that | 16:58 |
| greghaynes | mordred: oh, right, were exposing clients entirely | 16:58 |
| mordred | greghaynes: so for now doign this seems safer - but I think it's TOTALLY something we should think about a design for | 16:58 |
| greghaynes | mordred: ok, yea, nodepool for now | 16:58 |
| *** asahlin has quit IRC | 16:59 | |
| greghaynes | pabelanger: Yea, I think what clarkb said is probably right though - first iteration as a normal old job should work fine and be a ton easier to debug / figure out | 16:59 |
| *** ashleighfarnham has joined #openstack-infra | 17:00 | |
| jeblair | flip214: to use the existing templates, your project should have been called "stackforge/devstack-plugin-drbd" | 17:00 |
| *** asahlin has joined #openstack-infra | 17:00 | |
| clarkb | mordred: nope no "manager died." string in the test output, this is weird | 17:00 |
| jeblair | flip214: since you named it "stackforge/drbd-devstack" you will need to make your own job definitions | 17:00 |
| clarkb | mordred: curious to see what you have found out, otherwise my next step may be pdb | 17:00 |
| jeblair | flip214: see this line: http://git.openstack.org/cgit/openstack-infra/project-config/tree/jenkins/jobs/devstack-gate.yaml#n1934 | 17:01 |
| pabelanger | greghaynes, clarkb: yes, I agree. Just making sure we are all on same page | 17:01 |
| clarkb | mordred: oh except I see a none type is not callable exception in your provider manager hange | 17:01 |
| *** amotoki_ has quit IRC | 17:01 | |
| jeblair | flip214: you could also rework that template to be more flexible for project names | 17:01 |
| flip214 | jeblair: can I rename a project? I didn't know about that requirement. | 17:02 |
| *** HeOS has quit IRC | 17:02 | |
| clarkb | mordred: the config value passed to shade there is supposed to have a callable member | 17:02 |
| jeblair | flip214: yes, it will take a few weeks. | 17:02 |
| pabelanger | So, next step, installing packaging tool bits. To me, having puppet do that makes sense, eg: install mock, sbuild. However, I don't know if we have any jobs that are launching puppet as the first step for a jenkins slave? | 17:02 |
| pabelanger | other wise, we put the bits into a bash script | 17:02 |
| clarkb | mordred: http://paste.openstack.org/show/240318/ | 17:02 |
| pabelanger | which, is meh for me | 17:02 |
| greghaynes | pabelanger: most people seem to just have a bash script in a job | 17:02 |
| *** deepakcs has quit IRC | 17:02 | |
| greghaynes | pabelanger: there is work on using bindep as a general solution, but thats not done yet | 17:03 |
| clarkb | not using puppet is more friendly to the masses | 17:03 |
| flip214 | jeblair: ;[ | 17:03 |
| clarkb | but if puppet is the best tool then ok | 17:03 |
| pabelanger | greghaynes, right | 17:03 |
| *** hemna is now known as hemnafk | 17:03 | |
| mordred | clarkb: right. so - the question is how is self.provider existing without getting a cloud_config object on it | 17:03 |
| flip214 | jeblair: would you recommend creating a new project with the correct name instead? | 17:03 |
| mordred | clarkb: which may just mean that a fake isn't faking enough things | 17:03 |
| clarkb | mordred: no I think the type is wrong entirely | 17:03 |
| *** shardy has quit IRC | 17:03 | |
| clarkb | mordred: oh wait its none so duh | 17:04 |
| mordred | right | 17:04 |
| mordred | now - you may be right about the type ... | 17:04 |
| pabelanger | clarkb, well, to me puppet makes it easy to add support for more OS's. But that is my personal opinion. Not a deal breaker for me to use a bash script launched from jenkins job. | 17:04 |
| clarkb | pabelanger: sure except tomorrow you will be field questions from all sides on why you didn't use chef/ansible/salt/cfengine/etc | 17:05 |
| pabelanger | right | 17:05 |
| fungi | flip214: please don't create new repos to correct misnaming. if it really needs to be renamed then put it in the queue to go through the renaming process, but in the meantime adding a new job-template or making the existing one more flexible should be a much faster solution | 17:05 |
| clarkb | pabelanger: and in a couple months there will be a separate branch for all the builds to implement in ^ | 17:05 |
| clarkb | pabelanger: if you just use bash no one seems to complain, its really weird | 17:05 |
| jeblair | flip214: no. i'd recommend: 1) update the template to allow repo names that don't fit the pattern (do this in any case to keep things moving for now). 2) optionally rename your repo if you want to be like the others. 3) if you do 2, consider putting the template back after the rename | 17:05 |
| pabelanger | clarkb, Ya, i see that. | 17:05 |
| pabelanger | okay, so for the simple fact we don't run puppet apply on jenkins jobs launching, is likely a reason not to use puppet. However, it _could_ be added | 17:06 |
| flip214 | jeblair: fungi: thank you for your help. | 17:06 |
| flip214 | If only I knew all the interrelations and dependencies here... | 17:07 |
| clarkb | mordred: my guess is that we don't have clouds.yaml in place and without it oscc doesnt dothe right thing merging items? | 17:07 |
| mordred | oscc should TOTALLY do the right thing without a clouds.yaml in place | 17:07 |
| clarkb | mordred: then get_one_cloud returns None | 17:07 |
| mordred | everythign in the universe breaks if it doesn't | 17:07 |
| *** dboik has quit IRC | 17:07 | |
| flip214 | thanks a lot, will try that when I'm back. | 17:07 |
| jeblair | flip214: you're welcome; sorry that didn't get caught in review. | 17:08 |
| *** fhubik is now known as fhubik_afk | 17:08 | |
| mordred | get_one_cloud will overlay the passed in parameters on top of an empty cloud if there is no clouds.yaml | 17:08 |
| clarkb | mordred: hrm ya get_one_cloud returns a CloudConfig always | 17:08 |
| *** sshnaidm has quit IRC | 17:09 | |
| *** annegentle has joined #openstack-infra | 17:10 | |
| mordred | clarkb: so - I'm getting a None returned from self._getClient() in provider_manager | 17:10 |
| mordred | clarkb: inside of resetClient | 17:10 |
| mordred | oh - hang on - might be debugging bong | 17:11 |
| *** dboik has joined #openstack-infra | 17:12 | |
| *** fhubik_afk is now known as fhubik | 17:12 | |
| *** ociuhandu has joined #openstack-infra | 17:12 | |
| *** tiswanso has quit IRC | 17:12 | |
| *** baoli_ has quit IRC | 17:12 | |
| *** annegentle has quit IRC | 17:13 | |
| clarkb | mordred: random thought oscc should be a dep in requiremnts.txt | 17:13 |
| *** markus_z has quit IRC | 17:14 | |
| *** sdake_ has joined #openstack-infra | 17:14 | |
| *** julim has joined #openstack-infra | 17:14 | |
| pabelanger | greghaynes, are the puppet infra bits will on the dib after the puppet element is run? | 17:14 |
| clarkb | mordred: also the inequality check on CloudConfig objects isn't going to work I don't think | 17:14 |
| pabelanger | I _think_ so | 17:14 |
| mordred | clarkb: yes. I agree | 17:15 |
| mordred | clarkb: why not? | 17:15 |
| clarkb | pabelanger: ya we don't clean up /etc/puppet/modules or the puppet install | 17:15 |
| mordred | clarkb: well, I mean, it will once my patch to add equality support to CloudConfig objects lands | 17:15 |
| clarkb | mordred: bceause its just going to compare objects but we always get a new object so it will always be unequal | 17:15 |
| clarkb | oh is that part of the oscc stack? | 17:15 |
| pabelanger | clarkb, okay, that might work for my POC | 17:15 |
| clarkb | mordred: I wonder if the unittests installing last oscc release is related? | 17:16 |
| clarkb | mordred: maybe try manually installing your oscc changes to the tox venv then rerun | 17:16 |
| mordred | clarkb: yes. I have already installed tip of oscc in my local tox env | 17:16 |
| *** dboik has quit IRC | 17:16 | |
| clarkb | kk | 17:16 |
| *** e0ne is now known as e0ne_ | 17:16 | |
| *** kutija has joined #openstack-infra | 17:16 | |
| *** gokrokve has joined #openstack-infra | 17:17 | |
| *** sdake has quit IRC | 17:17 | |
| greghaynes | clarkb: mordred do youall have a testname youall are using to debug? | 17:18 |
| *** SumitNaiksatam has quit IRC | 17:18 | |
| mordred | greghaynes: I'm running nodepool.tests.test_nodepool.TestNodepool.test_dib_node | 17:18 |
| clarkb | greghaynes: nodepool.tests.test_shade_integration.TestShadeIntegration.test_nodepool_provider_config | 17:18 |
| *** puranamr has joined #openstack-infra | 17:18 | |
| clarkb | I think ^ is the test that most closely targets the broken code | 17:18 |
| greghaynes | and thats just after checking out https://review.openstack.org/#/c/185972/3 ?e | 17:19 |
| clarkb | greghaynes: ya | 17:19 |
| clarkb | mordred: ok, there is a type issue with the config var | 17:20 |
| clarkb | mordred: shade expects it to be an OpenStackConfig but get_one_cloud returns CloudConfig | 17:20 |
| *** e0ne_ is now known as e0ne | 17:20 | |
| mordred | clarkb: aha! thank you | 17:20 |
| clarkb | mordred: I think thats where the None type comes from | 17:20 |
| fungi | jeblair: mordred: clarkb: SergeyLukjanov: jhesketh: pleia2: i'm going over https cert expirations coming up in july... jenkins.o.o is among them. any objections to me replacing that with a self-signed cert like we have on jenkinsNN.o.o servers? | 17:21 |
| clarkb | because CloudConfig.get_one_cloud is none and thats not callable | 17:21 |
| openstackgerrit | Matt Riedemann proposed openstack/requirements: Add tooz to projects.txt https://review.openstack.org/186090 | 17:21 |
| mordred | ++ | 17:21 |
| pleia2 | fungi: no objections | 17:21 |
| clarkb | fungi: the only concern there is possibly that we use that node for "trusted" builds, but its easy enough to confirm validity of artifacts otherwise (and https on jenkins.o.o likely doesn't help that too much) | 17:22 |
| *** fawadkhaliq has joined #openstack-infra | 17:22 | |
| *** krtaylor has quit IRC | 17:22 | |
| * jogo wonders how https://review.openstack.org/#/c/185365/ is different then all the other CI related stuff we have | 17:22 | |
| jeblair | fungi: nope | 17:22 |
| greghaynes | so, are the nodepool tests actually finding fails? | 17:22 |
| greghaynes | this is crazypants | 17:22 |
| *** puranamr has quit IRC | 17:22 | |
| clarkb | greghaynes: yup | 17:22 |
| clarkb | the test suite is actually reasonably robust at this point. Not perfect but our coverage is decent | 17:23 |
| *** fhubik has quit IRC | 17:23 | |
| fungi | jogo: looks like a nih zuul? | 17:23 |
| clarkb | fungi: jogo yes thats what it is | 17:23 |
| greghaynes | maybe request a project rename? | 17:23 |
| * greghaynes tries to be less snarky | 17:23 | |
| clarkb | I talked to them briefly about it, they didn't want to have to run jenkins (you don't have to), wanted lightweight python proess (we have that), so was trying to stear towards zuul but :/ | 17:23 |
| fungi | i find it amusing that changes proposed to the nih zuul will be coordinated by... actual zuul ;) | 17:24 |
| *** yamahata has joined #openstack-infra | 17:24 | |
| jogo | fungi: hehe | 17:24 |
| *** puranamr has joined #openstack-infra | 17:24 | |
| *** fhubik has joined #openstack-infra | 17:24 | |
| jogo | thanks, just checking if it was just me or not | 17:24 |
| jeblair | fungi: well, i mean, they would theoretically if we approved that change | 17:24 |
| clarkb | steer? english is hard | 17:24 |
| jeblair | i am inclined to -2 it. | 17:25 |
| *** e0ne has quit IRC | 17:25 | |
| clarkb | I think one is a cow the other is a driving direction thing | 17:25 |
| *** yamamoto has joined #openstack-infra | 17:25 | |
| *** baoli has joined #openstack-infra | 17:26 | |
| *** tiswanso has joined #openstack-infra | 17:26 | |
| mordred | jeblair: I'm inclined to abstain indefinitely | 17:26 |
| fungi | would be nice to see the rally-ci authors just help us solve their use case with zuul | 17:26 |
| jeblair | mordred: why abstain? if you do, someone will approve that change. | 17:26 |
| clarkb | oh I remember the other objects was the complicated config for zuul | 17:26 |
| clarkb | er other objection | 17:27 |
| jeblair | clarkb: where was that objection voiced? | 17:27 |
| clarkb | I think they are interested in a daemon that does the one thing that they want. | 17:27 |
| *** fawadkhaliq has quit IRC | 17:27 | |
| *** dizquierdo has quit IRC | 17:27 | |
| clarkb | jeblair: in a rally meeting a few weeks back, I can dig it up on eavesdrop | 17:27 |
| jeblair | well, i mean, they could have talked to us | 17:27 |
| *** mrunge has quit IRC | 17:27 | |
| fungi | i find rally confusing and hard to implement. i'm not going to let them know that, but i'm going to go ahead and write a replacement | 17:28 |
| clarkb | http://eavesdrop.openstack.org/meetings/rally/2015/rally.2015-04-21-17.02.log.txt there | 17:28 |
| *** ParsectiX has joined #openstack-infra | 17:29 | |
| *** yamamoto has quit IRC | 17:30 | |
| openstackgerrit | Merged openstack-infra/project-config: Clean up run-tox.sh to make it easier to reuse https://review.openstack.org/161414 | 17:30 |
| fungi | i already -2'd. rally is an official openstack project. ci systems are well outside the scope they presented to the tc | 17:30 |
| *** alexsyip has quit IRC | 17:30 | |
| *** emagana has joined #openstack-infra | 17:31 | |
| clarkb | mordred: I wonder if CloudConfig should go away and OpenStackConfig be the object | 17:31 |
| jeblair | fungi: well, their change had an error. it was _mostly_ proposing stackforge, except the acl was mistakenly in openstack/ | 17:31 |
| *** puranamr has quit IRC | 17:31 | |
| clarkb | mordred: and for backward compat just have CloudConfig inherit OpenStackConfig with no overrides | 17:32 |
| jeblair | fungi: even if we assume it's going to stackforge, i'm still opposed to it -- it duplicates a project in the infra program, which is official. | 17:32 |
| clarkb | mordred: or monkey patch it | 17:32 |
| *** puranamr has joined #openstack-infra | 17:32 | |
| clarkb | mordred: otherwise shade has to know about this in weird weays | 17:32 |
| jeblair | (also, i'm becoming increasingly -1 on new stackforge projects) | 17:32 |
| jeblair | fungi: i have -2d it on the grounds they should work with zuul. | 17:32 |
| *** alexsyip has joined #openstack-infra | 17:33 | |
| clarkb | mordred: it isn't clear to me how an openstack config is different than a cloud config within in the context of openstack client config. They should be the same? | 17:33 |
| mordred | clarkb: an OpenStackConfig represents the total amount of config found on the system. a CloudConfig is the config of one cloud | 17:34 |
| clarkb | mordred: hrm, in that case I think shade is expecting the wrong config object type? since thats config for one cloud | 17:35 |
| mordred | yes. I'm fixing right now | 17:35 |
| clarkb | cool | 17:35 |
| fungi | i agree they should at least discuss zuul-related options with us and clear up any misconception before running off to redesign it in a vacuum, but remain unconvinced that we should reject stackforge projects only on the grounds that they duplicate functionality of existing official projects (no matter how hypocritical the choice) | 17:35 |
| *** julim has quit IRC | 17:35 | |
| jogo | clarkb: what a classic line "the only problem with zuul is too complex for configuration" replace zuul with anything else | 17:35 |
| jogo | X is too big and complex, lets re-write it as Y and make it simple. 2 years later Y is too big and complex | 17:36 |
| fungi | but yes if we start down the road of "all projects in our gerrit must join the big tent or get out" then the waters get a little more muddy there | 17:36 |
| jeblair | fungi: fungi " | 17:36 |
| jeblair | fung: "Where it makes sense, the project cooperates with existing projects rather than gratuitously competing or reinventing the wheel" | 17:36 |
| dstufft | jogo: kill all configuration | 17:36 |
| jeblair | fungi: is the tc stance on such things | 17:36 |
| *** alexsyip has quit IRC | 17:37 | |
| mordred | jeblair: otoh, the TC has explicitly declined governance on stackforge. we have not yet gotten the agreement that stackforge is only a historical bug | 17:37 |
| jeblair | fungi: and yes, i think we should both apply that to stackforge (stackforge is not the place to be in openstack while escaping the requirements of being in openstack) | 17:37 |
| *** yfried has joined #openstack-infra | 17:38 | |
| mordred | jeblair: I think this might be an excellent motivating factor to get that agreement | 17:38 |
| fungi | jeblair: i'd love to see us start applying that to stackforge projects too, yes. do we need a tc resolution to grandfather existing stackforge projects into the big tent, or just shut it down officially where new project requests are concerned and grandfather the original ones into continuing their own wild west? | 17:38 |
| jogo | dstufft: everything should just have this UI: https://i.imgur.com/iMQcodA.jpg | 17:38 |
| jeblair | mordred: the TC has declined governance, but it also created stackforge | 17:38 |
| mordred | jeblair: yes indeed | 17:38 |
| dstufft | jogo: unironically agree | 17:38 |
| dstufft | jogo: too bad reality | 17:38 |
| mordred | jeblair: I agree with your position - I just think we need to actually get the TC to decide it | 17:38 |
| jogo | dstufft: silly reality | 17:38 |
| *** julim has joined #openstack-infra | 17:39 | |
| *** hemnafk is now known as hemna | 17:39 | |
| jeblair | mordred: so while i agree with you on moving forward for clarity, i also think that even in the current murky situation i think we can apply some judgement and say "it does not make sense for the project to negate its own goals by supporting this" | 17:39 |
| dstufft | jogo: I've recently taken a view point that every configuration option should be thought of as some sort of "reality debt" or something, and projects (not openstack specific, just projects in general) should strive to eliminate them where possible. Too often things, especially OSS things, use configuration as an excuse tnot to make a decision | 17:39 |
| mordred | jeblair: totally | 17:40 |
| mordred | jeblair: also, I reserve the right to -2 things because and only because I have exercised personal judgement | 17:40 |
| openstackgerrit | gordon chung proposed openstack/requirements: Pin pyngus version as python-qpid-proton is not in pypi https://review.openstack.org/186098 | 17:40 |
| jeblair | mordred: indeed | 17:40 |
| mordred | without ANY policy backing me up | 17:40 |
| jogo | dstufft: I don't disagree but the devil is in the details | 17:40 |
| mordred | I am more than happy if people want to complain | 17:40 |
| dstufft | jogo: for sure :) | 17:40 |
| jeblair | mordred, fungi: i will work on a resolution to clarify the future of stackforge, but i think the -2 should stand on this because i think the goals we have set as a community are important. | 17:41 |
| mordred | jeblair: agree | 17:41 |
| dstufft | jogo: maybe it was only me that the "OSS uses configuration to avoid making a choice" thing was a realization/revelation and not completely obvious :) | 17:41 |
| *** sputnik13 has joined #openstack-infra | 17:41 | |
| jogo | dstufft: so the inverse is a problem too: X doesn't support my use case and isn't willing to so I now will go and re-invent the wheel but with more knobs | 17:41 |
| *** teran_ has quit IRC | 17:42 | |
| devananda | question on reviewstats -- for projects that have discrete review teams for discrete subprojects, what's the best way to represent that? | 17:42 |
| fungi | jeblair: i'm willing to go along with whatever the tc decides on it, but also completely support your autonomy to -2 anything you judge problematic | 17:42 |
| devananda | eg, ironic has ironic-core, ironic-python-agent-core, ironic-discoverd-core, bifrost-core, and if you count drivers, also proliantutils-core | 17:43 |
| devananda | (those are all gerrit groups) | 17:43 |
| dstufft | jogo: yea :/ | 17:43 |
| mordred | I think there shold be One Clear Default value that makes sense for anything in a config | 17:44 |
| fungi | devananda: it's tricky. bknudson started https://review.openstack.org/139727 in an attempt to use my who-approves.py to dynamically generate/update review teams | 17:44 |
| mordred | requiring config for basic operation is fail | 17:44 |
| jeblair | dstufft, jogo: to that end, our configuration of zuul is exceedingly complex. however, the basic configuration of zuul is quite simple, and, mordred, yeah, the defaults are, i think, sane. | 17:44 |
| mordred | however, exposing config to solve jogo's thing | 17:44 |
| mordred | jeblair: ++ | 17:44 |
| *** SumitNaiksatam has joined #openstack-infra | 17:44 | |
| jeblair | our unit tests still test the same configuration we ran with zuul in 3 years ago. | 17:44 |
| devananda | fungi: oh that looks like fun :-/ | 17:44 |
| jogo | ++ for sane defaults, this really sounds like fear of using other peoples code because it is unknown | 17:45 |
| fungi | devananda: that is to say, it's currently not a solved problem. the infra projects currently just mark the infra-core team as core reviewers in reviewstats config, which is suboptimal for the reasons you mention | 17:45 |
| devananda | fungi: but doesn't address the question - it assumes an answer | 17:45 |
| *** dustins has quit IRC | 17:45 | |
| *** redixin has joined #openstack-infra | 17:45 | |
| devananda | yea. there are and will continue to be people with approval rights on subprojects, and more projects are adopting this "delegation of trust" model | 17:46 |
| bknudson | fungi: devananda: I didn't get too far on that... the issues that I ran into are mentioned at the top of the part: https://review.openstack.org/#/c/139727/1/tools/who-approves/reviewstats_projects.py | 17:46 |
| *** ildikov has joined #openstack-infra | 17:46 | |
| fungi | devananda: the real problem, i think, is that reviewstats assumes one particular team structure when across openstack we have a variety of team composition models | 17:46 |
| jeblair | jogo: and yeah, reading the meeting logs they clearly didn't know about major features of zuul. i honestly think they could do what they want with zuul with a minimum of fuss. which clarkb attempted to convey but they were not receptive. | 17:46 |
| bknudson | and it's the subprojects part that is not encoded anywhere. | 17:46 |
| devananda | fungi: yes | 17:46 |
| jeblair | they said they would look into it and regroup at the next meeting. they did not. | 17:46 |
| *** koolhead17 has quit IRC | 17:46 | |
| devananda | fungi: and different projects are adopting different approaches to suit their unique/evolved cultures | 17:46 |
| bknudson | Hopefully people are using reviewstats as a guide and not as having all the info. | 17:48 |
| fungi | devananda: i'm also not sure that it even makes sense to try to enumerate all the ways in which subteams might be structured and interrelated. the closest we can come with any accuracy is introspecting gerrit to identify accounts with workflow +1 privs on a per-repo basis | 17:48 |
| fungi | we can sort of infer team relationships from gerrit groups | 17:49 |
| greghaynes | mordred: hah, you renamed cloud to profile already | 17:49 |
| *** dboik has joined #openstack-infra | 17:49 | |
| mordred | greghaynes: :) | 17:49 |
| devananda | fungi: thus my question :) | 17:49 |
| *** emagana has quit IRC | 17:50 | |
| *** Sukhdev has quit IRC | 17:50 | |
| *** asselin__ has quit IRC | 17:51 | |
| devananda | fungi: for example, neutron seems to be splitting along the driver boundary, with one team per driver. OTOH, Ironic teams are split along API boundaries, with subteams for different services | 17:51 |
| fungi | devananda: but if we redesign, it probably makes the most sense to analyze these sorts of team structures objectively (things we can actually query/measure), rather than based on some assertion they make about their structures (which may be based entirely on human judgement and therefore can't be easily represented via metadata) | 17:51 |
| *** ivar-lazzaro has joined #openstack-infra | 17:51 | |
| devananda | fungi: I'm not sure yet what the desired representation even is, to be honest | 17:52 |
| devananda | so let's not redesign anything just yet :) | 17:52 |
| *** dannywilson has joined #openstack-infra | 17:52 | |
| fungi | devananda: third data point: the infra-council https://review.openstack.org/182811 | 17:52 |
| mriedem | fungi: we need a ninja merge to unblock all of stable/juno and stable/kilo | 17:53 |
| *** chlong has quit IRC | 17:53 | |
| devananda | hehe | 17:53 |
| jogo | jeblair: agreed | 17:53 |
| mriedem | fungi: clarkb: jeblair: https://review.openstack.org/#/c/186084/ needs ninja merge | 17:53 |
| mriedem | dims_: dhellmann: harlowja_: ^ | 17:53 |
| fungi | mriedem: do you have a brief summary of the deadlock? | 17:53 |
| mriedem | fungi: yeah, https://etherpad.openstack.org/p/stable-tracker | 17:54 |
| mriedem | top thing under kilo | 17:54 |
| fungi | thanks | 17:54 |
| mriedem | and https://bugs.launchpad.net/python-tooz/+bug/1459322 | 17:54 |
| openstack | Launchpad bug 1459322 in tooz "stable/juno broken with zake 0.2.2 release" [Undecided,New] | 17:54 |
| mriedem | oh crap | 17:55 |
| fungi | mriedem: i'm having a hard time deriving a summary from that | 17:55 |
| mriedem | https://review.openstack.org/#/c/186084/ depends on a thing that | 17:55 |
| *** sdake has joined #openstack-infra | 17:55 | |
| *** dannywil_ has joined #openstack-infra | 17:55 | |
| mriedem | fungi: tooz has uncapped requirements on kazoo and zake in 0.12 which is what juno pulls in | 17:55 |
| mriedem | zake 0.2.2 released today with a block on the version of kazoo that is used in juno | 17:55 |
| fungi | that helps | 17:55 |
| mriedem | so we have to cap kazoo and zake in tooz on stable/juno | 17:55 |
| mriedem | and then left the cap on tooz in g-r on stable/juno to get a new release of tooz with those caps | 17:55 |
| mriedem | however, i either need to squash or re-order https://review.openstack.org/#/c/186084/ and it's dependent change for .gitreview | 17:56 |
| mriedem | dhellmann: do you have a preference there? ^ | 17:56 |
| *** dannywilson has quit IRC | 17:56 | |
| mriedem | *lift the cap | 17:56 |
| devananda | fungi: right - that's neat. and another good data point | 17:56 |
| *** aduarte has joined #openstack-infra | 17:57 | |
| devananda | fungi: in the meantime, I'd like to not have approvers on subprojects be not counted. if you have no opinions at the moment, i'll just throw something up that adds them, and we'll sort out details later | 17:57 |
| *** timcline has joined #openstack-infra | 17:57 | |
| *** sarob has joined #openstack-infra | 17:58 | |
| *** sdake_ has quit IRC | 17:58 | |
| fungi | devananda: i think it's entirely reasonable that if ironic is okay with the tracking disparity then their team in reviewstats represents all reviewers with approval on at least one repo within that project | 17:58 |
| mordred | devananda: http://stackalytics.com/report/contribution/ironic-group/30 | 17:59 |
| mordred | devananda: tracks contributions across all of the ironic projects | 17:59 |
| fungi | that looks surprisingly like reviewstats mostly reimplemented inside of stackalytics. neat | 17:59 |
| devananda | ohyah. except it looks like it's pulling from gerrit groups and handling inheritance | 18:00 |
| mordred | yes. people said they wanted the data that was in reviewstats. stackalytics has had that for at least 1.5 years | 18:00 |
| devananda | eg, it's showing TheJulia as a core, because bifrost-core is a subgroup of ironic-core | 18:00 |
| devananda | neat | 18:00 |
| mordred | well, actually | 18:00 |
| mordred | what it does is loko for people who have cast +2 or -2 votes | 18:00 |
| mordred | since you cannot do that if you are not core | 18:00 |
| devananda | mordred: on which projects? | 18:01 |
| mordred | devananda: on the projects defined in the governance repo as ironic's projects | 18:01 |
| devananda | oh | 18:01 |
| clarkb | mordred: stackalytics should serve http://en.wikipedia.org/wiki/Loco_moco | 18:01 |
| nikeshm | hi | 18:01 |
| * devananda forgets all about reviewstats | 18:01 | |
| devananda | this is actually smarter | 18:01 |
| nikeshm | in my nodepool vm | 18:01 |
| nikeshm | ram size is always 6gb | 18:01 |
| pleia2 | the stats are slightly different between stackalytics and reviewstats, so it's interesting to browse them both | 18:02 |
| nikeshm | either i am selecting 8gb ram or 16 ram flavor | 18:02 |
| *** dims_ has quit IRC | 18:02 | |
| fungi | clarkb: the tiki bar across the street from me has that, but they call it a "garbage plate" | 18:02 |
| nikeshm | patrickeast: hi | 18:02 |
| fungi | goes great with beer | 18:02 |
| mordred | clarkb: I want to eat that | 18:03 |
| *** achanda_ has joined #openstack-infra | 18:03 | |
| *** dims_ has joined #openstack-infra | 18:03 | |
| *** fhubik has quit IRC | 18:04 | |
| *** sarob_ has joined #openstack-infra | 18:04 | |
| mriedem | fungi: oooo | 18:04 |
| mriedem | upstate new york | 18:05 |
| mriedem | home of the garbage plate :) or so i'm told | 18:05 |
| * Clint squints. | 18:05 | |
| clarkb | mordred: clearly we need to find a way to go to hawii | 18:05 |
| pleia2 | ++ | 18:05 |
| clarkb | mordred: we can eat all of the amazing foods, but mostly just poke because it wins | 18:05 |
| *** achanda has quit IRC | 18:05 | |
| *** e0ne has joined #openstack-infra | 18:06 | |
| *** e0ne is now known as e0ne_ | 18:06 | |
| *** sarob has quit IRC | 18:06 | |
| * fungi loves poke salad | 18:06 | |
| *** ParsectiX has quit IRC | 18:06 | |
| fungi | which is two things in two completely different places in the world, but both great | 18:07 |
| *** e0ne_ is now known as e0ne | 18:07 | |
| clarkb | when I go to hawaii I buy a tub of poke and it it three meals a day | 18:08 |
| fungi | i would totally do that too | 18:08 |
| pleia2 | tub of poke <3 | 18:08 |
| *** Sukhdev has joined #openstack-infra | 18:08 | |
| fungi | much like i subsist on ceviche in other places | 18:09 |
| clarkb | fungi: and rum and coke | 18:09 |
| openstackgerrit | Spencer Krum proposed openstack-infra/project-config: Add beaker-localhost and beaker-nodepool hypervisors https://review.openstack.org/185754 | 18:09 |
| clarkb | fungi: you can get that in hawaii too | 18:09 |
| * fungi doubts the hawaiians call it a "cuba libre" | 18:09 | |
| openstackgerrit | K Jonathan Harker proposed openstack-infra/jeepyb: Swap logging levels for command and its output https://review.openstack.org/186110 | 18:10 |
| fungi | but as hawaii usually has decent sake on hand, i'd much prefer that over rum and coke | 18:10 |
| clarkb | fungi: I think they call it a mai tai and drop the coke | 18:10 |
| openstackgerrit | Merged openstack/requirements: Block oslo.vmware 0.13.0 due to a backwards incompatible change https://review.openstack.org/185748 | 18:10 |
| fungi | heh | 18:12 |
| clarkb | though I suppose mai tai is more than just lime | 18:13 |
| *** Sukhdev has quit IRC | 18:13 | |
| Clint | and a cuba libre is more than just rum&coke | 18:14 |
| clarkb | ya thats there the lime comes in | 18:14 |
| fungi | indeed. it's also false hope and dashed dreams | 18:14 |
| *** doug-fish has quit IRC | 18:14 | |
| * Clint sheds a solitary tear | 18:15 | |
| fungi | mriedem: did you come to a conclusion on the tooz change? | 18:15 |
| clarkb | fungi: mriedem rereading the tooz thing, couldnt we just update the pin in tooz and release? | 18:15 |
| clarkb | I am not grokking why it requires force merging | 18:16 |
| *** krtaylor has joined #openstack-infra | 18:16 | |
| mriedem | clarkb: in g-r? | 18:16 |
| clarkb | mriedem: and in tooz | 18:16 |
| mriedem | circular dep | 18:16 |
| clarkb | mriedem: where? | 18:16 |
| mriedem | https://review.openstack.org/#/c/186065/ | 18:16 |
| mriedem | that's g-r on stable/juno | 18:16 |
| mriedem | we need https://review.openstack.org/#/c/186084/ to release tooz 0.12.1 | 18:16 |
| mriedem | which will make https://review.openstack.org/#/c/186065/ pass | 18:16 |
| clarkb | oh because we failed at pinning :/ | 18:16 |
| mriedem | yes | 18:16 |
| mriedem | we borked ourselves | 18:16 |
| mriedem | fungi: i'm good with https://review.openstack.org/#/c/186084/ if you're ok with merging that and the dependent change which is just a .gitreview update | 18:17 |
| clarkb | ya, I see now | 18:17 |
| mriedem | for the new branch | 18:17 |
| mriedem | dhellmann: are you happy with this now? https://review.openstack.org/#/c/186084/ | 18:17 |
| clarkb | if we force merge the requirements update can we then let 186084 go through normally? | 18:18 |
| fungi | mriedem: well, you need to do something about the parent commit first if we're going to circumvent gating | 18:18 |
| mriedem | clarkb: i think it has to be the other way around | 18:18 |
| dhellmann | mriedem: sorry, got side-tracked with another issue, catching up now | 18:18 |
| mriedem | clarkb: we need https://review.openstack.org/#/c/186084/ merged to tag and release tooz 0.12.1 | 18:18 |
| mriedem | clarkb: which will then make https://review.openstack.org/#/c/186065/ pass | 18:19 |
| dhellmann | mriedem: squashing the commits makes sense to me | 18:19 |
| mriedem | dhellmann: fungi: ok, squashing now | 18:19 |
| clarkb | mriedem: I think they will end up passing if either one merges first | 18:19 |
| *** maishsk_afk has quit IRC | 18:19 | |
| clarkb | mriedem: because you break the cycle in either case | 18:19 |
| mriedem | clarkb: not sure how that's true in the tooz patch | 18:19 |
| clarkb | ut I have to think about it a bit more and figure out where the fails are | 18:19 |
| *** BharatK has quit IRC | 18:20 | |
| clarkb | mriedem: http://logs.openstack.org/84/186084/2/check/gate-tempest-dsvm-neutron-src-tooz/d369f8a/logs/devstacklog.txt.gz#_2015-05-27_17_50_16_856 | 18:20 |
| fungi | unless the requirements change passing needs a tooz release first | 18:20 |
| dhellmann | it will | 18:20 |
| dhellmann | if we can get tooz itself unblocked, we can land the requirements changes cleanly | 18:21 |
| clarkb | dhellmann: right my suggestion is to force merge the requirements change then actually test that tooz can be unblocked | 18:21 |
| mriedem | dhellmann: done https://review.openstack.org/#/c/186084/ | 18:21 |
| dhellmann | clarkb: interesting | 18:21 |
| clarkb | dhellmann: bceause force merging requirements and letting jenkins test tooz is preferable to making a tooz release then finding out we haven't fix it | 18:21 |
| clarkb | dhellmann: because if that happens then we do this dance again | 18:21 |
| clarkb | but I am not sure if we can test it that way because requirements needs a new release of tooz in the first place | 18:22 |
| *** dboik has quit IRC | 18:22 | |
| clarkb | (consuming from releases vs consuming from git is what makes this complicated) | 18:22 |
| mriedem | clarkb: but http://logs.openstack.org/84/186084/2/check/gate-tempest-dsvm-neutron-src-tooz/d369f8a/logs/devstacklog.txt.gz#_2015-05-27_17_50_16_856 is not pulling in the capped zake from the change itself | 18:22 |
| clarkb | mriedem: ya which tells me there may be something else at fault | 18:22 |
| mriedem | so even if we force merge the g-r raised cap on tooz, there is no tooz 0.12.1 to test it | 18:22 |
| *** kmartin has quit IRC | 18:22 | |
| clarkb | mriedem: because thats the src job | 18:22 |
| clarkb | mriedem: right but the src job is the one failing | 18:23 |
| mriedem | clarkb: i think that's because ceilometer is pulling in tooz 0.12 | 18:23 |
| mriedem | http://logs.openstack.org/84/186084/2/check/gate-tempest-dsvm-neutron-src-tooz/d369f8a/logs/devstacklog.txt.gz#_2015-05-27_17_47_30_787 | 18:23 |
| mriedem | that's another part of this onion, once we've updated g-r we have to sync g-r to ceilometer on stable/juno and merge that too | 18:23 |
| clarkb | hrm, in that case I think the src job is broken but we can come back to that later | 18:23 |
| mriedem | yeah, those are weeds i'm not familiar with | 18:24 |
| clarkb | so ya I think you may be correct about needing tooz release and force merge especially if the src job has that interaction | 18:25 |
| openstackgerrit | Merged openstack-infra/project-config: Upload ical output to log server https://review.openstack.org/186081 | 18:26 |
| * dhellmann wonders if the src job installs a package, then uninstalls it and replaces it with src later | 18:26 | |
| clarkb | dhellmann: oh I bet thats it, but since it doesn't do a pip install -U the old conflicting version is still in place | 18:27 |
| clarkb | this is a corner case that lifeless' pip fixes should fix I think | 18:27 |
| dhellmann | or it hasn't gotten to that point when it fails | 18:27 |
| clarkb | as it should detect that the resolution of that package has changed and install a newer version | 18:27 |
| dhellmann | we should make sure we install the libs from src before anything else in devstack | 18:27 |
| openstackgerrit | Merged openstack-infra/tripleo-ci: Remove nova temprevert https://review.openstack.org/182695 | 18:27 |
| clarkb | dhellmann: ++ | 18:28 |
| *** freerunner has quit IRC | 18:28 | |
| dhellmann | but in the short term, what's safer? force merging the requirements update? and does that actually let us land the tooz fix? | 18:28 |
| mriedem | dhellmann: that doesn't help | 18:28 |
| *** apmelton is now known as apmelton_away | 18:29 | |
| clarkb | dhellmann: I believe mriedem is correct and it won't help | 18:29 |
| dhellmann | so we have to do both? | 18:29 |
| mriedem | we need ninja merge tooz fix, release tooz 0.12.1, recheck g-r change, sync to ceilometer in juno and merge that | 18:29 |
| dhellmann | force both, that is? | 18:29 |
| mriedem | then garbage plates all around | 18:29 |
| dhellmann | k | 18:29 |
| clarkb | dhellmann: no, the original plan of ninjamerge tooz, release 0.12.1 should work | 18:29 |
| mriedem | dhellmann: only one force | 18:29 |
| clarkb | yup | 18:29 |
| dhellmann | cool | 18:29 |
| fungi | sounds right. so who wants to push the button of shame this time? shall i, or are you doing that now clarkb? | 18:30 |
| clarkb | fungi: I can do it unless you want to | 18:30 |
| *** dboik_ has joined #openstack-infra | 18:30 | |
| fungi | go for it | 18:30 |
| clarkb | (at this point I understand and am happy to) | 18:30 |
| clarkb | kk | 18:30 |
| clarkb | though one more question | 18:31 |
| clarkb | mriedem: dhellmann the zake pin is essentially doing what g-r does which we now recognize as broken | 18:31 |
| clarkb | kazoo too | 18:31 |
| dhellmann | harlowja_: ^^ | 18:32 |
| clarkb | should we be using <X.Y.Z instead of <=X.Y.Z-1 for those? | 18:32 |
| *** ashleighfarnham has quit IRC | 18:32 | |
| dhellmann | are we pinning or excluding? | 18:32 |
| *** ashleighfarnham has joined #openstack-infra | 18:32 | |
| mriedem | https://review.openstack.org/#/c/186084/3/requirements-py3.txt | 18:32 |
| dhellmann | capping | 18:32 |
| clarkb | dhellmann: ya but the cap is inclusive, which means if we need to release a point relase for fixes we have to do the dance again | 18:33 |
| fungi | jeblair: is christopheraedo who you were talking to about puppeting apps.openstack.org? | 18:33 |
| dhellmann | yeah | 18:33 |
| mriedem | i just copied what was in g-r, but yeah | 18:33 |
| mriedem | i could see zake<0.2 | 18:33 |
| dhellmann | mriedem: I think clarkb is right | 18:33 |
| *** BharatK has joined #openstack-infra | 18:33 | |
| clarkb | mriedem: oh if this is already in g-r then lets fix it after we are unwedged | 18:33 |
| *** dustins has joined #openstack-infra | 18:33 | |
| mriedem | clarkb: the tooz caps are the same as what they are in g-r on stable/juno | 18:34 |
| clarkb | as we can do that with normal process, if no one disagress I can force merge now | 18:34 |
| jeblair | fungi: i think so, mordred? ^ | 18:34 |
| mriedem | since stable/juno for tooz was just created today and isn't synced up | 18:34 |
| fungi | jeblair: mordred: also was the conversation just "thins would be a good next step" or do you know of any work already started in that direction? | 18:34 |
| mordred | jeblair: yes | 18:34 |
| *** kmartin has joined #openstack-infra | 18:34 | |
| clarkb | mriedem: dhellmann: that sound reasonable? | 18:34 |
| mordred | fungi: no work that we know of | 18:35 |
| dhellmann | clarkb: yep | 18:35 |
| fungi | k, thx | 18:35 |
| clarkb | (I want to make sure I don't conflate fixing the wedge with making things better overall) | 18:35 |
| *** maishsk_afk has joined #openstack-infra | 18:35 | |
| dhellmann | clarkb: agreed | 18:35 |
| clarkb | ok merging now | 18:35 |
| *** Somay has joined #openstack-infra | 18:36 | |
| clarkb | dhellmann: mriedem done | 18:37 |
| openstackgerrit | Anita Kuno proposed openstack-infra/irc-meetings: Correct an error in a meeting description. https://review.openstack.org/186075 | 18:37 |
| dhellmann | clarkb: thanks | 18:37 |
| mriedem | clarkb: double thanks | 18:38 |
| *** jtomasek has joined #openstack-infra | 18:40 | |
| *** maishsk_afk has quit IRC | 18:43 | |
| *** mattfarina has joined #openstack-infra | 18:43 | |
| *** doug-fish has joined #openstack-infra | 18:44 | |
| openstackgerrit | Matt Riedemann proposed openstack/requirements: Leave wiggle room for zake and kazoo fix releases https://review.openstack.org/186123 | 18:44 |
| mriedem | dhellmann: clarkb: fyi https://review.openstack.org/#/c/186123/ | 18:45 |
| *** maishsk_afk has joined #openstack-infra | 18:45 | |
| *** amotoki_ has joined #openstack-infra | 18:46 | |
| *** maishsk_afk is now known as maishsk | 18:46 | |
| jogo | https://github.com/osops | 18:46 |
| mtreinish | dhellmann: were you really eager to announce the new tooz release? :) | 18:47 |
| jogo | even being shared out by foundation staff | 18:47 |
| pleia2 | huh | 18:47 |
| *** achanda_ has quit IRC | 18:47 | |
| mriedem | mtreinish: that's automated :) | 18:47 |
| *** otter768 has joined #openstack-infra | 18:47 | |
| mordred | wow | 18:47 |
| *** sdake_ has joined #openstack-infra | 18:47 | |
| jogo | mtreinish: you should add your own words to the script | 18:47 |
| mtreinish | mriedem: heh, I know I'm just giving dhellmann a hard time :) | 18:47 |
| mriedem | still have to wait for it to hit the mirrors | 18:48 |
| mordred | hogepodge: heya | 18:48 |
| mriedem | 'sorry about burning your house down, here is a new release of x! | 18:48 |
| jogo | mordred: context http://lists.openstack.org/pipermail/openstack-operators/2015-May/007060.html | 18:48 |
| *** BharatK has quit IRC | 18:48 | |
| *** Sukhdev has joined #openstack-infra | 18:49 | |
| clarkb | jogo: are you on that list? you can respond with the elk stack we use | 18:50 |
| jeblair | clarkb: he already did! :) | 18:50 |
| clarkb | which includes filters for most openstack services | 18:50 |
| clarkb | awesome | 18:50 |
| clarkb | I see it now | 18:50 |
| fungi | yeah, the last time this same topic came up on the ops list i provided links to our logstash filters | 18:50 |
| clarkb | (maybe I should subscribe to that list too) | 18:50 |
| *** mattfarina has quit IRC | 18:50 | |
| *** HeOS has joined #openstack-infra | 18:50 | |
| pleia2 | Clint: wow, nice work digging up these ancient reviews (clearly it's been some time since I've made it that far back in my queue) | 18:50 |
| *** maishsk has quit IRC | 18:50 | |
| *** amotoki_ has quit IRC | 18:51 | |
| jeblair | clarkb: and the PR to the osops repo was to add a link to the stack in infra | 18:51 |
| fungi | a new thread seeking logstash filters seems to crop up about every 6 months on that ml | 18:51 |
| jogo | I was just about to say that myself | 18:51 |
| *** maishsk has joined #openstack-infra | 18:51 | |
| *** sdake has quit IRC | 18:51 | |
| jogo | fungi: but what I find interesting is github osops | 18:51 |
| Clint | pleia2: well, pabelanger was grexing | 18:51 |
| clarkb | speaking of github, I am (trying) to get this etherpad-lite fix merged and I have no idea why people like this workflow | 18:52 |
| *** otter768 has quit IRC | 18:52 | |
| pleia2 | Clint: rightly so :) | 18:52 |
| clarkb | grabbing the PR is really hard | 18:52 |
| fungi | jogo: some of our operator community seem allergic to doing things which might make them seem like developers, but apparently pushing their configs into a github repo isn't one of those things | 18:52 |
| clarkb | and even though a PR is an issue you can add a PR to an issue | 18:52 |
| pabelanger | Clint, pleia2 well, mostly because openstack-infra-packagers group does not exist yet! Hopefully somebody will get +2 support to help merge that stuff | 18:53 |
| pleia2 | I believe it does some magical linking if you mention the issue in the commit message | 18:53 |
| clarkb | thankfully I have sorted most of that out now and am only running into my lack of js knowledge | 18:53 |
| pleia2 | but I don't like the workflow either | 18:53 |
| clarkb | pleia2: it does, but PRs are issues | 18:53 |
| clarkb | pleia2: so why can't I just say attach this code to issue X when making a pull request | 18:53 |
| pleia2 | ah, right | 18:53 |
| *** gokrokve has quit IRC | 18:53 | |
| clarkb | in fact funny story, there are hacks in our PR closer bot because of this | 18:54 |
| clarkb | we disable issues, except there is no close PR api call bceause they are issues, you are supposed to close the issue instead | 18:54 |
| *** mattfarina has joined #openstack-infra | 18:54 | |
| * pleia2 nods | 18:55 | |
| clarkb | which confuses the python lib, this leads to getting the PR as a PR then pretending its an issue and closing it | 18:55 |
| * clarkb notes that the osops filters are already very similar to ours | 18:56 | |
| *** gokrokve has joined #openstack-infra | 18:56 | |
| openstackgerrit | George Peristerakis proposed openstack-infra/system-config: Moved the server class out of the status class https://review.openstack.org/186127 | 18:56 |
| *** dboik_ has quit IRC | 18:57 | |
| openstackgerrit | Matthew Treinish proposed openstack-infra/subunit2sql: Add release notes for recent releases https://review.openstack.org/183657 | 18:57 |
| *** dboik has joined #openstack-infra | 18:57 | |
| clarkb | fun thing I have noticed about this test suite, is they have flaky tests too :) sometimes 3 tests fail but usually 2 tests fail | 18:58 |
| *** e0ne has quit IRC | 18:59 | |
| *** bdemers has joined #openstack-infra | 19:00 | |
| *** achanda has joined #openstack-infra | 19:00 | |
| *** fawadkhaliq has joined #openstack-infra | 19:01 | |
| *** fawadkhaliq has quit IRC | 19:06 | |
| *** abregman has joined #openstack-infra | 19:06 | |
| *** sdake has joined #openstack-infra | 19:06 | |
| *** eharney has quit IRC | 19:08 | |
| dhellmann | mtreinish: have you ever seen me eager about anything? ;-) | 19:09 |
| *** sdake_ has quit IRC | 19:10 | |
| *** mattfarina has quit IRC | 19:12 | |
| *** fawadkhaliq has joined #openstack-infra | 19:14 | |
| *** eharney has joined #openstack-infra | 19:15 | |
| *** jtomasek has quit IRC | 19:16 | |
| *** gyee has joined #openstack-infra | 19:16 | |
| openstackgerrit | Monty Taylor proposed openstack-infra/nodepool: Consume os-client-config directly https://review.openstack.org/185972 | 19:17 |
| openstackgerrit | Monty Taylor proposed openstack-infra/nodepool: Debug dying task managers https://review.openstack.org/186079 | 19:17 |
| jeblair | dhellmann: i've seen you eager to _avoid_ things :) | 19:17 |
| mordred | clarkb: ok - I fixed some things ^^ but I'm back to not being able to see why the providerManager is going away | 19:17 |
| mordred | jeblair: I rebased on clark's patch to capture that, but there is still something going wrong that's not logging | 19:18 |
| mordred | jeblair: I'm not convinced that's a bug in nodepool per-se though - I _think_ it may be fake constructure | 19:18 |
| kashyap | Hi folks, maybe I'm missing something very trivial, `git review` fails at this step - `git rebase -p -i remotes/gerrit/master` when I try to submit a review from a local branch. | 19:18 |
| kashyap | Any hints what am I missing? http://paste.openstack.org/show/240517/ | 19:19 |
| clarkb | mordred: you should look at the provider config test | 19:19 |
| clarkb | mordred: that one should blow up if providers are still breaking (I would hope) | 19:19 |
| clarkb | kashyap: it means your change has a merge conflict and you should resolve it | 19:20 |
| *** yamamoto has joined #openstack-infra | 19:20 | |
| *** spredzy is now known as spredzy|afk | 19:20 | |
| clarkb | kashyap: if you don't want to resolve it yet (note that gerrit cannot merge if you push this way) you can `git review -R` to push it up without a rebase | 19:20 |
| *** nolafleur has quit IRC | 19:20 | |
| kashyap | clarkb: Oh, I could swear I did rebase it? But I came back to this commit after a while, let me check again. | 19:20 |
| *** kmartin has quit IRC | 19:20 | |
| jeblair | oh, maybe we should detect that case and have git-review say "your change has a merge conflict and you should resolve it, if you don't want to resolve it yet (note that gerrit cannot merge if you push this way) you can `git review -R` to push it up without a rebase" | 19:20 |
| *** jtomasek has joined #openstack-infra | 19:21 | |
| clarkb | jeblair: ya that may be a friendler message, and possibly git rebase --abort for the user too | 19:21 |
| kashyap | jeblair: Yeah, current error seems to come from - https://github.com/openstack-infra/git-review/blob/master/git_review/tests/test_git_review.py#L150 | 19:21 |
| mordred | clarkb: which test is the provider config test? | 19:21 |
| clarkb | mordred: nodepool.tests.test_shade_integration.TestShadeIntegration.test_nodepool_provider_config | 19:22 |
| mordred | oh - got it | 19:22 |
| mordred | that works | 19:22 |
| clarkb | so after all this fighting node.js I finally checked out HEAD^ and reran test, those fai ltoo | 19:22 |
| clarkb | so I think their tests are broken | 19:22 |
| jeblair | clarkb: ga | 19:22 |
| jeblair | clarkb: the abort does happen | 19:23 |
| clarkb | jeblair: hrm? | 19:24 |
| jeblair | oh, wait, i may not fully understand that yet. | 19:24 |
| clarkb | jeblair: sorry I am etherpad-lite debugging too | 19:24 |
| jeblair | clarkb: (i'm switching back to git-review. i'm going to implement my suggestion. i'll let you know when i have something.) | 19:24 |
| clarkb | because I want to get our fixed merged and I was asked to make sure their tests pass (they don't) and just now realized thats because master doesn't work | 19:25 |
| anteaya | mtreinish: this patch adds a neutron linuxbridge job to tempest's experimental queue, could you add it to your review queue please? https://review.openstack.org/#/c/184830/ | 19:25 |
| jeblair | clarkb: *nod* | 19:25 |
| clarkb | jeblair: oh you mean abort in git-review, I thought it didn't abort but maybe I need to upgrade locally | 19:25 |
| jeblair | clarkb: it sometimes does. i'm working out the threads now. | 19:25 |
| mtreinish | anteaya: sigh, yet another dsvm job on the overloaded tempest queue | 19:27 |
| clarkb | mordred: so that may have been one issue but not the only issue. Now to track down that no longer running traceback | 19:27 |
| mordred | yah | 19:27 |
| mtreinish | sc68cal: why not just do the sorta reckless thing and switch it over without telling anyone | 19:27 |
| mordred | clarkb: gut hunch is that we're not faking out the right things | 19:27 |
| *** ashleighfarnham has quit IRC | 19:27 | |
| mordred | clarkb: greghaynes figured it out last time - maybe he's got the magic touch | 19:28 |
| * mordred looks at greghaynes ... | 19:28 | |
| *** jsavak has quit IRC | 19:28 | |
| mtreinish | sc68cal: I would support that approach | 19:28 |
| clarkb | mordred: so that is raised if you call stop() on the manager | 19:28 |
| *** jsavak has joined #openstack-infra | 19:28 | |
| clarkb | mordred: which we would be calling around rebuilding the managers. So I think it is failing due to the inequalit checks not working right | 19:28 |
| mordred | oh! | 19:28 |
| anteaya | mtreinish: well I've been guiding him to the non-reckless approach | 19:29 |
| clarkb | mordred: we basically make new managers and the tests never have a manger long enough to effect change in the system | 19:29 |
| *** sarob_ has quit IRC | 19:29 | |
| mordred | clarkb: but if we're not faking that out appropriate ... let me add some prints ... | 19:29 |
| anteaya | mtreinish: what tempest queue would you prefer to see the job? it is non-voting | 19:29 |
| clarkb | mordred: I don't think we fake it at all | 19:29 |
| mordred | well, we need to now | 19:29 |
| clarkb | mordred: no | 19:30 |
| mordred | no? | 19:30 |
| clarkb | mordred: we need to make the inequality checks work so that we don't rebuild managers every pass through the loop | 19:30 |
| mordred | right | 19:30 |
| mordred | they work in os-client-config now | 19:30 |
| clarkb | mordred: and that shouldn't require faking | 19:30 |
| *** timcline has quit IRC | 19:30 | |
| *** ashleighfarnham has joined #openstack-infra | 19:30 | |
| mordred | what I'm saying is that I'm guessing somewhere now that we're actually directly instantiating that, that we're either faking too much or not enough things, such that the inequality checks are triggering | 19:30 |
| *** timcline has joined #openstack-infra | 19:31 | |
| *** mattfarina has joined #openstack-infra | 19:31 | |
| mtreinish | anteaya: it's on experimental which is the only one it makes sense. I'm just looking at the job count in all the queues and sighing | 19:31 |
| anteaya | mtreinish: I'm getting kind of tired of guiding someone in a process that is then circumvented/re-written due to overload | 19:31 |
| anteaya | mtreinish: ah yes, I definitely agree with your motivation for sighing | 19:31 |
| anteaya | and we again have the constant problem of it works for n things we we never imgaged that x^2 would want to use it | 19:32 |
| anteaya | n^2 | 19:32 |
| clarkb | mordred: I really think the fakes are sufficient | 19:33 |
| anteaya | mtreinish: I'm open to a discussion about how to solve the job count, if you want to have one | 19:33 |
| mordred | clarkb: I'm sure they are | 19:33 |
| *** abregman has quit IRC | 19:33 | |
| clarkb | mordred: we fake the clients themselves, the problem is the provider managers which are not faked are rebuilt each time | 19:33 |
| *** fawadkhaliq has quit IRC | 19:33 | |
| clarkb | mordred: and we shouldn't need to fake that out because its all internal process bits, no network or other service interaction there. | 19:33 |
| *** mattfarina has quit IRC | 19:34 | |
| mordred | well, other tahn the fact that it will search for env vars or system config files | 19:34 |
| mordred | but I don't think that's the problem | 19:34 |
| mordred | I'm starting from that inequality check and working outwards to see where things are falling over | 19:35 |
| electrical | Ola all im back. | 19:35 |
| mordred | clarkb: yes -that is showing up as unequal - hypothesis confirmed | 19:36 |
| mordred | now ... why | 19:36 |
| *** jgriffith has quit IRC | 19:36 | |
| mtreinish | anteaya: well in this case, I don't know why this job needs to be on tempest | 19:36 |
| openstackgerrit | James E. Blair proposed openstack-infra/git-review: Provide an explanation on failed rebase https://review.openstack.org/186139 | 19:36 |
| jeblair | clarkb, kashyap: ^ | 19:36 |
| *** jgriffith has joined #openstack-infra | 19:36 | |
| mtreinish | because running with a different driver hopefully shouldn't change neutron's behavior sufficiently to break a tempest run (assuming it's properly configured) | 19:37 |
| anteaya | mtreinish: okay well that would be a fix for now, if sc68cal agrees | 19:37 |
| anteaya | mtreinish: I can't assume such a thing with neutron | 19:37 |
| * kashyap clicks | 19:37 | |
| openstackgerrit | James E. Blair proposed openstack-infra/git-review: Provide an explanation on failed rebase https://review.openstack.org/186139 | 19:37 |
| anteaya | mtreinish: perhaps sc68cal can | 19:37 |
| jeblair | clarkb: forgot your co-authored-by ^ :) | 19:37 |
| clarkb | jeblair: ha | 19:37 |
| mtreinish | anteaya: oh, I know that's an incredibly naive statement I just made | 19:37 |
| mtreinish | because I've used neutron and worked with it before | 19:38 |
| mordred | Bad file descriptor (bundled/zeromq/src/epoll.cpp:81) | 19:38 |
| mordred | :) | 19:38 |
| mordred | clarkb: I got a new error ^^ ? | 19:38 |
| anteaya | mtreinish: but motivated by trying to reduce job count, I understand | 19:38 |
| clarkb | mordred: I have never seen that before | 19:38 |
| mordred | clarkb: I'm going to blame something else on that one | 19:38 |
| anteaya | mtreinish: yeah, so I guess I wait to hear from sc68cal and hear his thoughts | 19:38 |
| *** yamamoto has quit IRC | 19:38 | |
| mordred | clarkb: I'm also going to go with "maybe I didn't actually have the os-client-config patch installed like I thought" | 19:38 |
| *** yamamoto has joined #openstack-infra | 19:38 | |
| anteaya | since we do want neutron linuxbridge tested | 19:39 |
| clarkb | mordred: :) | 19:39 |
| anteaya | mtreinish: however that needs to happen | 19:39 |
| *** yamamoto has quit IRC | 19:40 | |
| mtreinish | anteaya: sure, but I just don't think running with linuxbridge on tempest commits adds anything. Neutron *should* be abstracting away the differences between backends | 19:41 |
| * mtreinish buries his head in sand | 19:41 | |
| mordred | mtreinish: well - except there are some differences and we're happy about them | 19:41 |
| *** timcline has quit IRC | 19:42 | |
| kashyap | I wish Gerrit allows one to correct grammatical mistakes after sending hit too soon on a comment. | 19:42 |
| mordred | mtreinish: like the difference between an install that needs floating-ips, for instance | 19:42 |
| *** timcline has joined #openstack-infra | 19:42 | |
| kashyap | hitting send too soon* ie. :-) | 19:42 |
| mordred | mtreinish: but I dont know that that would appear to tempest | 19:42 |
| mtreinish | mordred: no, I realize that. Look how many config options we have on tempest to try and map all the different neutron deploy configs out there | 19:42 |
| mordred | nod | 19:43 |
| clarkb | kashyap: I am actually a fan of immutable history there, its not always the cleanest ( and I make my fair share of mistakes ) but keeps everything honest and above board | 19:43 |
| clarkb | mordred: linux bridge would not affect need for floating ips | 19:43 |
| mordred | clarkb: zomg. I need to define __ne__ don't I? | 19:43 |
| clarkb | mordred: yes | 19:43 |
| mordred | *headdesk* | 19:43 |
| *** cody-somerville has joined #openstack-infra | 19:43 | |
| *** cody-somerville has quit IRC | 19:43 | |
| *** cody-somerville has joined #openstack-infra | 19:43 | |
| kashyap | clarkb: Yeah, I fully concur. It shows the human aspect too, as opposed to fixing everyone's OCD | 19:44 |
| clarkb | mordred: or use a "isEqual" metho | 19:44 |
| mtreinish | mordred: I'm just trying to figure out when we would need to run check experimental on a tempest commit to test a change against a linux bridge backed neutron | 19:44 |
| mordred | well, I've already got an __eq__ | 19:44 |
| clarkb | mordred: I personally am a fan of explicit methods simply because operator overloading is weird | 19:44 |
| clarkb | mordred: right and this is why its weird :) | 19:44 |
| mtreinish | and whether that edge case is justification enough for adding yet another job to the tempest queue | 19:44 |
| mordred | k. one sec | 19:44 |
| mordred | I mean, it'll be is_equal - but let's not quibble | 19:44 |
| pabelanger | woah, that took a few minutes! But local nodepool running again, now run with dibs | 19:46 |
| *** abregman has joined #openstack-infra | 19:46 | |
| clarkb | arg I really need to subscribe to this list, jogo maybe you can respond to the logstash thing and say pull requests welcom | 19:46 |
| clarkb | jogo: the only reason its greedy is because the data in the message is rather unstructured so being not greedy is hard | 19:46 |
| mordred | maybe point out that fixing the ones infra uses is important because they're the ones the dev use to find problems | 19:47 |
| mordred | so the more they're the same thing, the more the experience the operators have will be reflected back to the developers and the easier it will be to talk about the same problems | 19:47 |
| clarkb | and now I am subscribed | 19:48 |
| clarkb | mordred: +1000 | 19:48 |
| *** jgrimm is now known as zz_jgrimm | 19:49 | |
| openstackgerrit | Ryan Petrello proposed openstack-infra/project-config: Test pecan against barbican, magnum, and gnocchi. https://review.openstack.org/186141 | 19:49 |
| *** yamamoto has joined #openstack-infra | 19:50 | |
| jogo | clarkb: you want to respond as you are the expert? | 19:50 |
| clarkb | jogo: I can't because I only just subscribed | 19:50 |
| clarkb | otherwise you end up breaking the thread an all sorts of unfun | 19:51 |
| jogo | clarkb: ack, will respond | 19:51 |
| *** timcline has quit IRC | 19:51 | |
| clarkb | I am sure jeblair will say there is a way to get around that but my mail foo is such that if I attempt to I will likely just make a new dangling thread and it will be unfun | 19:52 |
| *** timcline has joined #openstack-infra | 19:52 | |
| jeblair | clarkb: easiest way is to ask someone to bounce the message to you | 19:52 |
| jeblair | clarkb: hardest (but best) way is what mordred wants to do -- set up an nntp server ;) | 19:52 |
| clarkb | also I have to pop out now or will be late for doctor | 19:52 |
| clarkb | hopefully back late this afternoon to check in on nodepool progress | 19:53 |
| jeblair | the doctor is usually late anyway (despite having a time machine) | 19:53 |
| jeblair | [it's also possible i'm confused about exactly what clarkb is doing when he goes to see 'the doctor'] | 19:54 |
| clarkb | :) | 19:54 |
| *** ociuhandu has quit IRC | 19:55 | |
| *** freerunner has joined #openstack-infra | 19:55 | |
| *** gokrokve has quit IRC | 19:55 | |
| jeblair | clarkb, fungi, mordred, marun: I merged https://review.openstack.org/161414 which is a potential "break everything on tomorrow's image builds" change; so if something goes wrong with tox jobs, we'll need to drop the new images. | 19:58 |
| fungi | jeblair: thanks for the heads up. i'll check when i wake up | 19:59 |
| jogo | clarkb: responded hopefully I didn't miss anything | 19:59 |
| *** freerunner has quit IRC | 19:59 | |
| *** mestery_ has joined #openstack-infra | 20:00 | |
| *** kmartin has joined #openstack-infra | 20:00 | |
| *** olaph is now known as olaf | 20:00 | |
| *** olaf is now known as Guest29419 | 20:00 | |
| *** Guest29419 is now known as olaph | 20:00 | |
| *** cody-somerville has quit IRC | 20:00 | |
| *** freerunner has joined #openstack-infra | 20:00 | |
| *** ajmiller_ is now known as ajmiller | 20:01 | |
| dhellmann | jeblair: you've seen to the truth of the matter regarding my eagerness | 20:01 |
| *** erikmwilson_ has joined #openstack-infra | 20:01 | |
| *** Rockyg has joined #openstack-infra | 20:01 | |
| *** jtomasek has quit IRC | 20:01 | |
| *** ashleighfarnham has quit IRC | 20:02 | |
| *** mestery has quit IRC | 20:03 | |
| *** tonytan4ever has quit IRC | 20:05 | |
| nikeshm | hi | 20:05 |
| nikeshm | i am using trusty-server-cloudimg-amd64-root.tar.gz image in nodepool | 20:06 |
| nikeshm | but nodepool vm is not having RAM greater than 6GB | 20:06 |
| nikeshm | is there any kernel-limit in trusty-server-cloudimg-amd64-root.tar.gz image | 20:07 |
| *** sdake_ has joined #openstack-infra | 20:07 | |
| nikeshm | with cirros image i am able to get RAM of any size | 20:07 |
| nikeshm | but with this nodepool image i am getting RAM of only 6GB | 20:08 |
| *** kgiusti has left #openstack-infra | 20:08 | |
| nikeshm | i tried many flavors with nodepool image | 20:08 |
| nikeshm | but its not taking RAM more than 6GB | 20:08 |
| *** jtomasek has joined #openstack-infra | 20:08 | |
| *** sdake has quit IRC | 20:11 | |
| *** tonytan4ever has joined #openstack-infra | 20:11 | |
| *** alexsyip has joined #openstack-infra | 20:12 | |
| *** achanda has quit IRC | 20:14 | |
| *** tiswanso has quit IRC | 20:14 | |
| *** tiswanso has joined #openstack-infra | 20:15 | |
| fungi | nikeshm: nodepool tells nova to boot instances from the image. in theory you should be able to nova boot that image and confirm whether or not your problem has anything to do with nodepool (i'm pretty sure it does not, but would need to know a lot more about how nodepool is configured to update your images to say that for certain) | 20:16 |
| asselin | nikeshm, try booting the base image in nova and see if that works as expected | 20:18 |
| *** esker has quit IRC | 20:19 | |
| asselin | nikeshm, actually if I remember right, nodepool dib picks the image for you...."release: trusty" http://paste.openstack.org/show/240583/ | 20:20 |
| *** ashleighfarnham has joined #openstack-infra | 20:20 | |
| *** slagle_ is now known as slagle | 20:20 | |
| nikeshm | asselin:right | 20:21 |
| openstackgerrit | Matthew Treinish proposed openstack-infra/subunit2sql: Add attachemnts table to database https://review.openstack.org/186151 | 20:21 |
| openstackgerrit | Matthew Treinish proposed openstack-infra/subunit2sql: Add support to subunit2sql for storing attachments https://review.openstack.org/186152 | 20:21 |
| mtreinish | pcrews: ^^^ I figured since your job change you dropped that, so I hacked something together. I still need to test it though | 20:22 |
| *** mrmartin has quit IRC | 20:26 | |
| *** julim has quit IRC | 20:34 | |
| *** amotoki_ has joined #openstack-infra | 20:34 | |
| *** akrivoka has quit IRC | 20:35 | |
| *** annegentle has joined #openstack-infra | 20:37 | |
| *** claudiub has quit IRC | 20:38 | |
| openstackgerrit | Anita Kuno proposed openstack-infra/irc-meetings: Remove on hold meeting https://review.openstack.org/186158 | 20:39 |
| *** amotoki_ has quit IRC | 20:39 | |
| *** tiswanso has quit IRC | 20:40 | |
| openstackgerrit | Sylvain Bauza proposed openstack-infra/irc-meetings: Rename Gantt meeting to Nova Scheduler https://review.openstack.org/186159 | 20:41 |
| *** stevemar has quit IRC | 20:43 | |
| *** dizquierdo has joined #openstack-infra | 20:44 | |
| *** ldnunes has quit IRC | 20:44 | |
| Somay | Hello everyone, I just started out as a summer intern with OpenDaylight on the Jenkins Job Builder project. I would be really grateful if someone reviewed the following patch - https://review.openstack.org/#/c/185445/ | 20:45 |
| openstackgerrit | Sridhar Ramaswamy proposed openstack-infra/project-config: Add tacker-horizon repo for Tacker VNF Manager UI https://review.openstack.org/186161 | 20:45 |
| *** esker has joined #openstack-infra | 20:46 | |
| anteaya | hi Somay I hope your internship is going well | 20:47 |
| anteaya | Somay: if you go into your gerrit ui and you are signed in, navigate to people > list groups and search for jenkins-job-builder-core | 20:49 |
| *** Sukhdev has quit IRC | 20:49 | |
| anteaya | Somay: while others can review your patch too, those are the folks you should spend some time getting to know | 20:49 |
| *** bswartz has quit IRC | 20:50 | |
| openstackgerrit | Merged openstack/requirements: Raise cap on tooz to <0.13 https://review.openstack.org/186065 | 20:51 |
| Somay | anteaya: Thanks, I’ll have a look there | 20:52 |
| anteaya | Somay: welcome | 20:52 |
| *** TravT_ has joined #openstack-infra | 20:53 | |
| electrical | Somay: i ave a question about your patch. Any specific reason to do negative comparisons? ( the not parts ) | 20:55 |
| openstackgerrit | Sumit Naiksatam proposed openstack-infra/irc-meetings: Remove Neutron Adv Services meeting https://review.openstack.org/186163 | 20:56 |
| electrical | just something I'm wondering because i personally try to avoid them :-) | 20:56 |
| Somay | electrical: Yes, the checkboxes in the Jenkins UI are stated as “Disable automatic site…” | 20:56 |
| *** baoli has quit IRC | 20:56 | |
| *** sridhar_ram has joined #openstack-infra | 20:56 | |
| Somay | and in JJB, I wished to keep the description as to enable the feature | 20:57 |
| electrical | Ah okay i see. Any specific reason for not calling them disable-automatic-site-archiving ? | 20:58 |
| Somay | Disable sounds good with a checkbox, but for yaml, we should have keywords which describe the feature, hence the negaton | 20:58 |
| Somay | I thought automatic-site-archiving is a better keyword | 20:58 |
| openstackgerrit | Sumit Naiksatam proposed openstack-infra/irc-meetings: Remove Neutron Adv Services meeting https://review.openstack.org/186163 | 20:58 |
| Somay | disable-automatic-site-archiving would be less user friendly IMHO | 20:59 |
| electrical | good point | 20:59 |
| Somay | and by the same argument, we should then have enable-disable for all keywords, which is not nice | 20:59 |
| electrical | just wanted to understand the reasoning behind it but you obviously thought about it :-) | 21:00 |
| Somay | yes, not an issue :) | 21:00 |
| *** jerryz has joined #openstack-infra | 21:00 | |
| *** soren_ has joined #openstack-infra | 21:01 | |
| *** ociuhandu has joined #openstack-infra | 21:01 | |
| electrical | I've seen developers do weird thing because they were stuck in a certain thought pattern ;-) | 21:01 |
| *** maishsk_ has joined #openstack-infra | 21:01 | |
| openstackgerrit | Sumit Naiksatam proposed openstack-infra/irc-meetings: Fix typo in name https://review.openstack.org/186167 | 21:01 |
| *** esker has quit IRC | 21:01 | |
| Somay | electrical: yes, that happens all the time. I agree with you | 21:02 |
| *** ociuhandu has quit IRC | 21:02 | |
| *** dustins has quit IRC | 21:02 | |
| *** ildikov has quit IRC | 21:03 | |
| *** maishsk has quit IRC | 21:03 | |
| *** maishsk_ is now known as maishsk | 21:03 | |
| *** sdake has joined #openstack-infra | 21:04 | |
| *** ociuhandu has joined #openstack-infra | 21:04 | |
| openstackgerrit | Jeremy Stanley proposed openstack-infra/project-config: Publish git-review documentation https://review.openstack.org/186168 | 21:05 |
| zxiiro | Any JJB reviewers available to review this patch, just needs one more +2 and we'd like to have it in sooner rather than later? https://review.openstack.org/#/c/185445/1 | 21:05 |
| zxiiro | clarkb: ^ ? | 21:06 |
| *** jsavak has quit IRC | 21:06 | |
| zxiiro | clarkb: woops sorry i copyed the wrong patch number i mean https://review.openstack.org/#/c/182116/ | 21:07 |
| openstackgerrit | Sumit Naiksatam proposed openstack-infra/irc-meetings: Fix GBP meeting details https://review.openstack.org/186171 | 21:07 |
| *** btully has quit IRC | 21:07 | |
| nikeshm | asselin: tried with trusty-server-cloudimg-amd64-disk1.img its working good with this,but nodepool is trying with trusty-server-cloudimg-amd64-root.tar.gz' | 21:08 |
| * zxiiro has had a long day :) | 21:08 | |
| *** sdake_ has quit IRC | 21:08 | |
| *** btully has joined #openstack-infra | 21:09 | |
| *** notnownikki has quit IRC | 21:10 | |
| *** samueldmq has quit IRC | 21:11 | |
| *** miqui has quit IRC | 21:11 | |
| *** mestery_ is now known as mestery | 21:12 | |
| electrical | I'm patiently waiting for 2 reviews as well :-) | 21:13 |
| *** dtantsur|afk has quit IRC | 21:13 | |
| * electrical thinks its time to implement a 'review friday' :p | 21:13 | |
| zxiiro | electrical: I usually patiently wait too but this particular patch we need sooner rather than later. I can't wait the usual 2-3 months that my patches usually take to have someone look at it... | 21:14 |
| *** notnownikki has joined #openstack-infra | 21:14 | |
| electrical | zxiiro: yeah indeed. | 21:14 |
| anteaya | electrical zxiiro Somay you have all reviewed each other's patches, yeah? | 21:15 |
| *** dizquierdo has quit IRC | 21:16 | |
| *** dkranz has quit IRC | 21:16 | |
| zxiiro | anteaya: I've reviewed Somay's patch since I'm his mentor :) | 21:16 |
| electrical | anteaya: i did review it but haven't left my +1 there yet. | 21:16 |
| anteaya | zxiiro: great | 21:16 |
| mordred | clarkb: ok - I just released os-client-config with the fixes needed for nodepool | 21:16 |
| openstackgerrit | Davide Guerri proposed openstack-infra/shade: Add port resource methods https://review.openstack.org/179968 | 21:16 |
| *** mwagner has quit IRC | 21:16 | |
| anteaya | electrical: wonderful, it helps to leave a comment that at least says "I've looked at this." | 21:16 |
| openstackgerrit | Monty Taylor proposed openstack-infra/nodepool: Consume os-client-config directly https://review.openstack.org/185972 | 21:17 |
| electrical | anteaya: will do | 21:17 |
| *** jtomasek has quit IRC | 21:17 | |
| mordred | clarkb: and I marked a version dep on the new version | 21:17 |
| anteaya | electrical: thank you | 21:17 |
| zxiiro | anteaya: we plan to provide many more patches over the summer | 21:18 |
| anteaya | zxiiro: awesome | 21:18 |
| *** bswartz has joined #openstack-infra | 21:18 | |
| zxiiro | I just wish there was a way to make a specific patch as something we need soon. most of the time I don't care how long it takes but for this particular one it's blocking us from some work at LF | 21:19 |
| anteaya | zxiiro: if you and Somay could also make a point of reviewing all the open jjb patches in gerrit it is possible one or both of you might be in a position to become a core reviewer on jjb, many considerations pending | 21:19 |
| anteaya | but it would speed up the time your patches spend in review | 21:19 |
| zxiiro | anteaya: I do do the occassional patch review and plan to increase my time as JJB si something we need | 21:20 |
| electrical | anteaya: i've been doing the same lately ( reviewing open changes ) | 21:20 |
| anteaya | zxiiro: the best way I know is to become a proficient reviewer | 21:20 |
| anteaya | zxiiro electrical awesome | 21:20 |
| anteaya | that really helps the rest of the jjb review team | 21:20 |
| waynr | anteaya: i'm interested in becomming a core reviewer on JJB although i tend to only review core library changes rather than module changes | 21:20 |
| anteaya | waynr: great | 21:20 |
| anteaya | waynr: I think a familiarity with the entire repo will help | 21:21 |
| anteaya | expertise in a certain area is great | 21:21 |
| anteaya | but it is good to have some of each | 21:21 |
| *** ildikov has joined #openstack-infra | 21:21 | |
| waynr | i am quite familiar with the entire repo, having implemented both yamlparser features and module features | 21:21 |
| anteaya | wonderful | 21:22 |
| zxiiro | anteaya: FWIW I've been contributing since December. Trying to get more active but it's hard to stay engaged when you feel ignored. | 21:22 |
| anteaya | electrical zxiiro Somay have you all reviewed waynr's patches too? | 21:22 |
| anteaya | zxiiro: I hear that | 21:22 |
| anteaya | zxiiro: this is why I suggest helping each other | 21:22 |
| anteaya | finding peers who have the same interest is important | 21:22 |
| electrical | anteaya: partially. We worked together on a weird bug that waynr in the end managed to solve :-) | 21:22 |
| anteaya | waiting for cores is often frustrating | 21:22 |
| zxiiro | Our priority has upped too since we hear JJB is becoming removed from OS's priorities from this past summit | 21:23 |
| anteaya | electrical: wonderful | 21:23 |
| zxiiro | I worry that any core reviewers are even still interested in reviewing and maintaining JJB if they are removing it from OpenStack's infra | 21:23 |
| anteaya | zxiiro: ahhhh, let's hear from jeblair on that as I am not feeling it is being removed as much as repurposed | 21:23 |
| electrical | anteaya: i'm working on getting JJB implemented at current job ( elastic ) so i got good reasons to get things reviewed faster ;-) hehe | 21:23 |
| anteaya | but it is possible I am incorrect | 21:23 |
| anteaya | electrical: sounds good | 21:24 |
| anteaya | our job > zuul workflow is definitely changing | 21:24 |
| * anteaya digs up the etherpad | 21:24 | |
| zxiiro | yeah that's what I got from one of the fishbowl sessions at the summit last week | 21:24 |
| anteaya | so here is the etherpad: https://etherpad.openstack.org/p/YVR-infra-zuulv3 | 21:25 |
| anteaya | zxiiro: you were at summit? | 21:25 |
| anteaya | did we meet? | 21:25 |
| mordred | yah - so we will be moving away from at at $future_date - however, enough people are happily using it that I'd expect we can make more people core on it and have it have a happy life | 21:25 |
| zxiiro | anteaya: indeed i was. not sure, i met zaro though | 21:25 |
| anteaya | zxiiro: ah glad you met zaro | 21:25 |
| zxiiro | i'll be at the Jenkins Conference in June too. | 21:26 |
| mordred | we certainly don't want to make life hard for people who love jjb - we think it's great too | 21:26 |
| anteaya | line 14 Stop using JJB in favor of ansible playbooks | 21:26 |
| zxiiro | mordred: that's good to know | 21:26 |
| waynr | interesting, how do ansible playbooks replace JJB? | 21:27 |
| anteaya | so yes according to the etherpad the intent is to move away from jjb and implement ansbile playbooks for jobs | 21:27 |
| anteaya | creating the config for a job | 21:27 |
| mordred | waynr: its all part of teh zuul v3 work | 21:27 |
| *** dannywil_ has quit IRC | 21:27 | |
| waynr | are there more details on the etherpad? | 21:27 |
| *** bdemers has quit IRC | 21:27 | |
| mordred | waynr: on infra's side, we basically use jenkins for 2 things currently: 1) log uploading 2) remote execution of shells | 21:27 |
| mordred | we don't use any of its other features | 21:27 |
| anteaya | so yes, stepping up involvement on jjb so that you can continue to use it for your work is a good idea | 21:27 |
| *** dannywilson has joined #openstack-infra | 21:27 | |
| *** peristeri has quit IRC | 21:27 | |
| mordred | and we're _very_ close to not using its log/artifact uploading either | 21:27 |
| mordred | waynr: let me find you the zuulv3 spec | 21:28 |
| waynr | oh is it in infra-specs? | 21:28 |
| anteaya | https://review.openstack.org/#/c/164371/3/specs/zuulv3.rst | 21:28 |
| mordred | waynr: https://review.openstack.org/#/c/164371/ | 21:28 |
| waynr | i have that open in a tmux session somewhere | 21:28 |
| mordred | yah | 21:28 |
| anteaya | line 334 | 21:28 |
| anteaya | 335 | 21:28 |
| mordred | but read the whole thing - it gives context on the why and everything | 21:28 |
| mordred | I do want to be clear that I think jenkins is an awesome technology and meets the needs of many people | 21:29 |
| waynr | okay this seems very important, as i have been working on getting zuul to work for $dayjob | 21:29 |
| mordred | we have just happened to hit the sscaling inflection points where that is not true for us any longer | 21:29 |
| waynr | (using github) | 21:29 |
| mordred | waynr: yes. I'm very excited about your github work | 21:29 |
| * mordred has evil plans for it | 21:29 | |
| anteaya | so waynr zxiiro electrical your goals to be more active in jjb are great, and it would serve all your purposes to get to know each other well and to review each other's patches regularly | 21:29 |
| electrical | anteaya: that's on my schedule for tomorrow afternoon :-) ( reviews ) | 21:30 |
| electrical | for now its bed time for me though :-) | 21:30 |
| waynr | anteaya: okay, I am reviewing Somay's patch now | 21:30 |
| anteaya | electrical: understood | 21:30 |
| anteaya | waynr: thank you | 21:30 |
| *** samueldmq has joined #openstack-infra | 21:30 | |
| electrical | Take care all. catch you all online tomorrow again | 21:30 |
| mordred | electrical: sleep well | 21:31 |
| waynr | it's sad that i can be more active on JJB on my vacation (right now) than i can while at $dayjob because i manage CI pipelines for several dozen projects using JJB | 21:31 |
| *** akshai has quit IRC | 21:31 | |
| anteaya | electrical: https://www.youtube.com/watch?v=Rp317mfo3Qg make reviewing a daily habit | 21:31 |
| openstackgerrit | Jeremy Stanley proposed openstack-infra/project-config: Publish git-review documentation https://review.openstack.org/186168 | 21:31 |
| zxiiro | anteaya: i think the reviews statitics is broken http://stackalytics.com/?metric=commits&user_id=thanh.ha@linuxfoundation.org&project_type=all&release=all | 21:31 |
| * anteaya feels bad she hasn't been true to that herself lately | 21:31 | |
| zxiiro | anteaya: it seems tho think i have no reviews when I know i've reviewed a few in the past | 21:32 |
| anteaya | zxiiro: it has you listed as an independent | 21:32 |
| anteaya | if you are Thanh Ha | 21:32 |
| openstackgerrit | Dawid Malinowski proposed openstack-infra/jenkins-job-builder: Extend support for join-trigger https://review.openstack.org/186176 | 21:33 |
| zxiiro | yeah i should probably submit a patch to stackalytics to set that right | 21:33 |
| anteaya | so for whatever reason stackalytics is confused about who you work with | 21:33 |
| anteaya | zxiiro: that would be my suggested course of action | 21:33 |
| Somay | Sorry, I wasn’t around. I read the whole conversation. Yes, we should review each other’s patches. It’s quite new to me that JJB is being removed from the priorities. Also, I’m quite new to JJB, so I’ll try to review as many patches as I can. | 21:34 |
| Somay | anteaya: thanks for the suggestion | 21:34 |
| anteaya | Somay: welcome | 21:35 |
| anteaya | Somay: you might find some value here, when getting started reviewing a patch: http://anteaya.info/blog/2013/03/21/reviewing-an-openstack-patch/ | 21:36 |
| Somay | Sure, i’ll read this | 21:38 |
| anteaya | I hope it is helpful | 21:39 |
| *** dpaterson has quit IRC | 21:39 | |
| *** dtantsur has joined #openstack-infra | 21:40 | |
| *** gordc has quit IRC | 21:41 | |
| openstackgerrit | Davide Guerri proposed openstack-infra/shade: Add port resource methods https://review.openstack.org/179968 | 21:41 |
| zxiiro | anteaya: done https://review.openstack.org/186178 | 21:41 |
| *** asahlin is now known as asahlin_afk | 21:42 | |
| zxiiro | are there infra meetings i should be aware of / attend? | 21:42 |
| Clint | every tuesday | 21:42 |
| openstackgerrit | Mateusz Matuszkowiak proposed openstack-infra/project-config: Added new SF repository for Calamari Fuel plugin https://review.openstack.org/186179 | 21:43 |
| zxiiro | clarkb: noted... i assume this one? https://wiki.openstack.org/wiki/Meetings/InfraTeamMeeting | 21:43 |
| *** dtantsur has quit IRC | 21:45 | |
| jeblair | electrical, waynr, zxiiro: yeah, i have a todo item to respond to the mailing list thread about the proposed jjb rewrite. the short version is that even if infra stops using jjb, i want to make sure the project continues. we will not pull support from jjb; instead we will work out the best way for jjb to continue to live beyond our use of it. because of the summit i have been too busy to give that the attention it deserves, but hope to d | 21:45 |
| anteaya | zxiiro: that is the one | 21:45 |
| openstackgerrit | Doug Hellmann proposed openstack-infra/yaml2ical: Allow extra settings for meetings to be used in templates https://review.openstack.org/186183 | 21:45 |
| jeblair | electrical, waynr, zxiiro: in other words, i don't think there's a hurry; we have time to work out a good transition plan | 21:46 |
| zxiiro | jeblair: thanks, in hindsight i probably should have introduced myself to you during the summit. Maybe next time :) | 21:46 |
| jeblair | zxiiro: see you in tokyo? ;) | 21:47 |
| zxiiro | jeblair: maybe, I'm also attending the Jenkins Conference next month if you're attending that | 21:47 |
| zxiiro | jeblair: I think you were the guy leading the fishbowl sessions? | 21:47 |
| *** ddieterly has quit IRC | 21:48 | |
| anteaya | zxiiro: is hard to miss, follow the bow tie | 21:48 |
| anteaya | jeblair is hard to miss | 21:48 |
| jeblair | zxiiro: some of them, yes. i also had a hat. | 21:48 |
| zxiiro | cool yeah i'm pretty sure I know who you are | 21:49 |
| jeblair | uh oh :) | 21:49 |
| *** mrmartin has joined #openstack-infra | 21:49 | |
| openstackgerrit | Doug Hellmann proposed openstack-infra/irc-meetings: Add meeting_id and agenda_url support https://review.openstack.org/186184 | 21:49 |
| jeblair | anteaya: ^ | 21:49 |
| *** dboik has quit IRC | 21:49 | |
| *** mriedem has quit IRC | 21:51 | |
| lifeless | is the requirements wedge fixed, or can I help? | 21:51 |
| anteaya | jeblair: doug's changes? | 21:51 |
| *** pc_m has quit IRC | 21:52 | |
| jeblair | anteaya: yeah, i think it may implement a feature you implicitly requested :) | 21:52 |
| waynr | jeblair: i am looking at the zuul3 spec right now, it looks very good! actually this is exactly the direction i was hoping zuul would go eventually (moving away from a reliance on jenkins as a job runner) | 21:52 |
| clarkb | I am not here but can we make sure packaging + delorean doesnt reinvent zuul | 21:52 |
| clarkb | because it sounds like they are reinventing zuul | 21:52 |
| fungi | lifeless: i think it's settled since the tooz patch and release... at least i haven't heard any further screaming but you might check in #openstack-stable | 21:52 |
| *** dizquierdo has joined #openstack-infra | 21:52 | |
| anteaya | jeblair: it does, I'll redo mine as a dependency of doug's patch | 21:53 |
| fungi | clarkb: i may have only skimmed... seemed like they said it grokked zuul refs and so could be run on our job workers | 21:53 |
| jeblair | waynr: cool, glad you like it. i think i have one or two things to tweak, but i think it's pretty close in principle to being ready to get started on. still time to leave feedback if you have some. | 21:53 |
| clarkb | no they have to add zuul support and it does its own triggering | 21:53 |
| clarkb | so no zuul at all you rundelorean | 21:53 |
| fungi | oh. ew | 21:53 |
| *** rguillebert has joined #openstack-infra | 21:53 | |
| fungi | derekh_: ^ ? | 21:54 |
| clarkb | at least thats how I read | 21:54 |
| *** dannywilson has quit IRC | 21:54 | |
| clarkb | "it monitors git repositories" | 21:54 |
| waynr | jeblair: yeah i might not be able to leave feedback until next week but i am somewhat concerned since $dayjob may need a transition period between jenkins and non-jenkins architecture if i can convince my team that zuul is the tool for us | 21:54 |
| *** dannywilson has joined #openstack-infra | 21:55 | |
| waynr | (concerned about backwards compatibility with the jjb/jenkins world that is) | 21:55 |
| *** mattfarina has joined #openstack-infra | 21:55 | |
| *** mjturek1 has quit IRC | 21:55 | |
| fungi | waynr: at least for now, we have a jenkins plugin which allows it to speak gearman, and use that to bridge zuul to jenkins. keep in mind that jenkins-only to jenkins+zuul is a transition we've already undergone | 21:56 |
| clarkb | also we dont need anew namespace or distro specific repos | 21:56 |
| * clarkb stops readinf email and goes back to hospitalling | 21:57 | |
| *** dtantsur has joined #openstack-infra | 21:57 | |
| jeblair | waynr: yeah, we've made a lot of transitions in a stepwise manner and supporting backwards compat (as fungi described), but for this one, i think it will be very hard to do that, so i am proposing a break. we will probably switch all our jobs at once. we will probably write a tool to do the bulk of the conversion automatically, but i don't know that it will end up being generalizable. | 21:57 |
| fungi | clarkb: right. derekh_ suggested asking infra, so i was going to wait for them to do that bit | 21:57 |
| clarkb | fungi well they have asked... anyways hard to be on irc, back later | 21:57 |
| jeblair | waynr: i think it will be possible/easy to switch projects one zuul shared-change-queue at a time | 21:57 |
| waynr | yeah it probably won't be generalizable given what little i know of openstack's JJB usage vs my own usage | 21:58 |
| jeblair | waynr: we might take advantage of that to switch infra first before openstack | 21:58 |
| waynr | (your auto conversion tool that is) | 21:58 |
| jeblair | that might work for you depending on how segmented your configuration is | 21:58 |
| fungi | clarkb: best of luck! stop looking at openstack things and focus on hospital | 21:58 |
| *** mwagner has joined #openstack-infra | 21:58 | |
| *** mrmartin has quit IRC | 21:58 | |
| *** dangers is now known as dangers_away | 21:58 | |
| waynr | i was actually hoping to convert all my JJB yaml to JJB python code using classes that could be used later to target something other than JJB for job definitions | 21:59 |
| anteaya | clarkb: what fungi said | 21:59 |
| derekh_ | clarkb: fungi yup its not currently ZUUL aware i.e. it currently only builds after a commit has gone into the branch its tracking, but I'm open to adding support for reading ZUUL REFs if it looks like we're going in that direction | 21:59 |
| jeblair | also, if someone comes up with a clever way to make the transition easier, i'm all for entertaining it. i just sensed that for this instance, it could greatly extend the implementation time and we would be able to more easily absorb a cutover rather than extended transition. | 21:59 |
| zxiiro | jeblair: when Zuulv3 is out would we still be able to use Jenkins along with it if we wanted to? or would support for it be dropped completely? | 22:00 |
| waynr | yeah i don't think my approach would help you necessarily, but I think that using python to define job configurations is something i want to continue doing beyond a JJB/jenkins world--i also considered implementing a zuul layout generator using these same classes | 22:00 |
| zxiiro | as far as I'm aware it's gearman that talks to Jenkins so that shouldn't be a problem no? | 22:01 |
| fungi | derekh_: yeah, if delorean could just be reduced into something which can run on demand on a job worker and let us fire it from zuul and pass in the usual refs and have it understand those, i think it ought to work fine. though at that point it seems like it's just a standardized set of package building instructions | 22:01 |
| anteaya | https://review.openstack.org/#/c/186184/1 failed with no logs | 22:01 |
| *** gokrokve has joined #openstack-infra | 22:02 | |
| waynr | zxiiro: in my understanding there are some important parameters passed from zuul to jenkins that tell jenkins where to get the git ref it uses for cloning the project being tested | 22:02 |
| jeblair | waynr: one thing i want to do in particular is try to reduce the repetitive constructs that made jjb necessary in the first place. we might fail to achieve perfection there and need something like what you describe :) | 22:02 |
| *** akshai has joined #openstack-infra | 22:03 | |
| *** mriedem has joined #openstack-infra | 22:03 | |
| anteaya | fungi: sorry fungi I rechecked your recheck | 22:03 |
| *** hrubi_ has quit IRC | 22:03 | |
| fungi | anteaya: zuul will ignore it--no worries | 22:03 |
| anteaya | k | 22:04 |
| *** dtantsur has quit IRC | 22:04 | |
| derekh_ | fungi: yes we should be able todo that, I don't see any reason it wouldn't work and your right that would just be a set of build instructions, but its only half of what delorean does, the other half is keeping track of builds for all of the projects | 22:04 |
| *** Sukhdev has joined #openstack-infra | 22:04 | |
| fungi | waynr: in fact, it's just adding those as parameters in jenkins which show up in the calling environment on the worker, and so it's really passing that to whatever your job payload/script is, just via jenkins | 22:04 |
| derekh_ | fungi: so that after each commit we have a repository containing all of the most recent successfully built packages | 22:05 |
| *** amotoki_ has joined #openstack-infra | 22:05 | |
| fungi | anteaya: https://jenkins03.openstack.org/job/gate-irc-meetings-tox-ical/17/console | 22:05 |
| jeblair | zxiiro: i believe it will theoretically be possible, but i think it would require a bit more work. my intention is for us to use ansible in openstack, and to design the system to support that use case, but also that part should be pluggable. i think that with some changes to the gearman-plugin, it could be done. | 22:05 |
| derekh_ | fungi: in other words you end up with a "current" repo that contains current openstack http://trunk.rdoproject.org/centos70/current/ | 22:05 |
| fungi | derekh_: got it. so customized publisher too | 22:06 |
| derekh_ | fungi: yup | 22:06 |
| *** bknudson has quit IRC | 22:06 | |
| waynr | jeblair: yeah that's my goal also--reduce repitition. i've found so often while developing templates and job groups using JJB yaml that what i really want is a turing complete language to allow me to achieve a greater degree of configuration efficiency using conditionals, subclasses, and more natural parameterization constructs | 22:06 |
| anteaya | fungi: is that a failure to connect to swift? | 22:07 |
| zxiiro | jeblair: guess we'll have to see, i discussed this a bit with zaro during the summit as well. I hope it can be done since we at LF do not yet have any plans to add Ansible to our setup but we do want to give zuul a try. | 22:07 |
| openstackgerrit | Davide Guerri proposed openstack-infra/shade: Add port resource methods https://review.openstack.org/179968 | 22:07 |
| fungi | anteaya: nope, http://logs.openstack.org/84/186184/1/check/gate-irc-meetings-tox-ical/ccc4920/console.html is working | 22:07 |
| fungi | anteaya: i think something's wrong with the swift directory index generation. maybe this is more widespread. checking now | 22:08 |
| jeblair | waynr: let's not put this in the critical path of zuulv3, but i could imagine making part of zuul pluggable so that instead of loading job definitions from yaml, it hands off to something else to produce the needed output. that could invoke the sort of thing you are describing. | 22:08 |
| fungi | also possible it's something gone wrong with serving directory indexes from os-loganalyze | 22:08 |
| jeblair | waynr: i think if we keep that in the back of our minds, we can probably avoid designing something that would break that, and then add it later. | 22:09 |
| dougwig | hi infra, can i get some project-config eyeballs on this experimental lbaas job? it's for moving the tempest tests out of tempest, per qa's request. https://review.openstack.org/#/c/184870/ | 22:09 |
| *** dtantsur has joined #openstack-infra | 22:09 | |
| zxiiro | gotta step out for dinner. If any JJB core reviewers has some time to review this patch I would be very appreciative ;) https://review.openstack.org/#/c/182116/ | 22:10 |
| *** hrubi has joined #openstack-infra | 22:10 | |
| *** amotoki_ has quit IRC | 22:10 | |
| waynr | jeblair: sounds good! i'll keep an eye on the spec and the upcoming zuul patches | 22:10 |
| jeblair | waynr: i'm anticipating this is a full 6-month cycle project. other people tell me they think it will take longer. :) | 22:11 |
| jeblair | (i don't believe them) | 22:11 |
| *** Rockyg has quit IRC | 22:11 | |
| mordred | jeblair: I've got it specced at 2 months in my head | 22:12 |
| openstackgerrit | Merged openstack/requirements: Add glance_store to projects.txt https://review.openstack.org/182674 | 22:12 |
| zaro | zxiiro: slow now because pelix is out for a week. | 22:12 |
| fungi | anteaya: aha! | 22:14 |
| zaro | jeblair, clarkb, fungi : i believe i was able to repro the gerrit issue on review-dev. http://paste.openstack.org/show/240745/ | 22:14 |
| *** timcline has quit IRC | 22:14 | |
| anteaya | fungi: that sounds like you found something useful | 22:14 |
| fungi | anteaya: it's caused by that change, but also an oversight on how we publish the test artifacts for that job | 22:14 |
| anteaya | wonderful | 22:14 |
| fungi | anteaya: we upload output/index.html which is generated as part of the job | 22:14 |
| anteaya | glad it flushed out an oversight | 22:14 |
| zaro | i'm thinking my next step is to validate this fix https://git.eclipse.org/r/#/c/48288/ | 22:15 |
| fungi | anteaya: _that_ is also what teh swift upload job would create to provide a directory tree index | 22:15 |
| anteaya | zaro: yay | 22:15 |
| zaro | unless you feel that we need to do something else? | 22:15 |
| anteaya | fungi: name clash | 22:15 |
| *** pradk has joined #openstack-infra | 22:15 | |
| fungi | anteaya: in this particular case, the proposed change is broken in such a way that output/index.html is 0 bytes in length | 22:15 |
| anteaya | not a helpful file length | 22:16 |
| fungi | anteaya: then the job helpfully uploads that as index.html for the logs files | 22:16 |
| anteaya | zaro: I'm supportive of your direction | 22:16 |
| anteaya | fungi: I think I follow | 22:16 |
| jeblair | zaro: awesome, i agree you have reproduced the error and next step is validate the fix. yay! | 22:16 |
| fungi | zaro: yeah, i saw it ended up in the 4.0.0 rc2 tag yesterday, so hopefully gerrit will also be picking that up soon | 22:16 |
| greghaynes | zaro: nice! | 22:16 |
| waynr | jeblair: if i can get my team on board with moving to zuul2 i should eventually be able to participate in zuul3 efforts | 22:17 |
| anteaya | waynr: one step at a time, motivating a team to change takes work | 22:17 |
| zaro | cool, now need to figure out how to package that change into gerrit | 22:17 |
| * mordred finds bottles of liquor help | 22:17 | |
| *** erikmwilson_ has quit IRC | 22:17 | |
| *** rguillebert has quit IRC | 22:17 | |
| jeblair | waynr: if offers of smoked meat help, i feel confident we can oblige | 22:18 |
| anteaya | waynr: we welcome your efforts but don't think moving a team in a different direction will be easy | 22:18 |
| jeblair | mordred: also that | 22:18 |
| mordred | mmm. smoked meat | 22:18 |
| waynr | anteaya: my team is more than ready to move away from our current Jenkins-centric CI architecture, the trick will be convincing them in the next month that zuul2 is the right next step for us | 22:18 |
| *** ashleighfarnham has quit IRC | 22:18 | |
| *** dguerri is now known as dguerri`away | 22:18 | |
| * fungi finds that smoked meat goes will with bottles of liquor | 22:19 | |
| fungi | er, well | 22:19 |
| anteaya | waynr: ah | 22:19 |
| anteaya | waynr: have any of your team spent any time in this channel? | 22:19 |
| waynr | we just finished up an 18-month long term plan that necessitates being able to scale beyond what Jenkins has to offer | 22:19 |
| waynr | anteaya: not that i am aware of, i finally got one of them to join me on JJB reviews last week though :) | 22:19 |
| anteaya | waynr: they are welcome to pop in and ask us our opinion of moving | 22:19 |
| anteaya | waynr: yay, progress | 22:20 |
| *** annegentle has quit IRC | 22:20 | |
| zxiiro | zaro: ah that explains why i couldn't find pelix in this channel :) | 22:21 |
| waynr | i am hoping to have a full proof of concept of zuul2 for our most common workflows by the end of June, this will be accompanied by a detailed technical overview/evaluation of zuul in the context of the user stories we have collected from our wider engineering org | 22:21 |
| zxiiro | fungi: zaro: Gerrit should pick it up soon. The JGit and Gerrit developers work very closely, some of them are even committers on both projects | 22:21 |
| anteaya | waynr: I hope you are getting all you need from us to accomplish your goals | 22:21 |
| anteaya | including offers of liquor and smoked meat | 22:23 |
| *** annegentle has joined #openstack-infra | 22:24 | |
| *** amitgandhinz has quit IRC | 22:24 | |
| fungi | jhesketh: i'm currently reading through zuul_swift_upload.py to figure out how to specify that i want to rename a file on upload, but if your around and have a moment to short-circuit me to the answer for that i would be most obliged | 22:27 |
| fungi | er, you're | 22:27 |
| fungi | oh, i bet i could just do that as part of the job instead | 22:28 |
| waynr | anteaya: ha! thanks, i will probably be increasingly active here asking questions for my POC work | 22:28 |
| mordred | fungi clarkb: I have pushed an 0.6.4 tag for shade which includes the swift upload fix | 22:28 |
| anteaya | waynr: of course, glad to know the overall plan | 22:28 |
| mordred | fungi, clarkb: I have also pushed a 1.1.0 tag for os-client-config which includes the oscc fixes needed | 22:29 |
| anteaya | mordred: if I read through the lines correctly it appears clarkb is at the hospital and latest report is he is offline for a bit | 22:29 |
| mordred | anteaya: yes | 22:29 |
| mordred | anteaya: I agree with your lines | 22:30 |
| anteaya | just in case you were expecting a reply | 22:30 |
| anteaya | kk | 22:30 |
| mordred | nope. just leaving him nuggets he can digest upon his return | 22:30 |
| anteaya | k | 22:30 |
| *** dizquierdo has quit IRC | 22:31 | |
| *** dannywil_ has joined #openstack-infra | 22:32 | |
| *** nelsnelson has quit IRC | 22:32 | |
| openstackgerrit | Jeremy Stanley proposed openstack-infra/project-config: Rename irc-meetings index.html in job logs https://review.openstack.org/186202 | 22:33 |
| fungi | anteaya: i think this ^ should do what we want | 22:34 |
| *** kmartin_ has joined #openstack-infra | 22:34 | |
| *** fawadkhaliq has joined #openstack-infra | 22:34 | |
| *** kmartin has quit IRC | 22:34 | |
| *** patrickeast_ has joined #openstack-infra | 22:34 | |
| * anteaya looks | 22:34 | |
| *** soren_ has quit IRC | 22:34 | |
| *** dimtruck is now known as zz_dimtruck | 22:35 | |
| *** dannywilson has quit IRC | 22:35 | |
| anteaya | looks like it should do the job | 22:35 |
| anteaya | mordred: can you cast an eye on 186202 please? | 22:36 |
| *** patrickeast has quit IRC | 22:36 | |
| *** patrickeast_ is now known as patrickeast | 22:36 | |
| *** tonyb has joined #openstack-infra | 22:37 | |
| tonyb | Am I in trouble? | 22:37 |
| anteaya | not at all | 22:37 |
| tonyb | \o/ | 22:37 |
| tonyb | How can I help? | 22:37 |
| anteaya | wanted to share the irc-meeting job situation | 22:37 |
| anteaya | https://review.openstack.org/#/c/186202/ | 22:37 |
| anteaya | if you could review that, that would be great | 22:37 |
| anteaya | basically the swift upload failed since the logs were called the same name as the file the swift job creates | 22:38 |
| anteaya | so that patch renames the file created by the ical job for the tests | 22:38 |
| anteaya | to avoid a name clash | 22:38 |
| *** fawadkhaliq has quit IRC | 22:38 | |
| fungi | well, the upload didn't exactly fail | 22:38 |
| openstackgerrit | Matt Riedemann proposed openstack/requirements: Leave wiggle room for zake and kazoo fix releases https://review.openstack.org/186123 | 22:39 |
| * anteaya listens while fungi fixes what she said | 22:39 | |
| fungi | in fact the upload successfully replaced the file index for the logs | 22:39 |
| jlvillal | I think this is a simple change to requirements: https://review.openstack.org/185770 But I would like to get feedback if this is the correct route. At the moment we are changing tox.ini in openstack/ironic to work-around the issue. | 22:40 |
| fungi | but ideally we'd like to get the file index for the job logs rather than the irc index being (sometimes unsuccessfully if the change is bad) generated by tox | 22:40 |
| tonyb | fungi: Yeah. I was reviewing a few irc-meetings changes and it was bit of a pain. | 22:41 |
| *** salv-orlando has joined #openstack-infra | 22:41 | |
| tonyb | fungi: Thanks for the fix. | 22:41 |
| anteaya | ah I hadn't looked at any irc-meeting logs before | 22:41 |
| anteaya | yes, http://logs.openstack.org/75/186075/4/check/gate-irc-meetings-tox-ical/e0043c6/ is kind of useless in debugging a job | 22:42 |
| tonyb | anteaya: +1 | 22:42 |
| *** derekh_ has quit IRC | 22:42 | |
| fungi | i'm hoping the approach of renaming it as the first publisher task will work around the fact that we need it to have that name when the scp publisher pushes it to the eavesdrop server but can't have that name when the swift publisher uploads it on check results | 22:42 |
| anteaya | tonyb: I just wanted to include you so you knew what was happening going forward | 22:42 |
| *** aysyanne has quit IRC | 22:43 | |
| anteaya | fungi: it will be fun to track | 22:43 |
| fungi | anyway, no fire, just resulted in some mild confusion over broken change https://review.openstack.org/186184 | 22:43 |
| anteaya | mordred: thank you | 22:44 |
| tonyb | Thanks all. I need to take the kids to school. | 22:44 |
| *** timcline has joined #openstack-infra | 22:44 | |
| anteaya | tonyb: enjoy | 22:45 |
| fungi | jlvillal: can you elaborate on "supposed to be installed by WSME, but for some reasons it was missed in py3 env"? | 22:45 |
| fungi | jlvillal: does wsme generally have this failing on py3k? | 22:45 |
| jlvillal | fungi: The openstack/ironic gate started failing. | 22:46 |
| *** annegentle has quit IRC | 22:46 | |
| fungi | dhellmann: ^ 185770 may point to a deeper problem | 22:46 |
| jlvillal | Adding the 'simplegeneric' to our tox.ini fixed it. I didn't do the initial patch. | 22:46 |
| jlvillal | fungi: I'm not sure why it started failing. I'm copying the text that was put into the patch which fixed the issue to make the gate work. | 22:47 |
| dhellmann | jlvillal: is simplegeneric in the requirements list for wsme? | 22:47 |
| fungi | jlvillal: right, but you asked what the correct approach is to solving this. i think it starts with having a solid understanding of what's broken | 22:47 |
| jlvillal | dhellmann: I think so. When I do the install without this fix on my system it works. | 22:47 |
| fungi | so we can figure out where it should be fixed | 22:47 |
| jlvillal | fungi: I am unsure what is the underlying cause. It is only happening on Python 3. | 22:48 |
| fungi | jlvillal: do you have a change proposed to trove to revert the workaround so i can see it failing? | 22:48 |
| dhellmann | jlvillal: ok, it shouldn't be necessary to add it to ironic's requirements, then. Do you have a link to logs for a failed job or a bug? | 22:48 |
| jlvillal | fungi: I do not. | 22:48 |
| fungi | that's probably a great place to start then. wip revert of the workaround change | 22:49 |
| jlvillal | dhellmann: It was failing around Tuesday of last week. So the hack was put into make it workaround. | 22:49 |
| fungi | when we see how it breaks, we'll know more | 22:49 |
| jroll | this may have been the one where us cores said "f it, let's fix the gate and then file a bug" | 22:49 |
| dhellmann | fungi: the dependency on simplegenerics is the same in requirements.txt and requirements-py3.txt (no range on either) | 22:49 |
| *** otter768 has joined #openstack-infra | 22:49 | |
| *** timcline has quit IRC | 22:49 | |
| jroll | that's totally my fault :( | 22:49 |
| *** kutija has quit IRC | 22:49 | |
| jlvillal | fungi: dhellmann: This is the patch: https://review.openstack.org/#/c/184429/ | 22:49 |
| dhellmann | jlvillal, jroll : ok, you should try reverting that and if you see it continue to fail we can help you debug it now that the summit is over | 22:50 |
| jroll | fungi: iirc, the theory (or actual behavior?) was it being uninstalled due to not being in requirements.txt | 22:50 |
| jroll | dhellmann: yeah, +1 | 22:50 |
| jroll | jlvillal: you got this? I was just passing by :) | 22:50 |
| jlvillal | dhellmann: fungi jroll : I will try a revert patch and see what happens. | 22:50 |
| dhellmann | jroll: hmm, I'm not aware of anything that uninstalls packages like you describe unless it's also trying to upgrade | 22:50 |
| * jlvillal has to walk home first. | 22:50 | |
| *** krtaylor has quit IRC | 22:50 | |
| jroll | sweet, thanks | 22:51 |
| dhellmann | jlvillal: ping me when you've got some logs and I'll see if I can help you out | 22:51 |
| *** afazekas has quit IRC | 22:51 | |
| jroll | dhellmann: I think there's code around "this is in reqs.txt but not in g-r" that does something like that | 22:51 |
| jlvillal | dhellmann: thanks | 22:51 |
| jroll | dhellmann: but then I went to a summit and forgot everything I had learned shortly before that :P | 22:51 |
| dhellmann | jroll: on a sync it does, but not when installing stuff | 22:51 |
| jroll | dhellmann: ok, my memory is poor atm | 22:52 |
| fungi | also that shouldn't in theory be happening on unit tests, which i think is the only place python 3.x comes into play? | 22:52 |
| dhellmann | jroll: and that's just to your project's requirements*.txt, not to wsme's -- pip should still be dragging in the secondary dependency | 22:52 |
| *** akshai has quit IRC | 22:52 | |
| dhellmann | fungi: good point | 22:52 |
| jroll | ah, true. hm. | 22:53 |
| * jroll shrugs | 22:53 | |
| jroll | let's see where the revert goes | 22:53 |
| fungi | but yeah, the quick-fix change lacks any debugging details about why it was implemented | 22:53 |
| anteaya | fungi: seems publishers can't talk shell: jenkins_jobs.errors.JenkinsJobsException: Unknown entry point or macro 'shell' for component type: 'publisher'. | 22:53 |
| *** camunoz has joined #openstack-infra | 22:53 | |
| jhesketh | Morning | 22:53 |
| anteaya | morning jhesketh | 22:53 |
| anteaya | and what great timing, we could use your help | 22:53 |
| anteaya | jhesketh: let us know when you are appropriately coffeed | 22:54 |
| fungi | anteaya: thanks. jhesketh have suggestions for how i can best do https://review.openstack.org/186202 ? | 22:54 |
| *** otter768 has quit IRC | 22:54 | |
| jroll | fungi: yeah, it's weird what happens when we talk about changes face-to-face :) | 22:54 |
| jhesketh | anteaya: heh, not sure how much coffee that'll require ;-). What's up? | 22:54 |
| jhesketh | fungi: looking | 22:54 |
| anteaya | jhesketh: right now we are getting http://logs.openstack.org/59/186159/1/check/gate-irc-meetings-tox-ical/77ac141/ as output for the job run, we would like to have the job logs | 22:55 |
| jhesketh | fungi: what you have done seems sensible to me | 22:56 |
| anteaya | publishers can't talk shell though | 22:56 |
| anteaya | jenkins fails on that approach | 22:56 |
| *** dannywil_ has quit IRC | 22:57 | |
| fungi | interestingly, it looks like a publisher can be a builder | 22:57 |
| anteaya | builders can talk shell | 22:57 |
| fungi | yeah, i'm just now reverse-engineering our publisher for swift uploads | 22:57 |
| jhesketh | Yeah we can use a publisher builders to do shell | 22:57 |
| fungi | postbuildscript builders | 22:58 |
| fungi | found it. thanks | 22:58 |
| fungi | the zuul-swift-devstack-logs-with-console publisher is the example i wanted | 22:58 |
| jhesketh | fungi: alternatively, and perhaps better, we can upload the whole output folder as a folder and train zuul-swift-upload to not generate an index of one already exists | 22:58 |
| openstackgerrit | Jeremy Stanley proposed openstack-infra/project-config: Rename irc-meetings index.html in job logs https://review.openstack.org/186202 | 22:59 |
| openstackgerrit | Adam Gandelman proposed openstack-infra/system-config: Add #akanda to openstack bot's channel list https://review.openstack.org/182368 | 23:00 |
| fungi | jhesketh: great idea. in the meantime there's the simple hack | 23:00 |
| lifeless | jeblair: dhellmann: IIRC you took photos of the whiteboard on the requirements BOF; could you send me copies? | 23:00 |
| *** angela-s has quit IRC | 23:00 | |
| openstackgerrit | Joe Gordon proposed openstack-dev/hacking: Add example to versioning section of README https://review.openstack.org/185252 | 23:01 |
| openstackgerrit | Joe Gordon proposed openstack-dev/hacking: Revise adding additional checks section of README https://review.openstack.org/185253 | 23:01 |
| openstackgerrit | Joe Gordon proposed openstack-dev/hacking: Take advantage of RST formatting for links https://review.openstack.org/185257 | 23:01 |
| *** doug-fish has left #openstack-infra | 23:01 | |
| fungi | jhesketh: though... i think we still want file indexes on those. for example it's nice to be able to browse the irc-meetings.ical file in the output directory even if generating output/index.html goes haywire and ends in a 0-byte file | 23:01 |
| *** Rockyg has joined #openstack-infra | 23:01 | |
| jhesketh | fungi: hmm, that's a good point. Maybe zuul-swift-upload needs a feature to detect and move indexes out of the way | 23:03 |
| fungi | you can still of course get to irc-meetings.ical by blindly feeding in the url in that case, but less confusing if just trying to follow the link to site-index.html results in a blank page (or in this recent case your browser offering to let you download a null-length file of unknown mime type) | 23:03 |
| *** sarob has joined #openstack-infra | 23:03 | |
| *** mriedem has quit IRC | 23:04 | |
| jeblair | lifeless: i don't think i did. i only have pictures of float planes. i believe dhellmann did. | 23:04 |
| *** gokrokve has quit IRC | 23:04 | |
| *** pradk has quit IRC | 23:04 | |
| openstackgerrit | Spencer Krum proposed openstack-infra/puppet-log_processor: Build worker config from template https://review.openstack.org/186214 | 23:05 |
| fungi | jeblair: you should have said "sure! hold on..." and then sent him pictures of float planes instead | 23:05 |
| anteaya | were they selfies with float planes? | 23:05 |
| jeblair | fungi, anteaya: there's a good chance that's actually what dhellmann was doing | 23:05 |
| *** chlong has joined #openstack-infra | 23:05 | |
| anteaya | ha ha ha | 23:05 |
| lifeless | jeblair: cool; I'll await my float plane photobomb | 23:06 |
| nibalizer | clarkb: interested in your feedback on 186214 | 23:07 |
| *** wuhg has joined #openstack-infra | 23:07 | |
| anteaya | nibalizer: he might or might not be offline for a while | 23:07 |
| nibalizer | yea for to do baby things | 23:08 |
| anteaya | just by way of managing your turn around expectations | 23:08 |
| anteaya | jhesketh: care to review now that Jenkins likes this? https://review.openstack.org/#/c/186202/ | 23:10 |
| *** sarob_ has joined #openstack-infra | 23:10 | |
| *** tonytan4ever has quit IRC | 23:10 | |
| mordred | lifeless: I took pictures in my mind | 23:10 |
| *** redixin has quit IRC | 23:10 | |
| anteaya | I'm not sure I want to see those | 23:11 |
| anteaya | do the lines stay still in your mind pictures? | 23:11 |
| openstackgerrit | James E. Blair proposed openstack-infra/project-config: Update infra project ACLs https://review.openstack.org/186217 | 23:11 |
| jeblair | that's a fun change ^ | 23:12 |
| *** sarob has quit IRC | 23:12 | |
| *** sarob_ is now known as sarob | 23:12 | |
| jeblair | i could do that as 120 changes if people prefer. | 23:13 |
| *** salv-orlando has quit IRC | 23:14 | |
| anteaya | I don't prefer 120 changes | 23:14 |
| anteaya | what do you want to do with projects that noone is really working on? | 23:14 |
| anteaya | are folks working on openstack-nose for instance? | 23:14 |
| *** koolhead17 has joined #openstack-infra | 23:15 | |
| *** koolhead17 has quit IRC | 23:15 | |
| *** koolhead17 has joined #openstack-infra | 23:15 | |
| jogo | mordred: wow this is painful: http://git.openstack.org/cgit/openstack/os-client-config/tree/os_client_config/vendors.py | 23:15 |
| jeblair | anteaya: we might consider retiring openstack-nose if we're done with it, but we can handle that separately. | 23:15 |
| jeblair | jogo: yeah, i believe it represents a failure to achive our goal as a project. | 23:15 |
| *** sdake_ has joined #openstack-infra | 23:15 | |
| *** notnownikki has quit IRC | 23:16 | |
| anteaya | jeblair: okay | 23:16 |
| mordred | jogo: yah | 23:16 |
| jogo | jeblair: agreed | 23:17 |
| openstackgerrit | Doug Wiegley proposed openstack-infra/project-config: Add second job for lbaasv1 API https://review.openstack.org/184870 | 23:17 |
| jeblair | (and to be really specific, i think it's in direct contradition to 'ubiquitous' in the mission statement) | 23:17 |
| jogo | mordred: good response to Keith | 23:17 |
| mordred | jogo: thanks! | 23:18 |
| mordred | jogo: hopefully it came across as a yes, and | 23:18 |
| *** heyongli- has joined #openstack-infra | 23:18 | |
| nibalizer | ya there are a couple puppet modules that could be nominated for attic or whatever other disposal process is available to us | 23:19 |
| *** yamamoto has quit IRC | 23:19 | |
| *** ChuckC has joined #openstack-infra | 23:19 | |
| nibalizer | but as jeblair said, different change | 23:19 |
| *** sdake has quit IRC | 23:19 | |
| clarkb | jhesketh: with the ongoing nodepool dib things then being afk I haven't gotten to the logs-dev stuff yet and likely wont today | 23:20 |
| jeblair | nibalizer: in infra, i expect it will be a README change to the project, and perhaps marking the project read-only if we feel really strongly about it. :) i don't buy into the idea that we need to change the repo name when we stop working on a project. | 23:20 |
| *** hdd has quit IRC | 23:21 | |
| fungi | or worse, when we have changes in governance of a project | 23:21 |
| * nibalizer nod | 23:21 | |
| jeblair | yeesh, indeed. | 23:21 |
| clarkb | nibalizer: 186214 doesn't produce a working worker config so not sure what the intent is there | 23:21 |
| anteaya | jeblair: I'm wondering about all the puppet modules | 23:22 |
| jeblair | anteaya: i think they're great | 23:22 |
| jeblair | and getting better all the time | 23:22 |
| clarkb | nibalizer: oh wait its the worker not the client I may be mistaken | 23:22 |
| anteaya | jeblair: I see the same puppet folks working on all of them, would it make sense to have a puppet-core team for all of them? | 23:22 |
| *** ZZelle_ has quit IRC | 23:22 | |
| *** TravT_ has quit IRC | 23:22 | |
| jeblair | anteaya: yes, but i'm going to do that with gerrit groups rather than acl files | 23:22 |
| openstackgerrit | Doug Hellmann proposed openstack-infra/release-tools: Add make_feature_branch.sh https://review.openstack.org/186222 | 23:22 |
| anteaya | jeblair: I'm just talking acl changes atm | 23:22 |
| nibalizer | clarkb: yea, the names in this section of -infra are not ideal | 23:22 |
| nibalizer | at least I get confused a lot | 23:23 |
| clarkb | nibalizer: gearmanisms | 23:23 |
| anteaya | oh so the puppet-core is a memeber of each individual puppet module | 23:23 |
| jeblair | anteaya: that's what i meant by the sentence "by using group inclusion in Gerrit" | 23:23 |
| jeblair | anteaya: yep | 23:23 |
| clarkb | nibalizer: so I think my initial review is why, the change doesn't capture any of that | 23:23 |
| anteaya | okay thanks | 23:23 |
| jeblair | anteaya: groups can be members of groups, so that's easy to do | 23:23 |
| clarkb | nibalizer: if I understand correctly this doesn't make the worker more or less configurable | 23:23 |
| jeblair | that's what i'm working on next, and want to get in place before we merge that change | 23:23 |
| clarkb | nibalizer: it just changes the api, so capturing the reason behind that would be good | 23:23 |
| anteaya | jeblair: yes, was just wondering what you were planning, so thanks for letting me know | 23:24 |
| jeblair | anteaya: but if there's someone who wants to be puppet-zuul core but doesn't care about any other modules, we can do that just as easily in the new system | 23:24 |
| nibalizer | clarkb: kindof, it moves the configuration closer to the thing doing the configuriaton | 23:24 |
| anteaya | right | 23:24 |
| *** hichihara has joined #openstack-infra | 23:24 | |
| anteaya | just thinking the puppet people tend to move as a group, that I see | 23:24 |
| clarkb | nibalizer: but is it worth breaking everyone using hte module? in theory just us but who knows | 23:25 |
| clarkb | nibalizer: good to justify that | 23:25 |
| nibalizer | clarkb: yea I can get in to that | 23:25 |
| nibalizer | oh gross puppet doesn't like variables with dashes | 23:25 |
| clarkb | nibalizer: correct | 23:25 |
| clarkb | this is why our projects are named puppet-foo_br | 23:25 |
| nibalizer | so things like output-host are gonna be output_host in puppet | 23:25 |
| nibalizer | :/ | 23:25 |
| clarkb | yes | 23:25 |
| nibalizer | i kenw about modules, I didn't think it mattered for variables | 23:26 |
| *** tim_o has quit IRC | 23:26 | |
| fungi | nibalizer: lots of languages dislike hyphens in variable names since long ago the authors of ascii decided to use one byte for both the subtraction sign and the hyphen | 23:26 |
| *** abregman has quit IRC | 23:27 | |
| jeblair | fungi: except of course the one true language | 23:27 |
| *** hdd has joined #openstack-infra | 23:27 | |
| fungi | yes | 23:27 |
| clarkb | fungi: in fairness old typewriters didn't have 1 keys (supposedly this is why 0 cames after 9 but they didn't have 0 keys either in many cases) | 23:28 |
| crinkle | yeah the module names and variable names are different stories | 23:28 |
| clarkb | fungi: so its good we got different characters in ascii for those | 23:28 |
| fungi | agreed. also glad lisp treats l/1 and 0/O differently ;) | 23:28 |
| jhesketh | clarkb: no trouble.. unfortunately I was unable to reproduce the issue locally | 23:29 |
| jeblair | fungi: i think you just poked the eyes out of anyone reading this in droid sans | 23:29 |
| clarkb | jeblair: I am reading in droid l and 1 are distinct but 0/O are not | 23:30 |
| fungi | ...and nothing of value was lost | 23:30 |
| *** mwhagedorn has quit IRC | 23:31 | |
| jeblair | clarkb: sure, but can you distinguish l/l ? | 23:31 |
| clarkb | jeblair: nope /me copies into vim to figure out what is going on there | 23:31 |
| * fungi just realized he lacks "the critic" on dvd. will correct that posthaste | 23:31 | |
| clarkb | vim assures me its the same ascii character | 23:31 |
| * jeblair goes back to work in the other desktop before clarkb figures it out | 23:31 | |
| clarkb | :P | 23:31 |
| openstackgerrit | Spencer Krum proposed openstack-infra/puppet-log_processor: Build worker config from template https://review.openstack.org/186214 | 23:32 |
| nibalizer | clarkb: updated | 23:32 |
| jeblair | fungi: it stinks! | 23:32 |
| clarkb | mordred: ok so oscc and shade have new releases with nodepool needed fixes, rechecking your nodepool change should pass now? | 23:32 |
| clarkb | mordred: but jenkins still says no | 23:33 |
| anteaya | jeblair: still on the acl change, why is openstack/openstack-planet acl part of that patch? | 23:34 |
| *** asselin_ has joined #openstack-infra | 23:34 | |
| jeblair | anteaya: it's an infra repo | 23:34 |
| mordred | clarkb: yah - I pushed that up and rechecked because I'm stumped again and was hoping it was just a local problem | 23:34 |
| clarkb | mordred: http://logs.openstack.org/72/185972/5/check/gate-nodepool-python27/968cb96/console.html#_2015-05-27_23_25_57_541 looks like a new error | 23:34 |
| clarkb | mordred: provider doesn't have region_name | 23:34 |
| mordred | clarkb: thank you | 23:35 |
| *** prad has quit IRC | 23:35 | |
| *** wenlock has quit IRC | 23:35 | |
| clarkb | then later mysql says deadlock detected but I think thats likely due to earlier errors | 23:35 |
| *** docaedo has joined #openstack-infra | 23:36 | |
| *** nelsnelson has joined #openstack-infra | 23:36 | |
| anteaya | jeblair: yes you an monty are core: https://review.openstack.org/#/admin/groups/93,members but so are anne reed and ttx | 23:37 |
| clarkb | mordred: so I think region_name needs to remain on the provider object in nodepool.yaml | 23:37 |
| clarkb | mordred: that is a small fix | 23:37 |
| mordred | clarkb: yah - that's totally fine | 23:37 |
| jeblair | anteaya: yes, and they'll end up in the new core group too | 23:37 |
| mordred | clarkb: lemme do that real quick | 23:37 |
| *** dims_ has quit IRC | 23:38 | |
| clarkb | jhesketh: interesting, maybe we broke the non swift case? did you test that too? | 23:38 |
| *** fawadkhaliq has joined #openstack-infra | 23:38 | |
| openstackgerrit | Monty Taylor proposed openstack-infra/nodepool: Consume os-client-config directly https://review.openstack.org/185972 | 23:38 |
| mordred | clarkb: ^^ I reorged slightly too - so that the patch is cleaner to look at and the provider mangling is in a function | 23:39 |
| *** krtaylor has joined #openstack-infra | 23:39 | |
| jlvillal | dhellmann, fungi, So reverting that simplegeneric patch failed the testing: https://review.openstack.org/#/c/186208/ | 23:39 |
| jlvillal | dhellmann, fungi, http://logs.openstack.org/08/186208/1/check/gate-ironic-python34/b3618cd/ | 23:40 |
| *** dims_ has joined #openstack-infra | 23:40 | |
| jhesketh | clarkb: yep | 23:40 |
| clarkb | mordred: yup looks good if jenkins is happy with it | 23:40 |
| jhesketh | clarkb: I only tried one file, but it worked from disk and swift | 23:40 |
| fungi | jlvillal: thanks. seeing now if i can interpret why | 23:41 |
| clarkb | jhesketh: maybe fungi remembers? iirc he and sdague were around for the breaking | 23:41 |
| jhesketh | clarkb: and the timestamps etc all marked-up correctly | 23:41 |
| *** ajmiller has quit IRC | 23:42 | |
| jlvillal | fungi, Thanks | 23:43 |
| nibalizer | clarkb: is the ascii diagram in http://docs.openstack.org/infra/system-config/logstash.html correct? | 23:43 |
| nibalizer | specificlly 1:1 ratio of logstash indexers to log-gearman workers? | 23:44 |
| nibalizer | it looks to me like there are 4 gear workers per logstash indexer | 23:45 |
| openstackgerrit | Merged openstack-infra/git-review: Isolate tests from user/system git configuration https://review.openstack.org/179954 | 23:45 |
| clarkb | nibalizer: its correct in that its not meant to imply any ratios | 23:45 |
| *** timcline has joined #openstack-infra | 23:46 | |
| clarkb | nibalizer: we run 4:1, you can run N:1 | 23:46 |
| fungi | dhellmann: jlvillal: it looks like wsme 0.6.4 is installed according to tox/py34-1.log and while it has simplegeneric in the install_requires within its setup.py that doesn't seem to get pulled in | 23:46 |
| anteaya | jeblair: okay there it is in governance: http://git.openstack.org/cgit/openstack/governance/tree/reference/projects.yaml#n595 | 23:46 |
| nibalizer | oooo | 23:46 |
| anteaya | I couldn't figure out what made it an infra repo | 23:46 |
| jlvillal | fungi, Yes, I think that is what we were seeing. And why they put in the change to tox.ini | 23:46 |
| jlvillal | fungi, Why it isn't being pulled in, not sure.... | 23:47 |
| nibalizer | I think "Currently we have multiple Logstash worker nodes that pair a log pusher with a Logstash indexer." is why I thought 1:1 | 23:47 |
| nibalizer | did it used to be 1:1 ? | 23:47 |
| lifeless | install_requires in setup.py? Is it a pbr using project? | 23:47 |
| fungi | lifeless: as of 0.6.4 it was not, no | 23:47 |
| lifeless | pbr sets install_requires from static data, setup.py's values will be ignored. | 23:47 |
| clarkb | nibalizer: it was at one point but not for very long | 23:47 |
| nibalizer | well, and the diagram | 23:47 |
| fungi | lifeless: no, wait it was indeed using pbr at that point | 23:47 |
| openstackgerrit | Jamie Lennox proposed openstack-infra/project-config: Ignore requirements job on keystoneauth branch https://review.openstack.org/186228 | 23:47 |
| fungi | lifeless: http://git.openstack.org/cgit/stackforge/wsme/tree/setup.py?id=0.6.4 for reference | 23:47 |
| fungi | lifeless: so perhaps new-world pbr has broken that | 23:48 |
| fungi | anteaya: it was previously an orphaned repo until infra adopted it out of the pile of unclaimed repos back when we started having formal programs | 23:49 |
| anteaya | ah okay thanks | 23:49 |
| fungi | anteaya: it wasn't a perfect fit for any particular program in my opinion, but it needed a home | 23:49 |
| anteaya | got it | 23:49 |
| anteaya | glad it has a roof and food | 23:49 |
| *** timcline has quit IRC | 23:50 | |
| fungi | and since it's configuration for a service being hosted on our servers, infra was a better home than the other options | 23:50 |
| *** ddutta has joined #openstack-infra | 23:50 | |
| *** yamamoto has joined #openstack-infra | 23:50 | |
| jeblair | also, some of us were reviewing it | 23:51 |
| fungi | yep, that too | 23:51 |
| lifeless | fungi: I don't think it ever worked | 23:51 |
| *** amotoki_ has joined #openstack-infra | 23:51 | |
| lifeless | fungi: we assign to the setup.py kwargs in utils.py | 23:51 |
| fungi | lifeless: dhellmann: jlvillal: so i wonder if tagging a wsme 0.7 is in order? | 23:51 |
| fungi | it looks much more pbresque | 23:52 |
| fungi | master does, i mean | 23:52 |
| fungi | 0.6.4 was ~6mos ago | 23:52 |
| *** nelsnelson has quit IRC | 23:52 | |
| fungi | it's going on 25 changes since then | 23:53 |
| lifeless | actually no, yes I'm sure pbr regressed this - but we never claimed to support it (and never had tests) - and update.py enforces a standard setup.py | 23:53 |
| lifeless | so while we might want to fix it (please file a bug), we should fix wsme IMO | 23:54 |
| dhellmann | lifeless: wsme isn't an openstack project, it's just trying to use some of the same tools | 23:54 |
| dhellmann | yeah, I agree | 23:54 |
| dhellmann | fungi, lifeless, jlvillal: how about if I make a note to release wsme on monday? do we think the current master fixes this? | 23:54 |
| fungi | dhellmann: it looks like it should. though who knows what new shadows lurk in the heart of 0.7.0 | 23:55 |
| jlvillal | fungi, dhellmann, if a new wsme fixes it then sounds good to me :) | 23:55 |
| dhellmann | fungi: right, so no releases this late in the week :-) | 23:55 |
| clarkb | mordred: down to one failure now http://logs.openstack.org/72/185972/6/check/gate-nodepool-python27/be02e34/console.html#_2015-05-27_23_44_50_777 so close | 23:55 |
| fungi | jlvillal: yeah, i guess keep your workaround in place and recheck 186208 once that happens | 23:55 |
| jlvillal | fungi, Okay. | 23:56 |
| *** amotoki_ has quit IRC | 23:56 | |
| dhellmann | jlvillal: do you have a bug open for this? | 23:56 |
| fungi | jlvillal: and if it passes, then update the commit message with some better detail about the wsme release and get it merged | 23:56 |
| samueldmq | anteaya, hello, could you please revisit 179661 ? (Experimental Devstack job for Identity v3 API only) | 23:57 |
| samueldmq | anteaya, you already had a +2 there, but it needed a rebase | 23:57 |
| samueldmq | anteaya, the devstack patch in which it depends on already started gate jobs | 23:57 |
| fungi | dhellmann: the change that added the workaround in ironic said a bug report was desired, but it was also during the summit so... | 23:57 |
| dhellmann | fungi: yeah | 23:57 |
| dhellmann | jlvillal: please file a bug against wsme so we have something to track this with | 23:57 |
| anteaya | samueldmq: sure | 23:58 |
| samueldmq | anteaya, great thanks | 23:58 |
| *** ddutta has quit IRC | 23:58 | |
| *** julim has joined #openstack-infra | 23:59 | |
| openstackgerrit | Spencer Krum proposed openstack-infra/puppet-log_processor: Build worker config from template https://review.openstack.org/186214 | 23:59 |
| anteaya | samueldmq: let's just get some fresh jenkins output on it first, I just rechecked the patch | 23:59 |
| samueldmq | anteaya, nice, makes sense | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
