| tristanC | clarkb: changing to git:// ref shouldn't be an issue, thanks for the suggestion | 00:10 |
|---|---|---|
| tristanC | clarkb: corvus: the main openshift specific bits is the role binding... i couldn't figure auth and acl with k8s and i only have access to openshift cluster | 00:12 |
| clarkb | tristanC: the image stuff is also very openshift specific there | 00:13 |
| clarkb | ImageStream and all that | 00:13 |
| tristanC | clarkb: oh, but that's the base job | 00:13 |
| clarkb | tristanC: ya as i explained to corvus I was mostly just surprised at how not k8s that ends up being | 00:14 |
| tristanC | corvus: i didn't realized the spec would need to be implemented on k8s first, i thought we could have both drivers | 00:14 |
| tristanC | clarkb: that's quite a challenging task to implement a native container workflow base job that will works for everyone | 00:15 |
| tristanC | there are lots of implementation details regarding build and deployment that is bound to be project specific | 00:17 |
| *** dtruong2 has joined #zuul | 00:25 | |
| *** dtruong2 has quit IRC | 00:36 | |
| tristanC | corvus: the current spec isn't very technical... are we expecting nodepool to have a k8s cluster role? | 00:39 |
| *** swest has quit IRC | 01:15 | |
| *** dtruong_ has quit IRC | 01:20 | |
| *** swest has joined #zuul | 01:30 | |
| *** openstack has joined #zuul | 02:51 | |
| *** ChanServ sets mode: +o openstack | 02:51 | |
| openstackgerrit | Merged openstack-infra/zuul master: Emit a log line containing version info around startup https://review.openstack.org/583639 | 03:23 |
| openstackgerrit | Ian Wienand proposed openstack-infra/zuul master: doc: Move zuul variable references to a section https://review.openstack.org/584779 | 03:27 |
| openstackgerrit | Ian Wienand proposed openstack-infra/zuul master: Add variables to project https://review.openstack.org/584230 | 03:27 |
| openstackgerrit | Ian Wienand proposed openstack-infra/zuul master: Use definition list for job status https://review.openstack.org/584780 | 03:27 |
| *** dtruong2 has joined #zuul | 03:34 | |
| *** dtruong2 has quit IRC | 03:38 | |
| *** dtruong2 has joined #zuul | 03:38 | |
| openstackgerrit | Ian Wienand proposed openstack-infra/zuul master: Rename didAnyJobFail() to hasAnyJobFailed() https://review.openstack.org/585134 | 03:39 |
| *** dtruong2 has quit IRC | 03:40 | |
| *** dtruong_ has joined #zuul | 04:06 | |
| *** nchakrab has joined #zuul | 04:45 | |
| *** spsurya_ has quit IRC | 04:48 | |
| *** spsurya_ has joined #zuul | 04:49 | |
| *** dtruong2 has joined #zuul | 04:57 | |
| *** spsurya_ is now known as spsurya | 05:07 | |
| *** dtruong2 has quit IRC | 05:21 | |
| *** dtruong2 has joined #zuul | 05:24 | |
| *** dtruong2 has quit IRC | 05:28 | |
| *** dtruong2 has joined #zuul | 05:30 | |
| *** dtruong2 has quit IRC | 05:30 | |
| *** dtruong2 has joined #zuul | 05:31 | |
| *** dtruong2 has quit IRC | 05:32 | |
| *** dtruong2 has joined #zuul | 05:33 | |
| *** dtruong2 has quit IRC | 05:33 | |
| *** swest has quit IRC | 05:45 | |
| *** swest has joined #zuul | 05:46 | |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Tolerate missing project https://review.openstack.org/579872 | 06:04 |
| openstackgerrit | Andreas Jaeger proposed openstack-infra/zuul-jobs master: Restore zuul_swift_upload.py script https://review.openstack.org/585814 | 06:07 |
| tobiash | corvus: responded on 582794 | 06:13 |
| openstackgerrit | Andreas Jaeger proposed openstack-infra/zuul-jobs master: Restore zuul_swift_upload.py script https://review.openstack.org/585814 | 06:20 |
| openstackgerrit | Merged openstack-infra/zuul master: doc: Move zuul variable references to a section https://review.openstack.org/584779 | 06:51 |
| tobiash | corvus: shall I +w 580967 and 582897 or do you want another vote from someone else? | 07:08 |
| openstackgerrit | Fabien Boucher proposed openstack-infra/zuul master: Remove unecessary shebang and exec bit https://review.openstack.org/583563 | 07:12 |
| Diabelko | clarkb: it's only about the defaults it seems - I've changed them and it works fine now; I want to prepare some dummy test scenario to see if with roles A, B, C (same as we had yesterday), the role B with no default values, will take default value from C, as it is the last one d | 07:24 |
| Diabelko | as it is the last one defined in the playbook * | 07:25 |
| *** pcaruana has joined #zuul | 07:25 | |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Restructure return values section in docs https://review.openstack.org/585981 | 07:32 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Restructure return values section in docs https://review.openstack.org/585981 | 08:03 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Support job pause https://review.openstack.org/585389 | 08:03 |
| *** jiapei has joined #zuul | 08:13 | |
| openstackgerrit | Merged openstack-infra/zuul master: Remove unecessary shebang and exec bit https://review.openstack.org/583563 | 08:29 |
| *** nchakrab has quit IRC | 08:35 | |
| *** sshnaidm|afk is now known as sshnaidm | 08:51 | |
| *** nchakrab has joined #zuul | 08:54 | |
| *** sambetts|afk has quit IRC | 08:57 | |
| *** sambetts_ has joined #zuul | 08:59 | |
| *** dtruong_ has quit IRC | 09:02 | |
| *** nchakrab has quit IRC | 09:55 | |
| *** dtruong_ has joined #zuul | 10:10 | |
| *** sshnaidm is now known as sshnaidm|afk | 10:13 | |
| *** nchakrab has joined #zuul | 11:06 | |
| *** Manjesh has joined #zuul | 11:18 | |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Support complex data structures as secrets https://review.openstack.org/585892 | 11:24 |
| mordred | corvus, pabelanger: reviewed https://review.openstack.org/#/c/546474 and left a question | 11:30 |
| *** panda|rover|off is now known as panda|rover | 11:37 | |
| *** Manjesh has quit IRC | 11:49 | |
| mordred | tobiash: I left a +2 and a comment on https://review.openstack.org/#/c/585981 | 12:08 |
| tobiash | mordred: I tried first with the sub section but to my eyes the section headers were too big | 12:09 |
| tobiash | I can change back to subsection if folks prefer that | 12:09 |
| mordred | tobiash: no - I think the bold was fine - mostly just wanted to bring it up in case | 12:12 |
| *** nchakrab has quit IRC | 12:12 | |
| *** nchakrab has joined #zuul | 12:13 | |
| *** rlandy has joined #zuul | 12:28 | |
| *** nchakrab has quit IRC | 12:30 | |
| *** jiapei has quit IRC | 12:33 | |
| *** openstackgerrit has quit IRC | 12:36 | |
| *** sshnaidm|afk is now known as sshnaidm | 12:48 | |
| *** maeca has joined #zuul | 12:56 | |
| *** samccann has joined #zuul | 13:23 | |
| *** nchakrab has joined #zuul | 13:28 | |
| tobiash | mordred: you're the openstack sdk expert | 13:28 |
| tobiash | mordred: do you know if it's possible to stream data during download when downloading objects? | 13:29 |
| tobiash | https://docs.openstack.org/openstacksdk/latest/user/proxies/object_store.html#openstack.object_store.v1._proxy.Proxy.download_object | 13:29 |
| tobiash | that seems to return the byte array directly | 13:29 |
| mordred | tobiash: you can using the get_object method that is directly on the connection object | 13:32 |
| mordred | tobiash: by passing either a filename (it'll stream to a file) or a file-like object to the outfile parameter | 13:33 |
| mordred | tobiash: I need to fix the docs to point people to the right places, but for now I recommend looking at https://docs.openstack.org/openstacksdk/latest/user/connection.html before looking at the individual proxy classes | 13:35 |
| tobiash | mordred: thanks, good to know | 13:35 |
| mordred | tobiash: if the file-like-object interface is unpleasing, we could probably add a wrapper method, like stream_object - that would create an in-memory file-like-object to pass to get_object and return an iterator wrapping it | 13:58 |
| tobiash | mordred: I'm trying to put this into a bytesio and forward this to the resonse in a wsgi app | 13:59 |
| tobiash | after reading the docs bytesio is probably not the thing I'm looking for | 13:59 |
| mordred | tobiash: for uploading large objects to swift that need to be broken in to segments, we made this: http://git.openstack.org/cgit/openstack/openstacksdk/tree/openstack/cloud/_utils.py#n665 | 14:02 |
| mordred | tobiash: which isn't exactly what you need for download- but makes me think that making a similar one for the streaming-download case is maybe not a terrible idea | 14:03 |
| tobiash | mordred: I wonder if that can be accomplished with some combination of streams in the io package of python | 14:04 |
| mordred | tobiash: probably - although now I'm looking at requests and it has a simpler interface that allows returning an iterator | 14:05 |
| mordred | lemme take a quick stab at a patch to sdk to expose something more directly | 14:05 |
| tobiash | mordred: http://paste.openstack.org/show/726695/ | 14:06 |
| tobiash | that is not quite what I expect from BytesIO :/ | 14:06 |
| *** _ari_ has quit IRC | 14:07 | |
| *** myoung has quit IRC | 14:07 | |
| *** pabelanger has quit IRC | 14:07 | |
| *** weshay_PTO has quit IRC | 14:07 | |
| *** pabelanger has joined #zuul | 14:07 | |
| *** _ari_ has joined #zuul | 14:07 | |
| tobiash | hrm, I think I need a pair of streams | 14:08 |
| pabelanger | mordred: looking, 546474 | 14:08 |
| *** weshay has joined #zuul | 14:08 | |
| *** myoung has joined #zuul | 14:12 | |
| tobiash | mordred: hrm, it looks like get_object pushes all data synchronously into the file object so that doesn't help me with streaming | 14:17 |
| pabelanger | mordred: replied, we are good on https://review.openstack.org/546474/ | 14:26 |
| tobiash | mordred: it might be better in my use case to directly use requests for the download | 14:28 |
| mordred | tobiash: how bout this: remote: https://review.openstack.org/586250 Add support for streaming object responses | 14:29 |
| mordred | tobiash: but also - if you need to use the request object directly - that should also be fine - conn.object_store.get() should return you a response | 14:29 |
| tobiash | ah cool | 14:30 |
| mordred | pabelanger: ++ | 14:31 |
| tobiash | mordred: that looks like exactly the thing | 14:33 |
| tobiash | I need | 14:33 |
| mordred | tobiash: conn.object_store.get or the patch? | 14:35 |
| tobiash | both ;) | 14:35 |
| tobiash | mordred: your patch returns a generator function right? is it possible to return (response_headers, <iterator>)? | 14:37 |
| tobiash | mordred: if I understood that right, you could move that for loop with the yield into its own function (_iterate) and return '(response_headers, iterate(response))' ? | 14:39 |
| mordred | tobiash: I thought about that - but the 'safe' way to deal with iter_content is to put the get call in a with: block so that the response is sure to get closed properly (otherwise you can apparently leak memory) | 14:40 |
| mordred | tobiash: and I don't think we can then return because it would make response go out of scope | 14:41 |
| mordred | BUT - it's entirely posible I'm wrong about that - I'm still only on the first coffee | 14:41 |
| tobiash | ah ok | 14:41 |
| mordred | tobiash: there is also a "get_object_metadata" method which returns what the headers would be | 14:41 |
| tobiash | too bad that this sacrifices the headers then | 14:41 |
| tobiash | mordred: yes, but that would be another rtt | 14:42 |
| mordred | yah | 14:42 |
| corvus | tobiash: i don't think we need more votes on 967 and 897; i approved. | 14:42 |
| mordred | I also considered yielding the headers as the first step | 14:42 |
| mordred | tobiash: do you think that would be weird to consume? | 14:42 |
| tobiash | yes | 14:42 |
| mordred | yeah. me too | 14:42 |
| mordred | corvus: if you get a sec, could you look at https://review.openstack.org/586250 and also the scrollback between tobiash and I and tell me if there is a better way you can think of? | 14:43 |
| mordred | tobiash: oh - also thought about making a raw parameter that would just return the response object | 14:43 |
| mordred | corvus: (and more tl;dr - what are scoping rules for with context_manager_thing() as foo: return foo ... does the foo become invalid becaue the return exits the with context? | 14:44 |
| mordred | especially if you're using the with to make sure foo closes properly (this is where C++ and std::smart_ptr is actually what I want) | 14:45 |
| tobiash | mordred: I think returning the response object would fit my use case best | 14:46 |
| tobiash | mordred: in that case I can setup the swgi response using the headers and after that directly yielding the content | 14:47 |
| tobiash | mordred: what I also saw is that the response already decodes the compressed file (which should be done by the browser in my case) | 14:49 |
| corvus | pabelanger: did you change anything between PS4 and PS5 of 546474? | 14:49 |
| *** nchakrab has quit IRC | 14:49 | |
| tobiash | mordred: what I'd like to do there is to forward the compressed stream with the correct headers so the browser decodes instead of my middleware | 14:50 |
| tobiash | I'll think later about that, have to run now | 14:51 |
| corvus | tobiash, mordred: weighed in on the "bold" question in 585981 :) | 14:53 |
| tobiash | Ok, will change that | 14:54 |
| pabelanger | remote: https://review.openstack.org/546474 Add support for Ansible extra-vars flag | 14:56 |
| pabelanger | corvus: mordred: had to rebase due to merge issue, but also added reno about new feature^ | 14:57 |
| pabelanger | corvus: sorry, I didn't see your +2 before pushing reno change | 14:57 |
| Shrews | grrr.... anyone remember how to stop unauthed irc users from pvt msg'ing??? | 14:58 |
| corvus | You can set /mode <YourNick> +R to avoid getting PMs from unregistered users. | 14:58 |
| Shrews | corvus: thx. spammers are now going pvt | 14:59 |
| Shrews | got 3 in the last 2 min | 14:59 |
| pabelanger | Yah, terrible this time around the spam | 15:00 |
| mordred | tobiash: remote: https://review.openstack.org/586262 Add method for returning a raw response for an object | 15:01 |
| corvus | mordred: 250 and 262 look reasonable to me | 15:02 |
| mordred | woot | 15:02 |
| * mordred needs to add tests and renos for them - but figured making sure they don't look insane first is a good idea | 15:03 | |
| corvus | tobiash: out of curiosity, why are you writing a wsgi to proxy this instead of just using mod_proxy or similar? | 15:03 |
| tobiash | corvus: I don't know how to make mod proxy work because of keystone auth | 15:04 |
| corvus | tobiash: your swift is private? | 15:05 |
| tobiash | I didn't find a way to make basic auth working in our swift | 15:05 |
| tobiash | Yes, that's private | 15:05 |
| tobiash | And a working solution would be wsgi that does auth against swift | 15:06 |
| mordred | kmalloc: ^^ when someone gets extra time, someone should write an apache module that could auth against keystone for doing things like authenticated mod_proxy in front of a swift container | 15:06 |
| corvus | tobiash: why does the swift container need to be private? do you want one container per tenant and to restrict access to that tenant, or...? | 15:07 |
| *** Guest19794 is now known as kklimonda | 15:13 | |
| *** kklimonda has quit IRC | 15:14 | |
| *** kklimonda has joined #zuul | 15:14 | |
| corvus | 00:39 < tristanC> corvus: the current spec isn't very technical... are we expecting nodepool to have a k8s cluster role? | 15:21 |
| corvus | tristanC: i don't understand that question. what do you mean "k8s cluster role" ? | 15:21 |
| Diabelko | clarkb: https://github.com/Diabelko/ansible-defaultvars-test | 15:29 |
| kmalloc | mordred: we have one, we support external auth, which can use mod_auth, mod_auth_ldap, etc | 15:30 |
| kmalloc | As long as remote_user is populated, and the auth request indicates external (I'll have to find docs) it works. | 15:31 |
| kmalloc | Oh | 15:32 |
| kmalloc | Sorry different. | 15:32 |
| kmalloc | Hmm. | 15:32 |
| mordred | kmalloc: yah. non-urgent- just wanted to put it across your brain-radar | 15:33 |
| *** pcaruana has quit IRC | 15:38 | |
| corvus | the nodepool docs don't appear to be up to date | 15:49 |
| corvus | the log is here: http://logs.openstack.org/ef/ef498011898893e739015d1581e955c7199e03d8/release/publish-zuul-docs/ba4c9a1/job-output.txt.gz | 15:49 |
| corvus | it seems to have put the docs in a 3.2.0 directory? | 15:50 |
| corvus | that's not how it's supposed to work | 15:50 |
| clarkb | corvus: I think the post job for the commit publishes to the latest/master location then the tags publish to the tag specific location | 15:52 |
| corvus | hrm, that may be teh first time we've run publish-zuul-docs on release... | 15:52 |
| pabelanger | corvus: mordred: tobiash: clarkb: anything else to review on 584656 (rsync bug fix)? 4 +2 currently | 15:52 |
| corvus | pabelanger: nope +3 | 15:53 |
| tobiash | corvus: our swift is backend infrastructure where normal users don't have access. Further we need to restrict access to tenants because we also work with external partners who should only see projects they need. The approach is to do authorization on apache revrrse proxies on a per path level like we do for zuul. | 15:54 |
| corvus | clarkb: hrm. unfortunately, that means our release notes are out of date until another patch lands | 15:54 |
| pabelanger | corvus: Yay | 15:55 |
| kmalloc | mordred: I did something like that in Lua/haproxy, I could revisit it (easier than Apache module) to start. | 15:55 |
| corvus | tobiash: got it. so atm, you need to write your own proxy because the thing mordred described doesn't exist. | 15:56 |
| corvus | or maybe it does :) ^ | 15:56 |
| tobiash | Maybe it does but I didn't find it in hours of googling | 15:57 |
| clarkb | corvus: we could have the tag job publish to the latest/master/$correctbranch location as well | 15:58 |
| clarkb | I just don't know that it does that today | 15:58 |
| *** myoung is now known as myoung|lunch | 16:00 | |
| *** sshnaidm is now known as sshnaidm|afk | 16:00 | |
| mordred | tobiash: updated https://review.openstack.org/586262 and https://review.openstack.org/586262 | 16:00 |
| * mordred afks for lunch | 16:00 | |
| tobiash | mordred: these links are identical ;) | 16:00 |
| clarkb | corvus: tobiash that doens't sound too disimilar from the rax CDN case | 16:02 |
| clarkb | where "normal users" cannot hit the objects unless going through the CDN | 16:02 |
| corvus | clarkb: except the cdn is public | 16:02 |
| clarkb | right but its a different url for read access than the url used for writes | 16:03 |
| corvus | clarkb: sure, but i'm failing to draw any conclusions from that... | 16:04 |
| clarkb | just that it is probably a reasonable use case to support in the role if possible. Rax isn't a complete outlier we can ignore for simplicity | 16:04 |
| corvus | clarkb: oh, i don't think we need to (or can do) anything in the role. there's no relationship between the url that tobiash wants to report and the url that the role stashed tho logs. so he'll just set the log_url himself after it runs. | 16:06 |
| corvus | clarkb: the rax case is, potentially, different, if rax can tell us the read location. | 16:06 |
| corvus | clarkb: put another way, in the rax case, rax is the oracle which knows the read location. in tobiash's case, tobiash is the oracle. | 16:06 |
| clarkb | gotcha | 16:07 |
| tobiash | corvus: yes, I guess that's fine for my use case. I just need the possibility to keep the container private | 16:07 |
| corvus | tobiash: one last idea -- i don't suppose you can use swift's container acls for your access control, and then simply put a dumb proxy (with no authentication) in front of it for the network access? | 16:08 |
| tobiash | corvus: no, the auth systems of swift and our ci are different | 16:10 |
| corvus | k | 16:11 |
| corvus | tobiash: is haproxy an option for you? would kmalloc's work be helpful? | 16:11 |
| tobiash | So even if acls would work (they don't atm) the users couldn't login | 16:11 |
| tobiash | We're using saml for auth and are tied to apache | 16:12 |
| kmalloc | Right. Hmm, let me check something. | 16:12 |
| tobiash | So haproxy would be between apache and swift just like the wsgi middleware | 16:12 |
| kmalloc | I could prob spin up a mod_lua script quickly. | 16:13 |
| kmalloc | tobiash: yeah haproxy can do thst. | 16:13 |
| kmalloc | I rewrote keystone middleware into haproxy with Lua as a poc ~2yrs ago | 16:14 |
| kmalloc | I could re-write it for modern keystone pretty quickly, and flesh it out. | 16:14 |
| logan- | https://github.com/morganfainberg/HAProxyKeystoneMiddlware/blob/master/ha_proxy_ksa.lua | 16:14 |
| kmalloc | A full Apache module would take me longer (but doable) | 16:14 |
| kmalloc | Yep. That won't really work anymore, but I could totally rework/finish it and even add testing. | 16:15 |
| kmalloc | If it helps folks. Let.me read scroll back to catch up. | 16:15 |
| *** openstackgerrit has joined #zuul | 16:19 | |
| openstackgerrit | Merged openstack-infra/zuul master: Fix delegate_to for ansible synchronize https://review.openstack.org/584656 | 16:19 |
| tobiash | kmalloc: thanks, I don't want to generate effort on your side and already have a poc wsgi proxy (without streaming yet) that will do for the start | 16:22 |
| tobiash | But I think that lua approach is the better one in the long run | 16:23 |
| kmalloc | Sure, but if it helps, I can totally spend some cycles bringing that back. :) | 16:23 |
| kmalloc | It was a fun day project to prove it worked (in lieu of ksm in a cloud) | 16:23 |
| *** samccann has quit IRC | 16:34 | |
| *** chandankumar is now known as chkumar|trekk | 16:47 | |
| *** myoung|lunch is now known as myoung | 17:35 | |
| *** nchakrab has joined #zuul | 17:44 | |
| *** maeca has quit IRC | 17:53 | |
| *** nchakrab has quit IRC | 17:59 | |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Restructure return values section in docs https://review.openstack.org/585981 | 18:18 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Support job pause https://review.openstack.org/585389 | 18:18 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Support job pause https://review.openstack.org/585389 | 18:19 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Restructure return values section in docs https://review.openstack.org/585981 | 18:20 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Support job pause https://review.openstack.org/585389 | 18:21 |
| tobiash | corvus, mordred: do you have problems with a merger? I got a merge failure on https://review.openstack.org/#/c/585389/ and I'm pretty sure that there was no merge conflict (gerrit didin't complain when pressing the rebase button) | 18:22 |
| clarkb | tobiash: I don't think that is a known issue but I will look at logs | 18:24 |
| tobiash | clarkb: happened also to 585892 this morning (check successful, gate directly after that merge failed) | 18:28 |
| clarkb | tobiash: http://paste.openstack.org/show/726704/ seems to be a valid merge failure, it doesn't like that file modes are changing | 18:31 |
| tobiash | clarkb: thanks for checking, interesting that gerrit doesn't care about this | 18:32 |
| tobiash | I actually thought that gerrit would be pickier than plain git | 18:33 |
| clarkb | I wonder if this is a jgit bug | 18:33 |
| clarkb | where it doesn't handle the mode change properly on a rebase and C git gets confused? | 18:33 |
| tobiash | maybe, or it just handles that correctly and git could be improved | 18:34 |
| tobiash | clarkb, corvus: what do you think about propagating the merge failure details to the user? | 18:50 |
| tobiash | I also often get questions from the users about the reason of merge failures... | 18:51 |
| clarkb | tobiash: I want to say there is an open story for doing that, I would be in favor | 18:51 |
| clarkb | tobiash: in this case I think the stderr output has the useful information | 18:52 |
| openstackgerrit | Merged openstack-infra/zuul master: Support complex data structures as secrets https://review.openstack.org/585892 | 18:57 |
| tobiash | Maybe next week I have time to implement that | 18:57 |
| openstackgerrit | Merged openstack-infra/nodepool master: Use pipelines for stats keys https://review.openstack.org/585577 | 18:57 |
| *** ianychoi_ has joined #zuul | 19:15 | |
| *** pcaruana has joined #zuul | 19:17 | |
| *** maeca1 has joined #zuul | 19:17 | |
| *** ianychoi has quit IRC | 19:18 | |
| *** elyezer has quit IRC | 19:20 | |
| *** elyezer has joined #zuul | 19:20 | |
| *** samccann has joined #zuul | 19:44 | |
| *** pcaruana has quit IRC | 19:45 | |
| *** maeca1 has quit IRC | 20:00 | |
| openstackgerrit | Merged openstack-infra/zuul master: Restructure return values section in docs https://review.openstack.org/585981 | 20:01 |
| *** hashar has joined #zuul | 20:02 | |
| corvus | tobiash: 585389 looks really good; i left one relatively minor comment | 20:09 |
| corvus | pabelanger, ianw: i just realized that we may want to coordinate your changes. pabelanger is adding extra-vars in https://review.openstack.org/546474 and ianw is adding project vars in https://review.openstack.org/584230 . logically, we should also add project extra-vars. | 20:13 |
| *** elyezer has quit IRC | 20:14 | |
| corvus | pabelanger, ianw: are one of you interesting in rebasing your change on the other and adding that, or do one of you want to volunteer to add that in a followup after both changes merge? | 20:14 |
| *** elyezer has joined #zuul | 20:14 | |
| pabelanger | corvus: what ever is easier for ianw, but agree with coordinating | 20:15 |
| pabelanger | about to head out for some food, I'll check back this evening once ianw is online | 20:16 |
| corvus | kk | 20:17 |
| mordred | pabelanger, ianw: also - I think there's gonna be a merge conflict between the extra-vars patch and the tobiash pause patch | 20:19 |
| mordred | (I can't remember what the code was I was looking at that touched the same thing - but it seemed like it was gonna conflict- possibly in the docs) | 20:19 |
| corvus | hrm, gerrit doesn't think so right now, but it's been wrong before :) | 20:20 |
| mordred | corvus: also I've been wrong before | 20:21 |
| tobiash | corvus: thanks for review, will fix tomorrow | 20:21 |
| corvus | mordred: i'm sure i don't remember | 20:21 |
| corvus | pabelanger, ianw: i've left +2s on both, so i'd suggest a followup patch that depends-on both. maybe flip a coin for who writes it. :) | 20:28 |
| corvus | pabelanger: re 584978 why is destdir confusing? | 20:31 |
| clarkb | corvus: its being used as the source dir in most of the tests | 20:31 |
| clarkb | I found that confusing (it works out because the same filesystem is used on the local and remote nodes so you can copy in betwee nthem all you want but if they were different hosts it would start to break down quickly) | 20:32 |
| corvus | clarkb: gotcha. i read "test_workspace_dir" as having something to do with the test infrastructure (we have tests that do that). or maybe the job workspace. maybe we could do "scratch_dir" or something? | 20:34 |
| clarkb | corvus: that wfm. | 20:34 |
| *** maeca has joined #zuul | 20:50 | |
| *** samccann has quit IRC | 20:59 | |
| *** hashar has quit IRC | 21:23 | |
| openstackgerrit | Merged openstack-infra/zuul master: Add support for Ansible extra-vars flag https://review.openstack.org/546474 | 21:34 |
| openstackgerrit | Merged openstack-infra/zuul master: Rename didAnyJobFail() to hasAnyJobFailed() https://review.openstack.org/585134 | 21:34 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: Add variables to project https://review.openstack.org/584230 | 21:47 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: Use definition list for job status https://review.openstack.org/584780 | 21:47 |
| mordred | ianw, corvus: ^^ zuul reported merge conflict - I rebased - no conflicts (shrug) and reapplied the +A | 21:48 |
| ianw | ok :) | 21:49 |
| mordred | corvus: https://review.openstack.org/#/c/582794/ has 2x+2 - any reason we should hold off pulling the trigger? | 21:56 |
| clarkb | mordred: the only thing I can think of is maybe it would cause memory problems again? and getting a tested release out without memory issues would be nice | 21:58 |
| mordred | clarkb: ++ | 21:58 |
| mordred | good call | 21:58 |
| mordred | also ... | 21:58 |
| mordred | corvus: clarkb and I have both +2'd https://review.openstack.org/#/c/95035/ (yes, that number is correct) - but since it's so old - do you want to look and/or pull the trigger? | 21:59 |
| mordred | SpamapS: ^^ that's a patch originally from you from 2014 - it seems to have been hacked on by tim chavez, pabelanger and mhu | 22:00 |
| * mordred now afks | 22:00 | |
| rcarrillocruz | tim! | 22:16 |
| rcarrillocruz | good times :-) | 22:16 |
| corvus | yeah, i'll give that a once (or 22nd?) over... | 22:24 |
| corvus | done | 22:27 |
| openstackgerrit | Merged openstack-infra/zuul master: Docs: add a cross-ref in secrets https://review.openstack.org/583301 | 22:32 |
| pabelanger | woot | 22:33 |
| *** pabelanger has quit IRC | 22:46 | |
| *** pabelanger has joined #zuul | 22:46 | |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul master: Report to gerrit over HTTP https://review.openstack.org/577027 | 22:51 |
| corvus | that's ready for actual review if it passes tests | 22:51 |
| corvus | ianw: did you see my note earlier about extra-vars and project vars? | 22:53 |
| clarkb | corvus: merge failed | 22:54 |
| clarkb | I wonder if there isn't something weird with the merges afterall | 22:54 |
| corvus | clarkb: well, it's an old change, it wouldn't surprise me | 22:54 |
| openstackgerrit | Merged openstack-infra/zuul master: Add variables to project https://review.openstack.org/584230 | 22:54 |
| clarkb | ah ok | 22:54 |
| corvus | hrm. cherry-picks okay. i'll check the logs. | 22:55 |
| clarkb | corvus: tobiash's merge failre was unhappy about a file mode change | 22:56 |
| corvus | ERROR: permissions conflict: 100755->100644,100755 in tests/base.py | 22:57 |
| clarkb | http://paste.openstack.org/show/726704/ same error as before but a different file | 22:58 |
| corvus | https://review.openstack.org/583563 | 22:58 |
| corvus | every outstanding change created before that change was merged is going to conflict | 22:58 |
| clarkb | ah | 22:59 |
| clarkb | if they touch those files (which they should if they are making changes that need test updates) | 22:59 |
| corvus | yeah | 22:59 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul master: Report to gerrit over HTTP https://review.openstack.org/577027 | 23:00 |
| corvus | mordred, tobiash: ^ we figured out why we're getting lots of apparently false merge conflicts | 23:00 |
| clarkb | and tobish reported that clicking the rebase button in gerrit wasn't sufficient | 23:01 |
| corvus | fascinating | 23:01 |
| clarkb | it did rebase fine according to gerrit but then zuul remained unhappy with it | 23:01 |
| openstackgerrit | Merged openstack-infra/zuul master: Use definition list for job status https://review.openstack.org/584780 | 23:01 |
| *** maeca has quit IRC | 23:12 | |
| *** tranzemc_ is now known as tranzemc | 23:23 | |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul master: Report to gerrit over HTTP https://review.openstack.org/577027 | 23:43 |
| clarkb | next release of pyyaml is going to make safe_load the default load() and danger_load() will exist for when you really want to load code out of yaml | 23:47 |
| clarkb | I don't think this affects our use of pyyaml | 23:47 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!