| *** openstack has joined #openstack-infra | 16:48 | |
| pleia2 | hooray! | 16:49 |
|---|---|---|
| clarkb | pleia2: safe for me to approve things that have the necessary reviews? | 16:49 |
| pleia2 | clarkb: yes | 16:50 |
| *** cppcabrera has joined #openstack-infra | 16:50 | |
| *** zaro0508 has joined #openstack-infra | 16:50 | |
| jeblair | clarkb: so had i an opportunity to review I4b83c78d83efdd9da8f99c72e85097370dffddc3 i would have said "-1 can you put this in the devstack-logs and devstack-logs-periodic" macro instead? | 16:51 |
| pleia2 | jeblair: hopefully I did the class thing right re: 37770 | 16:51 |
| openstackgerrit | A change was merged to openstack-infra/config: Add /cgit rewrite for git server docroot https://review.openstack.org/37623 | 16:51 |
| *** Protux has joined #openstack-infra | 16:52 | |
| jeblair | clarkb: from the shell logs in 37776, it looks like it's working as intended, so i think that's ready. | 16:52 |
| clarkb | jeblair: ok, I will propose a change to follow up and correct JJB if that helps | 16:52 |
| jeblair | clarkb: i'm on it. | 16:53 |
| *** ladquin has joined #openstack-infra | 16:54 | |
| *** derekh has quit IRC | 16:54 | |
| openstackgerrit | James E. Blair proposed a change to openstack-infra/config: Move devstack testr log copying to macros https://review.openstack.org/37942 | 16:54 |
| clarkb | pleia2: https://review.openstack.org/#/c/37751/ I left a note there (and approved it) | 16:58 |
| openstackgerrit | A change was merged to openstack-infra/jeepyb: Add mkdir of git directories to cgit script https://review.openstack.org/37751 | 16:59 |
| pleia2 | clarkb: good to know, thanks :) | 16:59 |
| openstackgerrit | Vijendar Komalla proposed a change to openstack/requirements: Add python-troveclient to requirements https://review.openstack.org/37686 | 17:00 |
| openstackgerrit | James E. Blair proposed a change to openstack-infra/config: Move devstack testr log copying to macros https://review.openstack.org/37942 | 17:00 |
| openstackgerrit | James E. Blair proposed a change to openstack-infra/config: Set the Zuul/Jenkins base log URL in a function https://review.openstack.org/36320 | 17:00 |
| *** niska` has quit IRC | 17:01 | |
| jeblair | clarkb, fungi: ^ can you review those two? I'd like to go ahead and merge them and start using the new feature. | 17:01 |
| pleia2 | btw, the openstack blog interviewed me (it's up now), I said lots of nice things about you coremudgeons | 17:01 |
| fungi | already looking, jeblair | 17:02 |
| clarkb | pleia2: did you call us coremudgeons? | 17:02 |
| pleia2 | clarkb: alas, no | 17:02 |
| clarkb | :( | 17:02 |
| pleia2 | jeblair did a couple posts before though! | 17:02 |
| clarkb | pleia2: fungi https://review.openstack.org/#/c/37770/4/modules/openstack_project/manifests/git.pp line 58, authorized_keys don't come with a gerrit url do they? should that variable be removed fro mthat line? | 17:03 |
| *** anteaya has joined #openstack-infra | 17:03 | |
| *** niska has joined #openstack-infra | 17:03 | |
| fungi | clarkb: pleia2: correct it should just be the key | 17:03 |
| pleia2 | clarkb: right (I swear I actually do know how to use ssh, this is what I get for copy pasta from a script that was using known_hosts to ssh in the opposite direction) | 17:03 |
| *** Protux has quit IRC | 17:03 | |
| pleia2 | :) | 17:04 |
| pleia2 | fixing | 17:04 |
| SpamapS | http://paste.openstack.org/show/40942 | 17:05 |
| SpamapS | halp ? | 17:05 |
| openstackgerrit | Elizabeth Krumbach Joseph proposed a change to openstack-infra/config: Add jeepyb & cgit repo creation script to git.o.o https://review.openstack.org/37770 | 17:05 |
| clarkb | SpamapS: upgrade distribute globally | 17:05 |
| clarkb | SpamapS: that appears to be the pain from distribute and setuptools merging | 17:06 |
| clarkb | (I think you can upgrade setuptools globally as well and get the same effect) | 17:06 |
| fungi | or upgrade setuptools. either way | 17:06 |
| *** nati_ueno has joined #openstack-infra | 17:06 | |
| * clarkb reviews jeblair's change now | 17:06 | |
| fungi | yeah. the problem arises when distribute gets upgraded as part of a set of packages and then other packages try to install/upgrade before setuptools upgrades | 17:07 |
| SpamapS | I have setuptools 0.6c11 and distribute 0.7. No change. | 17:07 |
| blamar | http://logs.openstack.org/37921/1/check/gate-nova-docs/27861/console.html new bug? doesn't seem to be affecting other jobs so might have been a blip? couldn't find an error like it on the rechecks page | 17:07 |
| fungi | SpamapS: is that the latest setuptools? | 17:08 |
| SpamapS | no | 17:08 |
| jeblair | blamar: yeah, blip, sorry. i'm moving the log storage around and permissions were wrong briefly | 17:08 |
| SpamapS | hm, why didn't -U pick it up? | 17:08 |
| SpamapS | ok now 0.7 of both | 17:08 |
| jeblair | blamar: i'd go with 'recheck no bug' since it's a known, nearly fixed problem. | 17:08 |
| blamar | jeblair: no worries, ty! | 17:08 |
| SpamapS | http://paste.openstack.org/show/40943 | 17:08 |
| SpamapS | slightly different similar fail | 17:09 |
| openstackgerrit | Russell Bryant proposed a change to openstack-infra/config: Update the default extension https://review.openstack.org/37755 | 17:09 |
| openstackgerrit | Russell Bryant proposed a change to openstack-infra/config: Update asterisk documentation https://review.openstack.org/37762 | 17:09 |
| SpamapS | all I wanted was a newer hacking. :-/ | 17:10 |
| openstackgerrit | Elizabeth Krumbach Joseph proposed a change to openstack-infra/config: Add jeepyb & cgit repo creation script to git.o.o https://review.openstack.org/37770 | 17:12 |
| fungi | SpamapS: it's an upstream python distribute/setuptools growing pain point, to be sure. we hit it a week or two ago and promptly puppeted all our servers to install the newest setuptools as a workaround | 17:12 |
| SpamapS | I have 0.9.6 | 17:12 |
| SpamapS | hm | 17:14 |
| *** linuxmohan has joined #openstack-infra | 17:14 | |
| *** ewindisch has quit IRC | 17:15 | |
| *** ewindisch has joined #openstack-infra | 17:16 | |
| clarkb | jeblair: reviewed. I found a typo that will need fixign | 17:16 |
| clarkb | fungi: I don't think that change has merged. I will hit the go button on it now | 17:16 |
| clarkb | unless we want to wait for zuul to be done | 17:16 |
| openstackgerrit | James E. Blair proposed a change to openstack-infra/config: Add voipms configuration https://review.openstack.org/37765 | 17:17 |
| fungi | clarkb: which change hasn't merged? | 17:17 |
| clarkb | fungi: https://review.openstack.org/#/c/36708/ | 17:17 |
| fungi | ahh, right, that one | 17:18 |
| *** cp16net is now known as cp16net|away | 17:19 | |
| *** sarob has joined #openstack-infra | 17:19 | |
| SpamapS | ok | 17:19 |
| SpamapS | so I had to rip out the system setuptools | 17:19 |
| *** jjmb has joined #openstack-infra | 17:19 | |
| openstackgerrit | James E. Blair proposed a change to openstack-infra/config: Set the Zuul/Jenkins base log URL in a function https://review.openstack.org/36320 | 17:19 |
| SpamapS | worth noting that if anybody else gets stuck | 17:19 |
| *** linuxmohan has quit IRC | 17:20 | |
| *** sarob has quit IRC | 17:20 | |
| clarkb | SpamapS: system being apt installed? | 17:20 |
| SpamapS | right | 17:20 |
| SpamapS | wtf | 17:20 |
| SpamapS | I thought we were going to ignore periods on commit messages?! | 17:20 |
| *** sarob has joined #openstack-infra | 17:21 | |
| SpamapS | ./setup.py:1:1: H803 git commit title ('Run forever if a command is configured.') should not end with period | 17:21 |
| SpamapS | gnar | 17:21 |
| clarkb | SpamapS: I think they added a rule so that you could choose one way or the other | 17:22 |
| clarkb | SpamapS: but you can easily add an ignore in tox.ini | 17:22 |
| SpamapS | that | 17:22 |
| SpamapS | is | 17:22 |
| SpamapS | stupid | 17:22 |
| SpamapS | lets just pick one way, kthxbai! | 17:22 |
| *** pabelanger has quit IRC | 17:22 | |
| SpamapS | (and the one way we picked on the mailing list was to ignore entirely. | 17:22 |
| SpamapS | but I digress | 17:22 |
| SpamapS | 'tis friday, and I have a talk to get ready for | 17:23 |
| * SpamapS ignores this insanity :p | 17:23 | |
| *** sarob has quit IRC | 17:25 | |
| *** sarob_ has joined #openstack-infra | 17:25 | |
| clarkb | we should probably change the mime type on our IRC log files | 17:26 |
| clarkb | so that the files open in your browser | 17:26 |
| pleia2 | chrome is ok with it | 17:27 |
| _TheDodd_ | Where can I find the actual gerrit hooks themselves? Such as change-merged, patchset-created, etc. | 17:28 |
| openstackgerrit | A change was merged to openstack-infra/jenkins-job-builder: xunit publisher was wrong with multiple types https://review.openstack.org/37634 | 17:28 |
| openstackgerrit | Sergey Lukjanov proposed a change to openstack-infra/config: Add #savanna to eavesdrop https://review.openstack.org/37945 | 17:28 |
| clarkb | _TheDodd_: in the jeepyb repo. https://github.com/openstack-infra/jeeypb | 17:28 |
| clarkb | oh nevermind | 17:29 |
| clarkb | _TheDodd_: the things that gerrit actually execs. One sec | 17:29 |
| pleia2 | ah, the cgit rewrite needs an apache restart | 17:29 |
| pleia2 | maybe make an exec that subscribes to the cgit apache file? | 17:29 |
| clarkb | _TheDodd_: https://github.com/openstack-infra/config/tree/master/modules/openstack_project/files/gerrit and https://github.com/openstack-infra/config/tree/master/modules/openstack_project/templates | 17:30 |
| _TheDodd_ | clarkb, yes. The actual executable files. I've been searching around quite a bit, but I can't seem find them. | 17:30 |
| _TheDodd_ | clarkb, thanks. | 17:30 |
| _TheDodd_ | I'll take a look at them. | 17:30 |
| jeblair | pleia2: if you use the apache::vhost module, it does all that for you | 17:30 |
| clarkb | pleia2: not an exec. Instead add an apache service and make it subscribe to that file | 17:30 |
| *** pabelanger has joined #openstack-infra | 17:30 | |
| clarkb | pleia2: or what jeblair suggests | 17:30 |
| mgagne | _TheDodd_: not sure if it's the correct file: https://github.com/openstack-infra/config/blob/master/modules/openstack_project/templates/gerrit_patchset-created.erb | 17:31 |
| *** ^demon has joined #openstack-infra | 17:31 | |
| *** ^demon has joined #openstack-infra | 17:31 | |
| clarkb | mgagne: that is one of them | 17:31 |
| clarkb | pleia2: chromium works too but firefox does not | 17:32 |
| clarkb | silly firefox | 17:32 |
| fungi | jeblair: spotted a tiny couple of omissions in 36320 | 17:32 |
| pleia2 | jeblair: yeah, poking around the module now | 17:33 |
| clarkb | pleia2: note we use a really old version because they made lots of incompatible changes and we never updated our end | 17:35 |
| mordred | clarkb: the apache module thing? | 17:36 |
| clarkb | mordred: ya | 17:36 |
| pleia2 | thanks | 17:36 |
| mordred | clarkb: we should really go back and figure that out at some point | 17:36 |
| clarkb | mordred: we should | 17:36 |
| *** Ryan_Lane has joined #openstack-infra | 17:36 | |
| *** Ryan_Lane has quit IRC | 17:36 | |
| *** Ryan_Lane has joined #openstack-infra | 17:36 | |
| openstackgerrit | James E. Blair proposed a change to openstack-infra/config: Set the Zuul/Jenkins base log URL in a function https://review.openstack.org/36320 | 17:36 |
| mgagne | pleia2: they refactored the whole module a couple of weeks ago | 17:36 |
| jeblair | fungi: thanks. i believe those snuck in while this was in review. | 17:36 |
| clarkb | pleia2: you can blame hunner | 17:36 |
| clarkb | pleia2: but they are good changse and hunner did the right hting. its just slightly painful for us | 17:36 |
| mordred | clarkb, jeblair: btw - I realized yesterday that once pleia2 get dib gating working, there's a decent path forward for us to test changes to the content of the nodes we use in d-g | 17:37 |
| mordred | clarkb, jeblair: I'm going to write down what I'm thinking there, but I think there's a couple of really sexy possibilities | 17:37 |
| jeblair | mordred: ok. what content? | 17:37 |
| mgagne | clarkb: lot of refactors coming I would say for a lot of modules, we got hit by the one for rabbitmq this week | 17:38 |
| mordred | jeblair: the stuff that's in the d-g machines- as opposed to the mechanism for launching them | 17:38 |
| jeblair | mordred: what stuff? | 17:38 |
| *** SergeyLukjanov has quit IRC | 17:38 | |
| mordred | jeblair: like the other day when we did the openjdk7 change and that affected the d-g nodes themselves | 17:39 |
| mordred | (that change was successful, clearly, just an example of the type of change I'm talking about) | 17:39 |
| *** nati_ueno has quit IRC | 17:40 | |
| jeblair | mordred: okay, well, maybe what you're writing down will help me appreciate what you're thinking; but my initial thought is that in general we don't attempt to gate the rest of the world... | 17:41 |
| clarkb | https://launchpad.net/ubuntu/precise/+source/openjdk-6 is still without a new release. ++ to upgrading to 7 | 17:41 |
| jeblair | mordred: and if it turs out our operating system is unable to host openstack development, we accept the risk that everything will stop working until we fix that, rather than try to artificially hold ourserves to a previous version | 17:42 |
| mordred | jeblair: yeah, working on writing it down right now. _definitely_ not talking about gating the rest of the world - purely talking about gating changes we decide to make. | 17:42 |
| openstackgerrit | will soula proposed a change to openstack-infra/jenkins-job-builder: Adding the Sauce OnDemand Plugin https://review.openstack.org/37947 | 17:42 |
| *** nati_ueno has joined #openstack-infra | 17:42 | |
| mordred | jeblair: we're missing each other. I'm sorry I brought it up without having writen down the thoughts first | 17:42 |
| jeblair | mordred: i'm not entirely sure we're missing each other. could you accomplish what you're thinking by adding 'puppet agent --test' to devstack-gate and running that in response to changes to infra/config? | 17:45 |
| ekarlso | win 3 | 17:45 |
| openstackgerrit | A change was merged to openstack-infra/config: Add jeepyb & cgit repo creation script to git.o.o https://review.openstack.org/37770 | 17:47 |
| openstackgerrit | A change was merged to openstack-infra/config: Move devstack testr log copying to macros https://review.openstack.org/37942 | 17:49 |
| reed | can we have stats for the IRC channel logs? | 17:49 |
| openstackgerrit | A change was merged to openstack-infra/config: Set the Zuul/Jenkins base log URL in a function https://review.openstack.org/36320 | 17:49 |
| jeblair | break-everything friday continues ^ | 17:49 |
| reed | oh, no | 17:49 |
| clarkb | jeblair: I am at my battle station | 17:50 |
| reed | no commits on friday :) | 17:50 |
| clarkb | reed: what kind of stats? | 17:50 |
| reed | stuff like http://stats.kopweb.net/ | 17:50 |
| clarkb | the logs are pretty raw, youwould have to do processing on them I think | 17:50 |
| reed | clarkb, yes, there are tools for that :) | 17:50 |
| jeblair | reed: i don't see why not. that sounds like a good low-hanging-fruit bug for openstack-ci | 17:52 |
| reed | jeblair, will file a request, thanks | 17:53 |
| jog0 | mordred: this entry point stuff is funny | 17:53 |
| openstackgerrit | A change was merged to openstack-infra/devstack-gate: Grab testr logs even when testr fails. https://review.openstack.org/37776 | 17:55 |
| openstackgerrit | will soula proposed a change to openstack-infra/jenkins-job-builder: Adding the Sauce OnDemand Plugin https://review.openstack.org/37947 | 17:59 |
| *** sandywalsh has quit IRC | 18:01 | |
| reed | https://bugs.launchpad.net/openstack-ci/+bug/1203137 | 18:03 |
| clarkb | wow pisg is mean | 18:03 |
| jeblair | clarkb: ? | 18:03 |
| *** amotoki has quit IRC | 18:03 | |
| clarkb | I ran it really quickly over today's infra log and got "Is jeblair stupid or just asking too many questions? 33.3% lines contained a question!" | 18:04 |
| *** SergeyLukjanov has joined #openstack-infra | 18:04 | |
| pleia2 | poor jeblair | 18:04 |
| jeblair | clarkb: i'm not sure we want that? | 18:04 |
| annegentle_ | clarkb: lol | 18:05 |
| clarkb | jeblair: yeah I don't think we want that either. but pisg was the example reed provided | 18:05 |
| clarkb | jeblair: it also spat out "pleia2 brings happiness to the world. 13.3% lines contained smiling faces. :)" | 18:05 |
| *** pabelanger has quit IRC | 18:05 | |
| annegentle_ | clarkb: lol? | 18:05 |
| pleia2 | go me | 18:05 |
| jeblair | clarkb: maybe we should look at something else? | 18:05 |
| annegentle_ | oh I crack myself up | 18:05 |
| annegentle_ | jeblair: ROTFL | 18:05 |
| jeblair | clarkb: or maybe it has an option not to be rude? | 18:05 |
| clarkb | jeblair: I think we should. I will update the bug that pisg is a little harsh | 18:05 |
| clarkb | jeblair: oh maybe /me reads more docs | 18:05 |
| pleia2 | I think it's configurable for those things | 18:05 |
| pleia2 | (I use it for a few non-foss channels) | 18:05 |
| reed | I put superseriousstats on the bug report | 18:05 |
| reed | pisg seems lame | 18:06 |
| reed | and I hate perl | 18:06 |
| pleia2 | hah | 18:06 |
| clarkb | reed: but it supports supybot out of the box so you never have to touch the perl | 18:06 |
| pleia2 | pisg is pretty solid though, it has been running on one of my servers for years without messing with it (I forgot it existed) | 18:06 |
| *** arborism has joined #openstack-infra | 18:06 | |
| reed | clarkb, it remains lame-oh | 18:06 |
| reed | :) | 18:07 |
| *** melwitt has joined #openstack-infra | 18:07 | |
| *** pcrews has joined #openstack-infra | 18:07 | |
| reed | there is irssistats too | 18:07 |
| pleia2 | it is very community amusing-oriented though, with quotes and all, not just raw stats | 18:07 |
| reed | less lame than pisg, it seems | 18:07 |
| reed | http://royale.zerezo.com/irssistats/ | 18:08 |
| reed | http://royale.zerezo.com/irssistats/irssi.php | 18:09 |
| jeblair | i think zuul may need a restart to pick up this config correctyl. | 18:09 |
| *** nati_ueno has quit IRC | 18:09 | |
| clarkb | jeblair: because it is a function? | 18:10 |
| clarkb | I have updated the bug with notes on pisg so that we don't have mean stats | 18:11 |
| jeblair | clarkb: i think so | 18:11 |
| jeblair | clarkb: i think the reload method may need to reimport imported modules | 18:11 |
| *** nati_ueno has joined #openstack-infra | 18:11 | |
| clarkb | jeblair: I think you are correct | 18:11 |
| fungi | i use reload(modulename) for that, though haven't checked to see if that's outdated methodology | 18:13 |
| jeblair | i rechecked things that were in the check queue | 18:13 |
| jeblair | fungi: yeah, zuul uses that in unit tests | 18:13 |
| fungi | ahh, good to know i'm not doing something too stupid where i've used it then | 18:14 |
| jeblair | the currently running jobs have a reasonable looking BASE_LOG_PATH | 18:14 |
| *** sandywalsh has joined #openstack-infra | 18:14 | |
| jeblair | and i just checked one job in jenkins and it uses BASE_LOG_PATH in the scp section | 18:15 |
| *** linuxmohan has joined #openstack-infra | 18:15 | |
| clarkb | \o/ | 18:15 |
| clarkb | jeblair: this should make our filesystems much happier | 18:15 |
| clarkb | thank you for getting it done | 18:15 |
| mordred | w00t! | 18:15 |
| jeblair | clarkb: yep. 6 months from now we'll be able to run 'ls'. :) | 18:16 |
| * mordred hands jeblair a cookie | 18:16 | |
| jeblair | COOKIE | 18:16 |
| clarkb | I should keep an eye on the logstash gearman client | 18:16 |
| jeblair | clarkb: does that need to be restarted? | 18:16 |
| *** jjmb has quit IRC | 18:16 | |
| clarkb | jeblair: it does, I am going to do it now | 18:18 |
| clarkb | I will lose whatever logs are queued in the gearman server. but I can live with that | 18:18 |
| *** linuxmohan has quit IRC | 18:20 | |
| *** jjmb has joined #openstack-infra | 18:20 | |
| jeblair | i believe there is a bug in that the requirements jobs seem to be always running, ignoring the file matching pattern. | 18:20 |
| pleia2 | bah, my special class for jeepyb is messed up, looking into | 18:21 |
| jeblair | however, the correct jobs appear to be voting, which is a similar configuration construct. | 18:21 |
| cppcabrera | fungi: The only catch with using reload(module) is that this doesn't work in Python 3. It that case, you'd have to `import imp; imp.reload(module)` | 18:21 |
| cppcabrera | s/It/In | 18:21 |
| *** kgriffs has joined #openstack-infra | 18:21 | |
| *** kgriffs has left #openstack-infra | 18:22 | |
| sdague | fungi: so you have some time to talk about the log filter / color stuff on the log server? | 18:22 |
| fungi | cppcabrera: indeed, i've run into that already since i was using it in a py3k-supporting project | 18:22 |
| *** linuxmohan has joined #openstack-infra | 18:22 | |
| fungi | sdague: yeah, i started seeing if i could figure out a good conditional rewrite to apply to that, but a lot of it will have to do with what makes clarkb happy/sad | 18:22 |
| jeblair | sdague: aside: you mentioned mod_python the other day; why not use that? | 18:23 |
| sdague | no worries | 18:23 |
| sdague | jeblair: because I wasn't sure if it was an option | 18:23 |
| fungi | sdague: we could go with a shadow path to the same filenames, or a separate vhost even if we wanted, rather than dealing with rewrites too | 18:23 |
| sdague | if you are cool with installing it in that environment, I can take a run at that | 18:23 |
| sdague | I was trying to make this a least intrusive as possible | 18:23 |
| *** cp16net|away is now known as cp16net | 18:24 | |
| fungi | sdague: i don't see any reason to forbid mod_python on static.o.o, as long as we're careful that it doesn't burst into flames when loaded | 18:24 |
| pleia2 | jeblair: my fix re: your first comment https://review.openstack.org/#/c/37770/3/modules/openstack_project/manifests/git.pp was all wrong, which class did you mean? | 18:24 |
| sdague | fungi: ok, it will add substantial weight to all the apache process, but if the server can handle it, I'm cool with figuring out how to do it that way | 18:25 |
| sdague | it should make things more flexible for sure | 18:25 |
| jeblair | i've been mulling over the idea of having log storing and log retrieving apps. i don't have a fully formed plan yet, but i think both could be useful in the long run, and this could be a start of the second. | 18:25 |
| *** dprince has quit IRC | 18:25 | |
| fungi | jeblair: swift backend! | 18:26 |
| jeblair | (in particular, i'm thinking that apps can help with alternate storage backends, additionaly security (containers for different kinds of artifacts), validation of upload permissions, etc) | 18:26 |
| * fungi ducks | 18:26 | |
| jeblair | fungi: exactly | 18:26 |
| clarkb | logstash is currently in a weird place where jobs that are finishing started before the JJB config changes so they don't have the new parameter | 18:26 |
| *** linuxmohan has quit IRC | 18:26 | |
| jeblair | clarkb: oh, yeah. | 18:27 |
| clarkb | but once they flush out everything should be happy on the logstash side | 18:27 |
| *** linuxmohan has joined #openstack-infra | 18:27 | |
| jeblair | so i think the bug in zuul relates to multiple metajobs matching; i think there may be a bit of last-job-wins | 18:28 |
| *** nati_ueno_2 has joined #openstack-infra | 18:29 | |
| jeblair | we should fix this immediately as it could mean some job configuration options are wrong | 18:29 |
| * jeblair hacks | 18:30 | |
| *** pabelanger has joined #openstack-infra | 18:33 | |
| *** krtaylor has quit IRC | 18:35 | |
| devananda | hi! anything particularly off with jenkins today? | 18:36 |
| mordred | devananda: there are some things that are being worked through right now | 18:38 |
| devananda | k | 18:38 |
| devananda | https://review.openstack.org/#/c/37763/ seems permanently LOST, but i've seen a few other things temporarily LOST recently too | 18:38 |
| openstackgerrit | Clark Boylan proposed a change to openstack-infra/config: Properly root log path in logstash gearman client. https://review.openstack.org/37953 | 18:39 |
| clarkb | ^ that fixes one problem with the logstash + new log path stuff | 18:39 |
| *** hashar has joined #openstack-infra | 18:39 | |
| clarkb | that is the only problem I currently see. I have manually applied that and disabled puppet. I will probably just push it thorugh after lunch if that hasn't happened sooner | 18:39 |
| notmyname | jeblair: do you have a twitter account? | 18:41 |
| jeblair | notmyname: nope | 18:41 |
| clarkb | devananda: we had a huge backlog of changes that couldn't go in around H2 and now that H2 is over the floodgates have opened | 18:42 |
| notmyname | jeblair: k. my searches were getting odd results (https://twitter.com/Blairgun) | 18:42 |
| notmyname | jeblair: ;-) | 18:42 |
| openstackgerrit | Elizabeth Krumbach Joseph proposed a change to openstack-infra/config: Fix jeepyb class issue on git.o.o https://review.openstack.org/37954 | 18:42 |
| *** boris-42 has quit IRC | 18:42 | |
| clarkb | devananda: these are good things that should make the gate and logging and stuff scale further | 18:42 |
| devananda | clarkb: awesome. yea, i've noticed you guys working on those things :) | 18:43 |
| clarkb | sdague: mtreinish did the stop logging glance sqlalchemy at DEBUG level change ever go through? | 18:43 |
| openstackgerrit | James E. Blair proposed a change to openstack-infra/zuul: Apply metajobs in order https://review.openstack.org/37955 | 18:45 |
| jeblair | clarkb, fungi, mordred: ^ | 18:45 |
| mordred | jeblair: looking | 18:45 |
| mordred | jeblair: looks reasonable to me | 18:46 |
| clarkb | reviewing | 18:47 |
| *** hashar has quit IRC | 18:48 | |
| fungi | lgtm | 18:48 |
| fungi | we need another quick outage as soon as we merge that, right? | 18:48 |
| jog0 | I have a strange question about copyright | 18:48 |
| jog0 | https://review.openstack.org/#/c/35380/7/nova/tests/virt/xenapi/test_driver.py | 18:48 |
| jog0 | see line 4 | 18:48 |
| jog0 | is that allowed? | 18:48 |
| jeblair | notmyname: i'm also not https://twitter.com/jeblair who totally gets what twitter is for. | 18:49 |
| clarkb | jeblair: +2'd will let you approve and control the push process | 18:49 |
| notmyname | jeblair: lol | 18:49 |
| clarkb | jog0: I am not a lawyer, but yes I think it is allowed. Open Source licenses don't invalidate copyright they depend on it | 18:51 |
| clarkb | jog0: so asserting that all rights are reserved by the copyright holder is fine | 18:51 |
| jeblair | jog0: it is... confusing. | 18:51 |
| jog0 | clarkb: they can reserve all rights doesn't conflict with open sourcing? | 18:51 |
| clarkb | jog0: no, because open sourcign depends on strong copyright... | 18:51 |
| jeblair | jog0: because it is clearly falsified by the next statements which declares that some rights, in fact, are not reserved. | 18:52 |
| mtreinish | clarkb: nope | 18:52 |
| clarkb | jeblair: apache2 doesn't undermine copyright rights though which that line refers to | 18:52 |
| jeblair | however, some licenses have that in their standard text | 18:52 |
| mtreinish | clarkb: https://review.openstack.org/#/c/37123/ still sitting there with 3 +1s | 18:52 |
| jeblair | eg, bsd. | 18:52 |
| clarkb | mtreinish: damn | 18:52 |
| openstackgerrit | Elizabeth Krumbach Joseph proposed a change to openstack-infra/config: Fix jeepyb class issue on git.o.o https://review.openstack.org/37954 | 18:53 |
| jog0 | jeblair: clarkb any links explaining this | 18:53 |
| jog0 | that I can defer to | 18:53 |
| fungi | jog0: it was a phrase required by the buenos aires convention and later obsoleted by the berne convention | 18:54 |
| jog0 | fungi: that means nothing to me | 18:54 |
| jog0 | :/ | 18:54 |
| * fungi gets international copyright reference out real quick | 18:54 | |
| mordred | jog0: http://en.wikipedia.org/wiki/All_rights_reserved | 18:55 |
| fungi | or wikipedia ftw | 18:55 |
| jeblair | jog0: i think the short version of what fungi is saying is that it doesn't matter, it's sort of a no-op. combine that with the fact that some floss licenses have it, and some don't, i think it's "mostly harmless". | 18:55 |
| mordred | "The requirement to add this notice became obsolete and essentially deprecated on August 23, 2000 when Nicaragua became the final member of the Buenos Aires Convention to also become a signatory to the Berne Convention. As of that date, every country that was a member of the Buenos Aires Convention (which is the only copyright treaty requiring this notice to be used) was also a member of Berne, which | 18:55 |
| mordred | requires protection be granted without any formality of notice of copyright." | 18:55 |
| jeblair | it is not in the standard apache license, nor is it in fsf licenses; bsd yes, and some others. | 18:56 |
| sdague | yeh, it doesn't really matter if the apache header is in the file | 18:56 |
| jog0 | ahh so it means nothing | 18:56 |
| jog0 | thanks | 18:56 |
| mordred | sdague: that's not the same discussion | 18:56 |
| jeblair | mordred: i think you had a parse error | 18:56 |
| sdague | mordred: yes, sorry, I think it was a parse error | 18:57 |
| jeblair | mordred: i think sdague said "as long as the apache header is in the file, also having ARR doesn't really matter" | 18:57 |
| sdague | correct | 18:57 |
| sdague | that's what I meant at least | 18:57 |
| mordred | ah. I thought he was bringing up the "do we need the headers in each file" thing again. :) | 18:57 |
| jeblair | (as compared to having ARR in the file and not having the apache header, which would be a bad situation) | 18:57 |
| sdague | mordred: no way, I just landed H102 | 18:57 |
| mordred | sdague: w00t | 18:57 |
| sdague | to enforce it | 18:58 |
| clarkb | jeblair: very bad | 18:58 |
| *** arborism has quit IRC | 18:58 | |
| *** ewindisch has quit IRC | 18:58 | |
| sdague | omg - http://status.openstack.org/zuul/ the devstack gate is empty | 18:59 |
| fungi | on purpose i hope | 18:59 |
| sdague | yeh, I guess it's the day of rest after H2 | 18:59 |
| *** krtaylor has joined #openstack-infra | 18:59 | |
| jeblair | yeah, i don't even think that's an error. :) | 18:59 |
| jeblair | well, it's break-everything day around here. :) | 18:59 |
| sdague | ok, fixed that. (sent in a devstack change) | 19:00 |
| sdague | something just doesn't feel right to me if that queue is empty | 19:00 |
| clarkb | I am going to grab lunch. Will pay attention to IRC so ping if things go further sideways | 19:00 |
| *** arborism has joined #openstack-infra | 19:01 | |
| openstackgerrit | A change was merged to openstack-infra/zuul: Apply metajobs in order https://review.openstack.org/37955 | 19:03 |
| *** hashar has joined #openstack-infra | 19:04 | |
| *** hashar has quit IRC | 19:05 | |
| *** hashar has joined #openstack-infra | 19:05 | |
| *** linuxmohan has quit IRC | 19:05 | |
| *** ewindisch has joined #openstack-infra | 19:07 | |
| *** anteaya has quit IRC | 19:18 | |
| *** nati_ueno has quit IRC | 19:21 | |
| *** nati_ueno has joined #openstack-infra | 19:22 | |
| *** sarob_ has quit IRC | 19:25 | |
| *** sarob has joined #openstack-infra | 19:26 | |
| *** sarob has quit IRC | 19:30 | |
| jeblair | restarted zuul | 19:31 |
| jeblair | rechecked/reverified | 19:31 |
| jeblair | the requirements jobs seem to be correctly not running this time | 19:32 |
| openstackgerrit | will soula proposed a change to openstack-infra/jenkins-job-builder: Adding the Sauce OnDemand Plugin https://review.openstack.org/37947 | 19:40 |
| *** mikal has quit IRC | 19:41 | |
| *** SergeyLukjanov has quit IRC | 19:41 | |
| *** vipul is now known as vipul-away | 19:41 | |
| *** SergeyLukjanov has joined #openstack-infra | 19:43 | |
| *** mikal has joined #openstack-infra | 19:43 | |
| *** ladquin is now known as ladquin_afk | 19:45 | |
| fungi | too awesome | 19:47 |
| *** vipul-away is now known as vipul | 19:52 | |
| *** sarob has joined #openstack-infra | 19:53 | |
| *** cppcabrera is now known as cppcabrera_afk | 19:53 | |
| jeblair | https://review.openstack.org/#/c/37329/ | 19:59 |
| jeblair | This change was unable to be automatically merged with the current state of the repository and the following changes which were enqueued ahead of it: 37207. Please rebase your change and upload a new patchset. | 19:59 |
| jeblair | first change to hit the new check! | 20:00 |
| sdague | sweet | 20:00 |
| fungi | i can't wait to see how much that improves gate throughput/parallelism | 20:02 |
| sdague | yeh, on busy days that's going to be huge | 20:02 |
| sdague | jeblair: did that also fix the kick out on first fail issue? | 20:02 |
| *** ewindisch has quit IRC | 20:03 | |
| jeblair | sdague: i believe that was fixed in the scheduler refactoring; i'm pretty sure we've spotted changes since being correctly being severed afteir their first job fails. | 20:03 |
| sdague | ok cool | 20:03 |
| *** ewindisch has joined #openstack-infra | 20:07 | |
| *** sarob has quit IRC | 20:11 | |
| *** sarob has joined #openstack-infra | 20:12 | |
| *** changbl has quit IRC | 20:12 | |
| *** linuxmohan has joined #openstack-infra | 20:16 | |
| *** sarob has quit IRC | 20:16 | |
| openstackgerrit | A change was merged to openstack-dev/hacking: Add information about the Google style guide https://review.openstack.org/36997 | 20:17 |
| *** plomakin has quit IRC | 20:18 | |
| *** plomakin has joined #openstack-infra | 20:19 | |
| *** linuxmohan has quit IRC | 20:22 | |
| openstackgerrit | A change was merged to openstack-infra/config: Add a node pattern for precise3k slaves. https://review.openstack.org/37609 | 20:24 |
| pleia2 | so, using apache::vhost is the way we should do things, but cgit doesn't strictly have a docroot (it's a cgi script) so I'll have to take time to look into what I should populate it with | 20:24 |
| pleia2 | but right now I need to head over to the airport | 20:25 |
| lifeless | ttx: mordred: which guys? | 20:25 |
| jeblair | pleia2: you can probably just set it to an empty dir in /var/www/something | 20:25 |
| jeblair | pleia2: have a nice flight! | 20:25 |
| pleia2 | jeblair: noted, thanks! | 20:26 |
| SlickNik | hey guys | 20:26 |
| clarkb | ingy says duplicate keys in yaml should be an error | 20:26 |
| SlickNik | What's the status of https://review.openstack.org/#/c/37903/ | 20:26 |
| clarkb | also .yaml is preferred to .yml | 20:26 |
| SlickNik | In the review it says: Starting gate jobs. http://status.openstack.org/zuul/ | 20:26 |
| clarkb | pleia2: see what we did for gerrit and jenkins | 20:27 |
| clarkb | pleia2: mordred passes in something like "INVALID ARGUMENT" | 20:27 |
| SlickNik | But if I look at that URL, there is no mention of 37903. | 20:27 |
| SlickNik | Any ideas what's going on? | 20:27 |
| jeblair | SlickNik: then it probably got lost in a recent zuul restart; 'reverify no bug' | 20:27 |
| jeblair | clarkb: "ingy says duplicate keys in yaml should be an error" context? | 20:28 |
| SlickNik | jeblair: Thanks for the update. Done! | 20:28 |
| clarkb | jeblair: pyyaml and consequently JJB allow dupliacate keys in your yaml and I think first one wins. eg if you define a template with the same name twice. Ingy was just in the office giving a talk on stuff and afterwards I asked him what the proper behavior should be and that is to be an error | 20:28 |
| jeblair | clarkb: *nod* | 20:29 |
| lifeless | mordred: clarkb: whats the account name for the infra HPCloud account? | 20:31 |
| lifeless | e.g. the one you want glance enabled on | 20:32 |
| clarkb | lifeless: openstackjenkins2 and openstackci2 but the second one needs mordred to bless it with not costing my credit card money | 20:33 |
| lifeless | openstackjenkins2 ack | 20:33 |
| clarkb | lifeless: in other news I Think I saw it mentioned that the double cuonting of tests doesn't affect just tempest | 20:34 |
| lifeless | ack | 20:35 |
| lifeless | will look into it | 20:35 |
| *** ewindisch has quit IRC | 20:36 | |
| lifeless | clarkb: or should I ask for both ? | 20:37 |
| lifeless | clarkb: note that HP Cloud's glance only does one-user or all-user images at the moment | 20:38 |
| lifeless | clarkb: so if we enabled openstackjenkins2 images uploaded there won't be visible to openstackci2 | 20:38 |
| lifeless | clarkb: looking for an ack - one or both ? | 20:38 |
| clarkb | lifeless: will enabling it break our ability to take snapshots and use hp provided images? | 20:39 |
| clarkb | lifeless: devstack-gate depends on those services so we don't want to flip the switch if we will berak that | 20:40 |
| clarkb | http://engineering.cloudscaling.com/happy-3rd-birthday-openstack/ | 20:40 |
| clarkb | I don't know how tripleo got in there but I will take the praise :) | 20:40 |
| lifeless | clarkb: no | 20:40 |
| clarkb | jeblair: fungi do you guys have an opinion on where we want to enable it? | 20:41 |
| clarkb | I am leaning towards both but lets see what they have to say | 20:41 |
| * fungi reads back up to see what "it"is | 20:41 | |
| lifeless | glance upload | 20:41 |
| jog0 | why don't we have a sqlalchemy-migrate launchpad thingy yet? | 20:41 |
| jog0 | for bug tracking etc | 20:41 |
| lifeless | HP cloud can permit end users to use it | 20:41 |
| lifeless | but you need to submit a support request | 20:41 |
| lifeless | and some internal foo | 20:42 |
| jeblair | clarkb, lifeless: sounds great; why? | 20:42 |
| lifeless | I've been chasing the thread to understand it | 20:42 |
| jog0 | clarkb: haha | 20:42 |
| lifeless | jeblair: tripleo gating will be massively easier if we can use the dib workflow directly. | 20:42 |
| lifeless | jeblair: which is glance based | 20:42 |
| jeblair | lifeless: is it also being enabled on rackspace? | 20:42 |
| *** hashar has quit IRC | 20:42 | |
| *** sarob has joined #openstack-infra | 20:42 | |
| lifeless | jeblair: Alex_Gaynor is chasing that | 20:42 |
| jog0 | jeblair mordred: ^, bug tracking for sqla-migrate https://review.openstack.org/#/c/37656/ | 20:43 |
| lifeless | jeblair: HPCloud support told me they plan to open glance up broadly eventually but no ETA yet | 20:43 |
| clarkb | the nova unittests over time graph is nice because that wouldn't be possible without testr | 20:43 |
| fungi | clarkb: if the question is whether we want an openstack feature available to all our accounts on one of our openstack-based public cloud providers, i'm pretty sure the answer is ye | 20:44 |
| fungi | s | 20:44 |
| fungi | (free newline for the taking) | 20:44 |
| clarkb | I will treasure that byte | 20:45 |
| jeblair | lifeless: ok, cool. i agree with clark: both, but disruption would be bad. and at the risk of reiterating something that may have been said already -- i'm not comfortable using it in production until we can use it globally | 20:45 |
| fungi | the safety factor could be mitigated by initially enabling it for openstackci2 and testing what we want there while it's not being used yet, then adding it for openstackjenkins2 once we think it checks out | 20:46 |
| lifeless | jeblair: so, I don't know how long the lead time is. I'd rather ask for it and not use it for a bit (or only use it for testing whether a new job works?) | 20:46 |
| lifeless | than not ask for it until its available everywhere. | 20:46 |
| jeblair | lifeless: yeah, i'm totally cool with getting it in there and starting to work with it. | 20:46 |
| *** plomakin has quit IRC | 20:47 | |
| lifeless | ok | 20:47 |
| jeblair | lifeless: i would personally love that feature. just want to be clear. :) | 20:47 |
| *** cp16net is now known as cp16net|away | 20:47 | |
| *** plomakin has joined #openstack-infra | 20:47 | |
| *** cp16net|away is now known as cp16net | 20:48 | |
| lifeless | jeblair: mordred: you have mail now about this. | 20:50 |
| *** sarob has quit IRC | 20:50 | |
| jeblair | fungi, clarkb, mordred: i'm going to get certs for review, jenkins, etherpad, wiki. | 20:52 |
| fungi | jeblair: great. i know we're down to about a week left now | 20:53 |
| jeblair | (those are the ones that expire in the next 6 months -- i'd like to align their renewals for efficiency (and safety)) | 20:53 |
| fungi | on a couple of them at least | 20:53 |
| openstackgerrit | Antoine Musso proposed a change to openstack-infra/storyboard: gitignore /*.egg https://review.openstack.org/37969 | 20:53 |
| jeblair | should i also get one for git.o.o? | 20:53 |
| clarkb | jeblair: ++ I was going to suggest that | 20:53 |
| jeblair | clarkb: suggest alignment or git? | 20:53 |
| fungi | jeblair: i think it would be awesome to do https on git.o.o | 20:53 |
| clarkb | jeblair: sure? though I foresee that git:// will be the popular protocol there which won't be secure in any way | 20:53 |
| clarkb | jeblair: alignment | 20:54 |
| jeblair | true. kernel.org's cgit service is behind https | 20:54 |
| jeblair | lifeless: thanks | 20:54 |
| pabelanger | git.o.o? Are back up for github or migrating everything to it? | 20:54 |
| fungi | pabelanger: our github presence is merely a mirror. this will be another, but more officially openstack, mirror | 20:55 |
| jeblair | pabelanger: we're going to refer to it in all our docs and scripts; we'll continue to mirror to github because it's useful for a lot of ppl. | 20:55 |
| fungi | pabelanger: afaik github mirrors will also remain in place | 20:55 |
| fungi | what jeblair said ;) | 20:55 |
| jeblair | 3 coremudgeons agree. :) | 20:55 |
| pabelanger | True, gerrit would be the official I guess, everything a mirror :) | 20:56 |
| fungi | jeblair: that is a strict majority. so be it | 20:56 |
| clarkb | pabelanger: it will allow us to host our own git daemon to use the git protocol | 20:56 |
| fungi | pabelanger: yep, we just don't want to point the entire world at gerrit for cloning/pulling because... java | 20:56 |
| clarkb | pabelanger: which will hopefully speed up some of our testing | 20:56 |
| jeblair | fungi, clarkb: 2 year 2048? | 20:57 |
| fungi | jeblair: sounds fine to me | 20:57 |
| *** sarob has joined #openstack-infra | 20:58 | |
| fungi | anything >1024 is safe for quite a while most likely | 20:58 |
| clarkb | jeblair: sounds good | 20:58 |
| fungi | modulo new and interesting ways to efficiently factor very large numbers | 20:58 |
| clarkb | fungi: you better be knocking on wood | 20:59 |
| pabelanger | clarkb: cool, does that mean by passing gerrit actually running the git server? | 20:59 |
| *** mgagne has quit IRC | 20:59 | |
| fungi | clarkb: i for one welcome our new factorizing masters | 20:59 |
| pabelanger | Trying to understand where the performance impact is | 20:59 |
| clarkb | pabelanger: we already bypass gerrit in most places and hit an http mirror so yes and yes | 20:59 |
| pabelanger | I see | 20:59 |
| clarkb | pabelanger: git over http is much faster now than it used to be but it is still the slowest protocol behind ssh and git | 20:59 |
| annegentle_ | does any other project in OpenStack delegate approvers for a particular directory within a repo? Or do they just start a new repo with "core" reviewers? Asking on behalf of the training manuals idea-guys. | 20:59 |
| clarkb | pabelanger: so the performance benefit is in talking a more efficient protocol when updating workspaces for jobs and so on | 21:00 |
| fungi | pabelanger: in many ways, this is just us moving the local apache-based mirror we already run on our gerrit server today to a separate host with a distinct name and nifty additional features | 21:00 |
| clarkb | annegentle_: I think that feature has been requested but gerrit doesn't support it directly | 21:00 |
| clarkb | annegentle_: instead these things are managed by social convention | 21:00 |
| pabelanger | clarkb: fungi: Cool! Need to logoff now, but will chat more later | 21:01 |
| pabelanger | cya | 21:01 |
| *** pabelanger has quit IRC | 21:01 | |
| fungi | annegentle_: in places where we need to ensure that, we go with separate repositories/projects with distinct groups | 21:01 |
| fungi | annegentle_: but i agree with clarkb, letting social pressures dictate that sort of thing leads to improved harmony amongst devs in most cases | 21:02 |
| openstackgerrit | Antoine Musso proposed a change to openstack-infra/storyboard: projects view now accepts projects with spaces https://review.openstack.org/37977 | 21:03 |
| *** ewindisch has joined #openstack-infra | 21:03 | |
| clarkb | I am going to try doing another code review pass before the weekend | 21:03 |
| clarkb | anything in particular folks want me to not miss? | 21:04 |
| fungi | clarkb: nothing specific from me | 21:04 |
| *** rcleere has quit IRC | 21:05 | |
| fungi | i'm heading out to grab some dinner in a bit, then coming back to add precise3k slaves and confirm pbr/oslo.config unit tests on them | 21:06 |
| openstackgerrit | A change was merged to openstack-infra/config: Properly root log path in logstash gearman client. https://review.openstack.org/37953 | 21:06 |
| openstackgerrit | Clark Boylan proposed a change to openstack-infra/jenkins-job-builder: Tests: Use a testscenarios for running tests. https://review.openstack.org/36990 | 21:08 |
| clarkb | fungi: jeblair: I am going to readvertise the series here https://review.openstack.org/#/c/14893 ^ should make them both mergable | 21:09 |
| lifeless | jeblair: fungi: clarkb: so we need to be in the us-east beta region for glance access. | 21:09 |
| lifeless | thats just another region which you can activate via the console | 21:09 |
| clarkb | lifeless: I don't think we want beta | 21:09 |
| clarkb | lifeless: oh | 21:09 |
| clarkb | lifeless: well, we want it to homogenous across the AZs but I suppose we can go ahead and enable it in one for testing | 21:10 |
| lifeless | right | 21:10 |
| clarkb | jeblair: fungi ^ sane? | 21:10 |
| fungi | so we can try out glance in us-east on our accounts, sounds like. which means it probably also is unlikely to break our production zones anyway. wfm | 21:10 |
| jeblair | clarkb, lifeless: what's the correlation between AZs and regions? | 21:10 |
| jeblair | "az1, az2, az3, us-east": one of these is not like the others. | 21:10 |
| lifeless | jeblair: us-west is the current hpcloud region | 21:11 |
| lifeless | https://account.hpcloud.com/services | 21:11 |
| clarkb | jeblair: az1 az2 and az3 are in us-west | 21:11 |
| jeblair | k, so az1-3 are in us-west | 21:11 |
| lifeless | US West AZ 1 | 21:11 |
| clarkb | jeblair: regions house AZs | 21:11 |
| jeblair | what azs are in us-east-beta? | 21:11 |
| lifeless | compute 13.5 is in us-east, and may have a different region/az layout | 21:11 |
| lifeless | jeblair: us-east | 21:11 |
| lifeless | jeblair: there is no us-east non-beta az atm | 21:12 |
| jeblair | ok. sounds nicely separate from everything we're doing now; enabling it sounds sane. | 21:12 |
| lifeless | cool | 21:12 |
| *** dirk has joined #openstack-infra | 21:12 | |
| lifeless | I suspect you may need to request access to us-east from both accounts before the glance request can go in. | 21:12 |
| jeblair | clarkb: want to take care of that? | 21:13 |
| clarkb | ya, I will do that now | 21:13 |
| clarkb | lifeless: there is no option to do this in the web console | 21:15 |
| lifeless | clarkb: https://account.hpcloud.com/services | 21:15 |
| lifeless | clarkb: bottom of the page 'Beta Services' | 21:15 |
| lifeless | clarkb: Compute US East v13.5 not activated Request Access | 21:15 |
| clarkb | gah why is it under beta | 21:15 |
| clarkb | put it in compute and note it is beta | 21:15 |
| lifeless | clarkb: possibly because naive users click on things without understanding | 21:16 |
| openstackgerrit | Dirk Mueller proposed a change to openstack-dev/hacking: Don't crash on tests when git is not installed https://review.openstack.org/37870 | 21:16 |
| lifeless | clarkb: possibly because they think naive users ... | 21:16 |
| lifeless | clarkb: who knows :> | 21:16 |
| *** ewindisch has quit IRC | 21:16 | |
| jog0 | mordred: ping on entrypoints | 21:17 |
| fungi | i can confirm naive users *definitely* click on things without understanding | 21:18 |
| *** cp16net is now known as cp16net|away | 21:18 | |
| *** ewindisch has joined #openstack-infra | 21:19 | |
| fungi | every time i check to see whether any users have tried to add the internal system cla and usg cla in gerrit, in deference to the very shouty warnings on and in them, i find a handful | 21:19 |
| *** dhellmann is now known as dhellmann_ | 21:19 | |
| clarkb | lifeless: requests have been submitted for accounts and I have noted the desire to use glance in the requests | 21:19 |
| fungi | though i think in that case it's mostly people with a very tenuous grasp of english getting desperate when they start having issues with the icla | 21:20 |
| *** krtaylor has quit IRC | 21:21 | |
| lifeless | clarkb: cool | 21:21 |
| * fungi dines... bbl | 21:21 | |
| lifeless | clarkb: the request for glance is in, I believe. jeblair and mordred are being copied on everythnig | 21:22 |
| clarkb | lifeless: perfect, thank you | 21:22 |
| lifeless | is there an alias for 'this gets to -infra' that I could use rather than a cc-list ? | 21:22 |
| clarkb | lifeless: not really | 21:22 |
| clarkb | there is the infra maillist but it requires membership | 21:22 |
| lifeless | just saying:> | 21:23 |
| clarkb | pleia2: does your plane have wifi? | 21:27 |
| clarkb | pleia2: I have questions | 21:27 |
| *** rnirmal has quit IRC | 21:27 | |
| clarkb | meh I will leave them in the review | 21:27 |
| clarkb | jeblair: any particular reason for not approving https://review.openstack.org/#/c/35360/3 ? (maybe the soft freeze) | 21:31 |
| *** mrodden has quit IRC | 21:31 | |
| reed | do you know what a deployment advisor is? https://ask.openstack.org/question/3232/deployment-advisor-in-openstack/ | 21:31 |
| jeblair | clarkb: so you and fungi could look at it | 21:32 |
| clarkb | jeblair: ok, I will let fungi eye it over after dinner | 21:33 |
| *** sarob has quit IRC | 21:33 | |
| *** sarob has joined #openstack-infra | 21:33 | |
| Shrews | mordred: or any pbr guys... this look familiar to anyone? http://paste.openstack.org/show/40967/ | 21:34 |
| *** ewindisch has quit IRC | 21:35 | |
| clarkb | Shrews: ya, I think that is a case of needing latest setupttools/distribute | 21:36 |
| *** sarob has quit IRC | 21:36 | |
| clarkb | Shrews: the merging of those two projects has caused some headaches and upgrading smooths it out | 21:36 |
| *** sarob_ has joined #openstack-infra | 21:36 | |
| *** Ryan_Lane has quit IRC | 21:36 | |
| pleia2 | clarkb: hey, I have a half hour before boarding | 21:37 |
| Shrews | clarkb: setuptools was upgraded in an attempt to fix it. should distribute be installed as well? | 21:37 |
| pleia2 | clarkb: yeah, so they are required for one of the .erb files that jeepyb populates | 21:37 |
| clarkb | Shrews: they should be the same at this point, SpamapS mentioned that he had to remove the apt installed version of setuptools too | 21:38 |
| clarkb | pleia2: thanks | 21:38 |
| jeblair | can't we just define them inside the class? | 21:38 |
| pleia2 | jeblair: what class? | 21:38 |
| jeblair | as in, if they aren't used, why do they need to be class parameters? | 21:38 |
| jeblair | pleia2: openstack_project::git | 21:39 |
| clarkb | jeblair: we can do that | 21:39 |
| LinuxJedi | Shrews: just tried upgrading distribute too, still not working :( | 21:39 |
| *** vipul is now known as vipul-away | 21:41 | |
| *** avtar has quit IRC | 21:41 | |
| jog0 | Shrews: I think I did a rm -rf .tox and upgraded pip and distribute etc to the latest and reran | 21:41 |
| clarkb | jeblair: fungi mordred I am writing up a list of changes that I plan on merging this afternoon if no additional reviews are posted https://etherpad.openstack.org/infra-reviews | 21:42 |
| clarkb | *no additional negative reviews | 21:42 |
| *** ewindisch has joined #openstack-infra | 21:42 | |
| *** _TheDodd_ has quit IRC | 21:42 | |
| LinuxJedi | jog0: this is on a deployment, no .tox | 21:43 |
| *** lcestari has quit IRC | 21:43 | |
| jeblair | clarkb: lgtm | 21:43 |
| *** gyee has quit IRC | 21:43 | |
| *** mriedem has quit IRC | 21:43 | |
| jog0 | LinuxJedi: ouch | 21:44 |
| LinuxJedi | Shrews: figured out a fix, but it is sort of dangerous | 21:45 |
| jog0 | anyone ex Drizzlers here at the moment? | 21:45 |
| LinuxJedi | jog0: me, Shrews, mordred, pcrews | 21:45 |
| Shrews | 3 of those are quite busy, atm | 21:46 |
| jog0 | so the sqlite migration snafu | 21:46 |
| jog0 | since someone stuck me as a sqlalchemy-migrate maintainer I started digging into this one | 21:46 |
| *** vipul-away is now known as vipul | 21:47 | |
| LinuxJedi | jog0: we are in the middle of a huge deployment, give us a couple of hours | 21:47 |
| jog0 | LinuxJedi: sure | 21:47 |
| jog0 | the TL;DR is can we actaully replace sqlite with drizzle as our testing backend | 21:47 |
| jog0 | but it can wait till monday | 21:47 |
| openstackgerrit | Elizabeth Krumbach Joseph proposed a change to openstack-infra/config: Fix jeepyb class issue on git.o.o https://review.openstack.org/37954 | 21:54 |
| *** odyssey4me has quit IRC | 21:54 | |
| *** ewindisch has quit IRC | 21:55 | |
| *** koolhead17 has quit IRC | 21:56 | |
| *** sarob_ has quit IRC | 21:57 | |
| *** koolhead17 has joined #openstack-infra | 21:57 | |
| openstackgerrit | Sergey Lukjanov proposed a change to openstack-infra/config: Add more jobs for Savanna projects https://review.openstack.org/37987 | 21:58 |
| *** sarob has joined #openstack-infra | 21:58 | |
| clarkb | https://pypi.python.org/pypi/xattr/0.7.0 is a new version that appears to have broken the gate | 22:01 |
| clarkb | sdague: jeblair fungi ^ | 22:01 |
| clarkb | not sure how that snuck paste the mirror jobs | 22:02 |
| clarkb | s/paste/past/ | 22:02 |
| *** vipul is now known as vipul-away | 22:02 | |
| *** sarob has quit IRC | 22:02 | |
| clarkb | well thats unfortunate https://jenkins.openstack.org/job/periodic-mirror-python27/140/console | 22:03 |
| *** bknudson has joined #openstack-infra | 22:04 | |
| * clarkb attempts a fix | 22:05 | |
| mordred | jog0: yes. I agree. I thnk it will be easy | 22:06 |
| *** cp16net|away is now known as cp16net | 22:06 | |
| jog0 | mordred: got another q for you: https://pypi.python.org/pypi/zc.buildout/2.2.0 | 22:07 |
| jog0 | that came recomended as a what to try instead of entry_points via pkg_resources | 22:07 |
| jog0 | which slow down the more things are in python | 22:07 |
| mordred | oh god. no | 22:08 |
| jog0 | so the question is have you ever heard of / used that? | 22:08 |
| mordred | we are not switching to a completely new build and packaging system | 22:08 |
| mordred | :) | 22:08 |
| clarkb | mordred: jeblair why do C libs that python pacakges need to compile against not get isntalled by puppet on the d-g nodes? | 22:08 |
| jog0 | running 'time python -c "import pkg_resources"' on a devstack instance takes 0.2 seconds or so | 22:08 |
| mordred | clarkb: because we expect devstack to request them | 22:08 |
| clarkb | mordred: jeblair: I think the fix for xattr here is to make sure that libffi-dev is installed on the nodes and I am trying to figure out if it needs to go on the d-g hosts too | 22:08 |
| mordred | jog0: grok proble. I'd rather go back to hand-coded bin/ scripts | 22:09 |
| jog0 | in tripleo it takes .08 seconds in the bigger venvs | 22:09 |
| clarkb | mordred: ok, so devstack will need to be patched independently if necessar | 22:09 |
| mordred | than switch to zc.buildout | 22:09 |
| jog0 | mordred: thats what I figured | 22:09 |
| mordred | jog0: zc.buildout is a parallel system to pip/virtualenv | 22:09 |
| mordred | it would change EVERYTHING | 22:09 |
| jog0 | ohhh hahaha | 22:09 |
| *** krtaylor has joined #openstack-infra | 22:10 | |
| openstackgerrit | Clark Boylan proposed a change to openstack-infra/config: Install libffi-dev on our slaves. https://review.openstack.org/37990 | 22:10 |
| clarkb | mordred: jeblair fungi ^ once that is merged and applied the mirror script will need to be rewritten | 22:10 |
| jog0 | mordred: I included you on the thread that lifeless started about this | 22:10 |
| clarkb | mordred: jeblair fungi also why is the mirror script uploading stuff if it fails to install packages? I thought it was supposed to protect us from that | 22:10 |
| lifeless | jog0: ? | 22:11 |
| * clarkb investigates that problem | 22:11 | |
| lifeless | mordred: what project were you referring to above btw ? | 22:11 |
| jog0 | lifeless: the rootwrap thing | 22:11 |
| lifeless | jog0: ah yes | 22:11 |
| mordred | clarkb: can you describe the problem you are seeing a little more? | 22:12 |
| clarkb | mordred: https://jenkins.openstack.org/job/gate-keystone-pep8/6418/console is happening because https://jenkins.openstack.org/job/periodic-mirror-python27/140/console did not properly fail | 22:12 |
| lifeless | mordred: ttx: to be clear, I"m not talking about rootwrap, something else - about 5 horus back | 22:13 |
| mordred | clarkb: ok. so we didn't successfully install libffi | 22:13 |
| lifeless | you say I could get a release 'of that lib' | 22:13 |
| mordred | on the mirror builder | 22:13 |
| lifeless | but not what lib that is. | 22:13 |
| mordred | but then we uploaded libffi to our mirror? | 22:14 |
| clarkb | mordred: correct, but we should never have rsync'd xattr to the mirror | 22:14 |
| clarkb | mordred: no we uploaded xattr to the mirror which had dependencies that failed to install | 22:14 |
| mordred | AH | 22:14 |
| mordred | weird | 22:14 |
| clarkb | the dependencies failed to install because libffi-dev was not present | 22:14 |
| mordred | how did xattr get ... oh jeez | 22:14 |
| * mordred groks the logic error | 22:14 | |
| clarkb | mordred: so immediate fix is the one I have proposed above | 22:14 |
| clarkb | but we need to fix the mirror jobs too | 22:14 |
| mordred | well, it's not going to be a tiny change | 22:15 |
| clarkb | :/ | 22:15 |
| clarkb | maybe I will let it slide until later | 22:15 |
| mordred | not HUGE - but not a one liner | 22:15 |
| mordred | file a bug or whatnot - it's important | 22:15 |
| clarkb | will do | 22:15 |
| mordred | but yeah, right now, workaround is a good idea | 22:15 |
| clarkb | mordred: can I get reviews on that workaround? | 22:15 |
| pleia2 | ok, plane now, hopefully that change i made gets things sorted | 22:15 |
| mordred | +2'd | 22:15 |
| *** ladquin_afk has quit IRC | 22:16 | |
| mordred | clarkb: because we process the pip cache... | 22:16 |
| mordred | clarkb: things will get downloaded | 22:16 |
| mordred | into the cache | 22:16 |
| *** linuxmohan has joined #openstack-infra | 22:16 | |
| mordred | oh wait. | 22:16 |
| mordred | no, that's not right | 22:16 |
| clarkb | mordred: is that cache persistent? it is right? | 22:16 |
| mordred | the double download/freeze cycle should handle this | 22:17 |
| mordred | jog0: where's the thread lifeless started about this? | 22:18 |
| *** Ryan_Lane has joined #openstack-infra | 22:19 | |
| clarkb | https://bugs.launchpad.net/openstack-ci/+bug/1203198 | 22:19 |
| jog0 | mordred: I sent it to your private email | 22:20 |
| *** linuxmohan has quit IRC | 22:20 | |
| mordred | jog0: gotit | 22:20 |
| *** datsun180b has quit IRC | 22:21 | |
| clarkb | I have approved my workaround. I will run puppet on the mirror hosts once that is in and happy then trigger the mirror jobs again | 22:21 |
| clarkb | by the time that is all done normal puppet should've puppetted the other slaves | 22:21 |
| jog0 | http://www.mail-archive.com/distutils-sig@python.org/msg17454.html | 22:21 |
| jog0 | mordred: thats where it came from | 22:22 |
| mordred | jog0: ok. I just responded | 22:22 |
| openstackgerrit | A change was merged to openstack-infra/config: Install libffi-dev on our slaves. https://review.openstack.org/37990 | 22:22 |
| mordred | jog0: I believe we could VERY easily fix this in pbr | 22:23 |
| *** mrodden has joined #openstack-infra | 22:23 | |
| clarkb | #status notice Projects with a dependency on xattr are failing tests due to unresolved xattr dependencies. Fix should be in shortly | 22:23 |
| openstackstatus | NOTICE: Projects with a dependency on xattr are failing tests due to unresolved xattr dependencies. Fix should be in shortly | 22:23 |
| mordred | jog0: still specifying the entrypoint console_script as we do now - but intercept it and install a boilerplate script like the entrypoints one | 22:23 |
| mordred | jog0: but that does an import and not a pkg_resources lookup | 22:23 |
| jog0 | mordred: I like the words 'Very easily fix' | 22:23 |
| mordred | jog0: I did not want to be that specific on the mailing list | 22:24 |
| *** mrodden1 has joined #openstack-infra | 22:24 | |
| mordred | jog0: http://paste.openstack.org/show/40974/ is what testr does | 22:25 |
| jog0 | cool that should be much faster | 22:25 |
| mordred | jog0: I think we can do that without checking in 1000 boilerplate crappers into our trees | 22:25 |
| mordred | jog0: gimme a sec | 22:25 |
| openstackgerrit | A change was merged to openstack-infra/jenkins-job-builder: Tests for publishers https://review.openstack.org/14893 | 22:26 |
| jog0 | well we only *really* need it or any binary that is short lived and we expect to call more then once a second | 22:26 |
| *** linuxmohan has joined #openstack-infra | 22:26 | |
| jog0 | for any* | 22:26 |
| openstackgerrit | A change was merged to openstack-infra/config: Add python unittests to JJB. https://review.openstack.org/37617 | 22:27 |
| mordred | jog0: sure. but it's a super-simple interface - so if we solve it once, we might as well just solve it across the board | 22:27 |
| openstackgerrit | A change was merged to openstack-infra/jenkins-job-builder: Tests: Use a testscenarios for running tests. https://review.openstack.org/36990 | 22:27 |
| *** mrodden has quit IRC | 22:27 | |
| jog0 | mordred: true, although I wonder if there is any side effect IE why use pkg_resorces for this at all | 22:27 |
| *** ewindisch has joined #openstack-infra | 22:28 | |
| jog0 | maybe it does some neato trick or something | 22:28 |
| mordred | jog0: why us? or why in general? | 22:28 |
| mordred | it does not | 22:28 |
| mordred | http://paste.openstack.org/show/40975/ | 22:28 |
| mordred | there is what you get from entrypoints | 22:28 |
| mordred | the main difference is checking for version | 22:28 |
| jeblair | clarkb: does the mirror script ever exit nonzero? | 22:29 |
| mordred | but, you know what? I don't think we need to check that at runtime. it's silly | 22:29 |
| openstackgerrit | A change was merged to openstack-infra/config: Fix blueprint and bug parser to require a separator https://review.openstack.org/36853 | 22:29 |
| *** prad has quit IRC | 22:29 | |
| jog0 | mordred: and that magic load_entry_point command | 22:30 |
| jog0 | but yes, why in general and why us | 22:31 |
| *** linuxmohan has quit IRC | 22:31 | |
| jog0 | I just don't want to fix one problem and create a worse one | 22:31 |
| mordred | jog0: why in general is so that we have the bulk of our code in nova/* | 22:31 |
| mordred | and not have real logic in a bin/*py script | 22:31 |
| mordred | where it's hard to unittest | 22:31 |
| clarkb | jeblair: no I don't think it does | 22:31 |
| clarkb | jeblair: it will need to track the status of each install and if one fails exit(1) | 22:32 |
| mordred | or, well, really, it's the same answer for both :) | 22:32 |
| jog0 | mordred: I understand why we do it, but why is there taht load_entry_point magic | 22:32 |
| mordred | well, that's just how setuptools and friends do this | 22:32 |
| clarkb | jeblair: maybe. I think we wanted a single failure to not break everything else | 22:32 |
| clarkb | except a single failure did that by getting past the mirror script | 22:32 |
| dstufft | in a console_script? | 22:32 |
| mordred | dstufft: yeah | 22:32 |
| dstufft | Because setuptool supports multi version installs | 22:32 |
| dstufft | so you have multiple versions of things installed to the same python | 22:32 |
| dstufft | and the load_entry_point stuff activates the appropiate versions | 22:33 |
| jog0 | dstufft: so even multiple versions of dependencies and whatnot? | 22:33 |
| dstufft | jog0: sure | 22:33 |
| dstufft | Installing with pip forces multi version stuff to go away because it's a tremendous hack | 22:33 |
| mordred | dstufft: where would it put such a script? | 22:33 |
| *** ^demon has quit IRC | 22:33 | |
| mordred | ah | 22:33 |
| clarkb | puppet is running on the mirror slaves now | 22:33 |
| jog0 | dstufft: ahh so that the magic it adds | 22:33 |
| dstufft | there's actually a discussion on distutils-sig about this right now | 22:33 |
| mordred | ok. so, the fact that we're pip heavy means that the benefit of the complexity is lost on us | 22:33 |
| *** ^demon has joined #openstack-infra | 22:34 | |
| jog0 | wait, so entry points do somethign that pip prevents from happening | 22:34 |
| jog0 | mordred: python fail | 22:34 |
| dstufft | entry points themselves are just progrmatic hooks | 22:34 |
| jog0 | that have support for multiple versions of everything which presumablly is why importing pkg_resources is slow: it walks the full tree | 22:35 |
| dstufft | Oh is this a speed thing | 22:35 |
| jog0 | dstufft: *nod* | 22:35 |
| clarkb | mirror job builds have started | 22:35 |
| dstufft | It's looking in every .egg-info directory for a entry_points.txt | 22:35 |
| clarkb | hopefully this will all be behind us in about 10 minutes | 22:35 |
| jog0 | in devstack: time python -c "import pkg_resouces" | 22:36 |
| dstufft | which it then loads | 22:36 |
| dstufft | loads ~= opens + reads | 22:36 |
| dstufft | it's an ini file I believe | 22:36 |
| dstufft | does openstack support windows? | 22:36 |
| jog0 | dstufft: http://paste.openstack.org/show/40952/ | 22:36 |
| clarkb | dstufft: some pieces of it | 22:36 |
| clarkb | dstufft: a small subset | 22:36 |
| dstufft | If you don't need windows support you can approximate the same thing without a dependency on pkg_resources just by shipping a python script using the distutils scripts element | 22:37 |
| *** danger_fo is now known as danger_fo_away | 22:37 | |
| dstufft | the problem is windows | 22:37 |
| dstufft | becasue you need an exe for it to work on windows | 22:37 |
| jog0 | dstufft: can I quote you on that "the problem is windows" | 22:37 |
| *** Ryan_Lane is now known as gerrit-wm | 22:37 | |
| mordred | we just shipped scripts until this last cycle | 22:38 |
| jeblair | pretty sure rootwrap isn't expected to work on windows | 22:38 |
| jog0 | jeblair: good point | 22:38 |
| mordred | which I believe will be the solution to our probleem | 22:38 |
| mordred | m | 22:38 |
| dstufft | If it doesn't need windows support you can just ship a regular python file without an extension and get an appromxiation of what console_scripts gives you without the machinary | 22:38 |
| jog0 | \o/ | 22:39 |
| mordred | right. I'm on it. I have a solutoin in mind. I am hacking on it right now | 22:39 |
| jog0 | faster | 22:39 |
| dstufft | if it does need windows you can still do that, you just need to ship an exe like setuptools has | 22:39 |
| dstufft | https://bitbucket.org/pypa/setuptools/src/9dc434ac0308749d564d721a19ee412c2e79754f/launcher.c?at=default | 22:40 |
| dstufft | is the exe setuptools ships | 22:40 |
| dstufft | well the code for it | 22:40 |
| *** ewindisch has quit IRC | 22:40 | |
| dstufft | basically it just loads a .py file named the same as the exe | 22:40 |
| mordred | I'm going to intercept the console_scripts on linux and ship a simple script | 22:40 |
| clarkb | well darn https://jenkins.openstack.org/job/periodic-mirror-python27/141/console | 22:40 |
| mordred | and let it passthrough on windows | 22:40 |
| *** gerrit-wm has quit IRC | 22:40 | |
| mordred | so that setuptools will do the right thing there | 22:40 |
| *** UtahDave has quit IRC | 22:40 | |
| mordred | and the day we care about speed on windows, we can revisit | 22:40 |
| clarkb | I am not super interested in making cffi compile on ubuntu precise. should we just pin xattr? | 22:41 |
| mordred | clarkb: yes | 22:41 |
| dstufft | clarkb: strange, cffi works fine for me on precise | 22:41 |
| dstufft | though I haven't tested xattr | 22:41 |
| *** Ryan_Lane has joined #openstack-infra | 22:42 | |
| clarkb | dstufft: is there another missing lib? | 22:42 |
| dstufft | mordred: fwiw distutils-sig will have a better solution to this sooner or later | 22:42 |
| * clarkb does a little more investigating | 22:42 | |
| dstufft | clarkb: Not sure offhand, however! | 22:42 |
| dstufft | paging Alex_Gaynor :D | 22:42 |
| dstufft | pretty sure he did the cffi port of xattr | 22:43 |
| Alex_Gaynor | dstufft: what'd I do? | 22:43 |
| mordred | dstufft: great! I'll have one today :) | 22:43 |
| dstufft | Alex_Gaynor: cffi + xattr borken for openstack | 22:43 |
| Alex_Gaynor | Can you better specific borken? | 22:43 |
| dstufft | https://jenkins.openstack.org/job/periodic-mirror-python27/141/console | 22:43 |
| Alex_Gaynor | I did port xattr to cffi | 22:43 |
| clarkb | Alex_Gaynor: first it was missing libffi-dev | 22:43 |
| mordred | Alex_Gaynor: | 22:43 |
| Alex_Gaynor | and then bob did a release | 22:43 |
| mordred | Alex_Gaynor: https://jenkins.openstack.org/job/periodic-mirror-python27/141/console | 22:43 |
| clarkb | Alex_Gaynor: but now with libffi-dev installed we get ^ | 22:43 |
| jog0 | mordred: nice solution | 22:44 |
| * clarkb enters a holding pattern in hopes Alex_Gaynor will fix all the things | 22:44 | |
| *** SergeyLukjanov has quit IRC | 22:44 | |
| Alex_Gaynor | hmm, looking | 22:44 |
| clarkb | it does the same on centos6 https://jenkins.openstack.org/job/periodic-mirror-python26/139/console | 22:45 |
| Alex_Gaynor | looks like XATTR_MAXNAMELEN, XATTR_NOFOLLOW, XATTR_NOSECURITY are only defined on some systems | 22:46 |
| clarkb | don't tell me you only tested on OS X :{ | 22:46 |
| clarkb | er :P | 22:46 |
| dstufft | There are other OS's besides OSX? :V | 22:47 |
| mordred | OSX is an OS? | 22:47 |
| clarkb | fwiw we only test on centos and precise and the poor BSDs and OS X get broken sometimes | 22:47 |
| *** cppcabrera_afk is now known as cppcabrera | 22:48 | |
| jeblair | clarkb: openstack runs on bsd and osx? | 22:48 |
| mordred | jeblair: no | 22:48 |
| clarkb | jeblair: vishy does all of his testing without virtualenvs on OS X | 22:48 |
| clarkb | and people use *BSD for stuff apparently | 22:48 |
| clarkb | so ya kinda | 22:48 |
| jeblair | clarkb: yes, but is there an OSX openstack cloud? public or private? | 22:48 |
| dstufft | FreeBSD supremacy | 22:48 |
| clarkb | jeblair: not that I am aware of | 22:48 |
| clarkb | Alex_Gaynor: so, I am guessing that won't be a quick fix? I suppose I should pin xattr for now? | 22:49 |
| mordred | jeblair: no. because not even apple uses it as a server os | 22:49 |
| jeblair | clarkb: we test all the platforms we support. | 22:49 |
| *** pabelanger_ has quit IRC | 22:49 | |
| clarkb | jeblair: good point | 22:49 |
| dstufft | mordred: you mean a bunch of mac minis shoved in a rack doesn't count as servers? | 22:49 |
| Alex_Gaynor | clarkb: Define quick? I think it'll be fixed in <1.5 hours :} | 22:49 |
| * mordred belly laughs | 22:49 | |
| *** gyee has joined #openstack-infra | 22:49 | |
| *** pabelanger has joined #openstack-infra | 22:49 | |
| jeblair | clarkb: +1pin | 22:50 |
| dstufft | I'd be happy if Ic ould legally run OSX on the cloud | 22:50 |
| clarkb | ok whipping that up now. Going to simply delete the package from the mirror first | 22:50 |
| *** cp16net is now known as cp16net|away | 22:50 | |
| mordred | dstufft: honestly - not trolling but a) why b) yeah, I spoke to the apple folks about that - their buildfarm guys agree with you :) | 22:51 |
| jeblair | clarkb, mordred: i'm trying to figure out exactly how wrong/right the mirror script was... | 22:51 |
| jeblair | clarkb, mordred: it's really nuanced... | 22:51 |
| vishy | clarkb: why ya gotta tell everyone about that | 22:51 |
| vishy | :) | 22:51 |
| dstufft | mordred: the clue is in the second point | 22:51 |
| clarkb | vishy: because its crazy :) | 22:51 |
| mordred | dstufft: yeah. I used to have osx builders in the drizzle build farm. of course, we had cloud builers for every other os | 22:51 |
| mordred | dstufft: but for osx, we had a mac pro tower in my living room | 22:52 |
| jog0 | btw there were apple people at the last summit planning on using OpenStack internally on Oracle RHEL | 22:52 |
| dstufft | mordred: my dream is to someday have an OSS buildfarm for PyPI, push up a sdist 2.0, plop out wheels for all the various archs | 22:52 |
| dstufft | mordred: this pretty much requires the ability to use a VM to run OSX though :[ | 22:52 |
| clarkb | mordred: jeblair: so actually Alex_Gaynor thinks <1.5 hours will be fixed I think I won't pin it for now and just rely on the fact that the mirrors don't update very often | 22:52 |
| jeblair | clarkb, mordred: on the one hand, 'external package changes and everything blows up and we pin the reqs repo' is something that we currently (say that) we want. | 22:52 |
| *** SergeyLukjanov has joined #openstack-infra | 22:52 | |
| mordred | dstufft: well, I might know some people who know a lot about really scalable build farms who take submissions from untrusted sources | 22:52 |
| clarkb | mordred: jeblair since I removed the package from the mirror by hand we should be good until Alex_Gaynor has pushed a fix | 22:52 |
| jeblair | clarkb, mordred: on the other hand, if this had been a reqs repo change, it may have "passed tests" and been permitted | 22:53 |
| *** vipul-away is now known as vipul | 22:53 | |
| mordred | jeblair: I think that in this case though the question is logically, how did xattr actually wund up in th eupload | 22:53 |
| mordred | aside from whether we want or don't want the break | 22:53 |
| mordred | as I understand the current mirror logic, it makes no sense to me | 22:53 |
| jeblair | clarkb, mordred: and as one of you pointed out earlier; if bar is permanently broken on pypi, we don't want ignore updates to foo. | 22:54 |
| mordred | jeblair: ++ | 22:54 |
| jeblair | mordred, clarkb: i think before we fix it, we need to start documenting exactly what we want it to do in all cases, maybe on this wiki page: https://wiki.openstack.org/wiki/Requirements | 22:54 |
| mordred | jeblair: +100 | 22:55 |
| jeblair | mordred, clarkb: because i'm almost certain we'll end up with wack-a-mole otherwise | 22:55 |
| Alex_Gaynor | clarkb: ok push is committed, I just realized I don't have pypi push rights, so I'm asking the maintainer | 22:55 |
| clarkb | jeblair: ++ | 22:56 |
| clarkb | Alex_Gaynor: ok, no rush now as I have a temprorary fix | 22:56 |
| dstufft | mordred: this is down the road stuff so it's not needed yet | 22:57 |
| mordred | dstufft: totes. just saying ... I might be interested in hacking on that with you | 22:57 |
| Alex_Gaynor | clarkb: release done! | 22:58 |
| dstufft | mordred: cool sounds good | 22:58 |
| *** nati_ueno has quit IRC | 22:59 | |
| openstackgerrit | Dirk Mueller proposed a change to openstack/requirements: Raise eventlet to 0.13.0 https://review.openstack.org/37996 | 22:59 |
| openstackgerrit | A change was merged to openstack-infra/config: Add tripleo-incubator to openstack. https://review.openstack.org/35360 | 23:00 |
| jeblair | ah, i'm having trouble getting ssl certs because the foundation executive director forwarded hostmaster@openstack.org to his account | 23:01 |
| clarkb | Alex_Gaynor: rerunning mirror jobs now | 23:03 |
| Alex_Gaynor | clarkb: cool | 23:03 |
| clarkb | jeblair: :/ | 23:03 |
| *** mrodden1 has quit IRC | 23:05 | |
| *** mrodden has joined #openstack-infra | 23:05 | |
| *** locke1051 has quit IRC | 23:05 | |
| lifeless | Alex_Gaynor: oh hi :) | 23:05 |
| *** locke105 has joined #openstack-infra | 23:05 | |
| lifeless | Alex_Gaynor: I've got a request happening to get the infra team access to our glance in the new US-East region; but jeblair etc can't use that in prod until all the infra clouds support it - so this is me pinging you about the same for Rackspace :> | 23:06 |
| lifeless | mordred: jeblair: I just had an evil idea; what if there was a takeovernode backend for glance | 23:08 |
| openstackgerrit | A change was merged to openstack-infra/config: Ensure latest setuptools is present. https://review.openstack.org/36708 | 23:08 |
| clarkb | Alex_Gaynor: running builds look much better | 23:10 |
| Alex_Gaynor | clarkb: <phew> | 23:10 |
| openstackgerrit | A change was merged to openstack-infra/config: Fix jeepyb class issue on git.o.o https://review.openstack.org/37954 | 23:11 |
| openstackgerrit | Clark Boylan proposed a change to openstack-infra/config: Tune ElasticSearch settings. https://review.openstack.org/37598 | 23:12 |
| lifeless | mordred: jeblair: specifically, a infra-owned glance instances running in a region; upload an image to it, and it spawns a vm, does takeovernode for you, snapshots, done. | 23:14 |
| clarkb | xattr 0.7.1 is in the mirror now. I think we should be fine | 23:15 |
| jeblair | lifeless: that sounds wonderfully rube-goldberg, but seems like it would work. | 23:16 |
| mordred | :) | 23:16 |
| mordred | I can't imagine ANYTHING that would go wrong ;) | 23:17 |
| lifeless | mordred: I can :> | 23:17 |
| dirk | clarkb: does xattr 0.7.1 fix http://logs.openstack.org/91/37991/2/check/gate-ceilometer-python26/3487/console.html ? | 23:17 |
| *** arborism has quit IRC | 23:18 | |
| *** openstackgerrit has quit IRC | 23:18 | |
| clarkb | dirk: it should | 23:18 |
| jeblair | certs acquired. | 23:19 |
| *** openstackgerrit has joined #openstack-infra | 23:19 | |
| clarkb | dirk: well that particular error was something else. there were two related failures but the other failure should also be correct now | 23:19 |
| zaro | clarkb: looks like this merged... https://review.openstack.org/#/c/37529/ | 23:19 |
| openstackgerrit | A change was merged to openstack-infra/config: Optimize ElasticSearch indices later. https://review.openstack.org/37397 | 23:19 |
| clarkb | dirk: tl;dr try it again and let us know if you see any more breakage | 23:19 |
| zaro | clarkb: now what do i need to do to tag? | 23:19 |
| openstackgerrit | A change was merged to openstack-infra/config: Tune ElasticSearch settings. https://review.openstack.org/37598 | 23:20 |
| clarkb | zaro: we will need to populate that group then you should be able to push tags | 23:21 |
| clarkb | zaro: I am doing about 4 different things at the moment but can add that to the list | 23:21 |
| zaro | clarkb: cool. thanks. | 23:21 |
| *** arborism has joined #openstack-infra | 23:22 | |
| clarkb | mordred: fungi: I think the jeepy install on review.o.o must've hit that setuptools thing | 23:23 |
| clarkb | mordred: fungi: can one of you guys confirm? and perhaps fix it? fixing it will make it so that gerrit groups are created as being self owned | 23:24 |
| *** nati_ueno has joined #openstack-infra | 23:25 | |
| jeblair | (i wish i could do file insertions in hieradata yaml; the ssl cert stuff is unweildy) | 23:26 |
| jeblair | (maybe i will write a python script to edit the yaml) | 23:26 |
| clarkb | jeblair: hiera does support a bunch of backends. there may be a clean way to do that | 23:26 |
| *** SergeyLukjanov has quit IRC | 23:28 | |
| jeblair | clarkb: there seems to be a 3rd party mysql one at least. | 23:29 |
| jeblair | i think i'll write a quick script for now though. | 23:29 |
| clarkb | zaro: can you tell me what your gerrit account id is for the account you are using? | 23:30 |
| clarkb | zaro: its under the settings page | 23:30 |
| clarkb | zaro: and if you like I can walk you through the process of tagging a thing on Monday | 23:30 |
| clarkb | zaro: its a bit painful to correct if smoething goes wrong | 23:30 |
| *** michchap has joined #openstack-infra | 23:31 | |
| fungi | osx/*bsd support... not for server projects, but it's great that people can potentially run the clients there | 23:37 |
| fungi | clarkb: review.o.o was where we hit it (and fixed it) originally. i'm wondering how setuptools/distribute could have downgraded since then | 23:39 |
| zaro | clarkb: account id is 6987 | 23:39 |
| zaro | clarkb: i will wait for your guidance. thanks. | 23:39 |
| jeblair | fungi: maybe we should start testing the clients on bsd | 23:40 |
| fungi | jeblair: i take it jonathan's e-mail isn't over quota and bouncing like todd's then, at least (lloyd said he got an over-quota bounce for hello@o.o which forwarded to todd@o.o) | 23:40 |
| jeblair | fungi: i unforwarded the account and asked the ca to resent the dlv emails | 23:41 |
| jeblair | fungi: and sent an email to jonathan letting him know | 23:41 |
| fungi | oh, even better ;) | 23:41 |
| clarkb | fungi: oh, maybe it was fixed and jeeepyb just wasn't updated? | 23:41 |
| clarkb | fungi: could be a different problem but my jeepyb fix to group creation doesn't seem to be on review.o.o | 23:41 |
| *** michchap has quit IRC | 23:42 | |
| * fungi looks | 23:42 | |
| clarkb | zaro: gearman-plugin-core is ready to go. We can push a tag on monday | 23:43 |
| openstackgerrit | Monty Taylor proposed a change to openstack-dev/pbr: Replace entry_points console_scripts https://review.openstack.org/38000 | 23:43 |
| *** gyee has quit IRC | 23:43 | |
| mordred | jog0, dstufft: ^^ | 23:43 |
| mordred | I need to add unittests | 23:43 |
| clarkb | jeblair: ninja email admin | 23:43 |
| mordred | but I wanted to show you the patch first | 23:43 |
| fungi | clarkb: ahh, looks like we have distribute installed from pip (and at the latest version) but setuptools only installed from ubuntu's precise deb | 23:44 |
| clarkb | #status notice Project affected by the xattr cffi dependency issues should be able to run tests and have them pass. xattr has been fixed and the new version is on our mirror. | 23:47 |
| openstackstatus | NOTICE: Project affected by the xattr cffi dependency issues should be able to run tests and have them pass. xattr has been fixed and the new version is on our mirror. | 23:47 |
| *** ^demon has quit IRC | 23:47 | |
| openstackgerrit | Jeremy Stanley proposed a change to openstack-infra/config: Pull toci into openstack-infra https://review.openstack.org/37432 | 23:47 |
| *** pcrews has quit IRC | 23:48 | |
| fungi | rebased due to merge conflict ^ | 23:48 |
| openstackgerrit | Monty Taylor proposed a change to openstack-dev/pbr: Replace entry_points console_scripts https://review.openstack.org/38000 | 23:49 |
| mordred | lifeless: ^^ also, if you have snarky or evil commends on the above patch, I'd appreciate them | 23:49 |
| *** dkehn is now known as dkehn_away | 23:49 | |
| lifeless | mordred: while I have your attention | 23:52 |
| mordred | lifeless: uhoh | 23:52 |
| lifeless | 04:26 < ttx> mordred: i don't know any of those guys. lifeless probably knows them though | 23:52 |
| lifeless | 04:27 -!- vogxn [~Adium@49.205.181.238] has quit [Quit: Leaving.] | 23:52 |
| lifeless | 04:28 < mordred> lifeless: ^^ if you could poke someone, a new release of that lib with that patch landed would | 23:52 |
| lifeless | be helpful | 23:52 |
| mordred | lifeless: yes please | 23:52 |
| lifeless | mordred: *which project* | 23:52 |
| mordred | lifeless: https://launchpad.net/django-openid-auth | 23:52 |
| lifeless | thank you | 23:52 |
| lifeless | I shall see | 23:52 |
| mordred | lifeless: if you know them - but I assume you know everyone :) | 23:53 |
| lifeless | mordred: review on the console scripts thing | 23:53 |
| mordred | lifeless: thank you | 23:54 |
| mordred | lifeless: good question. lemme check | 23:54 |
| lifeless | mordred: so, you need a release done ? | 23:54 |
| lifeless | [of d-o-a ? | 23:54 |
| mordred | lifeless: yes. it should get a usable thing for develop - that magic should be done by mucking with the lib install .pth stuff | 23:55 |
| mordred | lifeless: yes. the current one is broken | 23:55 |
| *** nati_ueno has quit IRC | 23:55 | |
| mordred | lifeless: https://code.launchpad.net/~mordred/django-openid-auth/zip-safe/+merge/175875 | 23:56 |
| mordred | lifeless: setuptools can't autosense the use of data files in the tree because the calls to use them are inside of django, not in the current source tree | 23:56 |
| mordred | so being explicit is needed if you want to be able to install_requires on that package | 23:56 |
| clarkb | I have reenabled puppet in all of the places that I had turned it off (mirror slaves, logstash.o.o, and the elasticsearch nodes) | 23:58 |
| *** retr0h_ is now known as retr0h | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!