| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Extend github testing using app auth https://review.openstack.org/563297 | 00:15 |
|---|---|---|
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Test parent-change-enqueued with github https://review.openstack.org/563242 | 00:15 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Optionally record created github clients in tests https://review.openstack.org/563402 | 00:15 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Limit search scope of getChangesDependingOn to tenant https://review.openstack.org/563196 | 00:15 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Remove getPullBySha from FakeGithubConnection https://review.openstack.org/563416 | 00:15 |
| *** eumel8 has joined #zuul | 00:31 | |
| *** eumel8 has quit IRC | 01:16 | |
| *** D3VIATION has joined #zuul | 01:25 | |
| *** swest has quit IRC | 01:56 | |
| *** swest has joined #zuul | 02:12 | |
| *** rlandy|rover|bbl is now known as rlandy|rover | 02:26 | |
| *** D3VIATION has quit IRC | 02:27 | |
| *** rlandy|rover has quit IRC | 03:27 | |
| *** eumel8 has joined #zuul | 03:35 | |
| *** pcaruana has joined #zuul | 04:24 | |
| *** pcaruana has quit IRC | 04:29 | |
| *** openstackgerrit has quit IRC | 04:34 | |
| *** D3VIATION has joined #zuul | 05:27 | |
| *** D3VIATION has quit IRC | 05:31 | |
| *** dkranz has quit IRC | 06:06 | |
| *** quiquell|off is now known as quiquell | 06:15 | |
| *** jamielennox has joined #zuul | 06:31 | |
| *** eumel8 has quit IRC | 06:41 | |
| *** pcaruana has joined #zuul | 06:54 | |
| *** dkranz has joined #zuul | 06:58 | |
| *** ssbarnea_ has joined #zuul | 07:28 | |
| *** jpena|off is now known as jpena | 07:37 | |
| *** sshnaidm|off is now known as sshnaidm | 07:59 | |
| *** gtema has joined #zuul | 08:01 | |
| *** toabctl has quit IRC | 09:08 | |
| *** johanssone has quit IRC | 09:41 | |
| *** johanssone has joined #zuul | 09:44 | |
| *** ssbarnea_ has quit IRC | 10:42 | |
| *** jpena is now known as jpena|lunch | 10:58 | |
| *** jpena|lunch is now known as jpena|off | 11:10 | |
| *** eumel8 has joined #zuul | 11:52 | |
| *** jpena|off is now known as jpena | 11:58 | |
| *** quiquell is now known as quiquell|lunch | 12:12 | |
| *** openstackgerrit has joined #zuul | 12:24 | |
| openstackgerrit | Artem Goncharov proposed openstack-infra/zuul master: github approval type is approved https://review.openstack.org/570378 | 12:24 |
| *** rlandy has joined #zuul | 12:28 | |
| *** rlandy is now known as rlandy|rover | 12:30 | |
| *** eumel8 has quit IRC | 12:32 | |
| *** quiquell|lunch is now known as quiquell | 12:41 | |
| *** eumel8 has joined #zuul | 12:48 | |
| openstackgerrit | Artem Goncharov proposed openstack-infra/zuul master: correct github review approval type is `approved` https://review.openstack.org/570378 | 12:53 |
| *** myoung|off is now known as myoung | 13:03 | |
| *** D3VIATION has joined #zuul | 13:09 | |
| odyssey4me | hi folks - is there a way for nodepool to execute a user_data file when building a node? | 14:55 |
| odyssey4me | I'm thinking of a situation where a cloud-image is used, not a diskimage | 14:55 |
| clarkb | odyssey4me: you would need to use a cloud init image or other image with a tool like cloud init | 14:56 |
| clarkb | then set the metadata on the image? | 14:56 |
| clarkb | We may have only done that for ssh keys now that I think about it | 14:56 |
| odyssey4me | yeah, not a whole lot of options in https://docs.openstack.org/infra/nodepool/configuration.html#cloud-images | 14:57 |
| *** acozine1 has joined #zuul | 14:59 | |
| gtema | odyssey4me: if you want I can share a link to 'elements' I was using to build image with cloud_init - there are some tricks, which costed me some time | 14:59 |
| odyssey4me | gtema in this case I'm using a cloud-image, not diskimage-builder... so thanks, but building the image myself with cloud-init doesn't help this use case at all :) | 15:01 |
| gtema | ok | 15:02 |
| clarkb | the issue is nodepool doesnt currebtly have a mechanism to set the user data attribute on nodes | 15:02 |
| clarkb | probabpy straightforward to add as long as shade does yhe base 64 encoding properly but will require cbanges to nodepool I think | 15:02 |
| odyssey4me | clarkb ok, thanks for the confirmation - that might just be something we work on | 15:03 |
| tobiash | adam_g: I think it could be a valid use case for requiring comments with certain text in it. So I think we should at least not remove that feature. We could extend it now or at a point in time when someone needs it. | 15:04 |
| *** gtema has quit IRC | 15:14 | |
| *** gtema has joined #zuul | 15:15 | |
| *** pcaruana has quit IRC | 15:19 | |
| *** electrofelix has quit IRC | 15:32 | |
| *** GonZo2000 has joined #zuul | 15:36 | |
| *** eumel8 has quit IRC | 15:38 | |
| *** quiquell is now known as quiquell|off | 15:42 | |
| *** wei_liu_ has joined #zuul | 15:44 | |
| *** wei_liu_ has quit IRC | 15:50 | |
| *** wei_liu_ has joined #zuul | 15:51 | |
| *** eumel8 has joined #zuul | 15:54 | |
| *** wei_liu_ has quit IRC | 15:54 | |
| *** sshnaidm has quit IRC | 15:56 | |
| *** wei_liu_ has joined #zuul | 15:58 | |
| *** Jacktt has joined #zuul | 16:01 | |
| *** wei_liu_ has quit IRC | 16:01 | |
| *** Jacktt has quit IRC | 16:06 | |
| *** dmellado has quit IRC | 16:09 | |
| *** gouthamr has quit IRC | 16:09 | |
| pabelanger | odyssey4me: clarkb: we did have the concept of a ready-script before in nodepool but was deprecated during zuulv3 development, that might have been a way in the past | 16:14 |
| odyssey4me | pabelanger yeah, seems to me that using the user_data for that sort of thing would be more cloudy ;) | 16:15 |
| odyssey4me | that way it's not nodepool that executes anything except a node build | 16:15 |
| mordred | odyssey4me: shade should totally support that - and if it doesn't it's definitely a bug | 16:16 |
| mordred | odyssey4me: although fwiw personally I stay as far away from cloud-init user_data scripts as humanly possible | 16:16 |
| odyssey4me | we had a another little bit of fun - if you have time for a story, I'll tell it :) | 16:16 |
| mordred | (that's saying yes, we should totally support it if you want to use it, just as long as nothing grows an actual _dependency_ on using it) | 16:17 |
| odyssey4me | mordred oh agreed, but it's handy to get a package installed for jenkins to be able to use the node - this is part of the jenkins <-> nodepool use-case which we're using | 16:17 |
| mordred | ++ | 16:17 |
| mordred | totally | 16:17 |
| odyssey4me | so, I'm not sure how universal this is, but rax pubclous has a seperate quota for onmetal vs normal instances | 16:17 |
| odyssey4me | nodepool, however, only cares about the instance quota | 16:18 |
| odyssey4me | we had one provider with two pools - one instances, and one onmetal | 16:18 |
| odyssey4me | node requests for the onmetal nodes were being blocked because provisioning the node would exceed quota | 16:19 |
| odyssey4me | except that the onmetal quota would not be exceeded :p | 16:19 |
| odyssey4me | oh, and btw - this was a fun one to work out: https://github.com/openstack-infra/nodepool/blob/db87b73a508a91043d80109980b29a72cefd3c72/nodepool/driver/openstack/handler.py#L258-L259 | 16:20 |
| odyssey4me | if it's not non-negative, then return false | 16:21 |
| corvus | odyssey4me: do you know how we can differentiate the two quotas? | 16:22 |
| *** pcaruana has joined #zuul | 16:23 | |
| *** pcaruana has quit IRC | 16:32 | |
| odyssey4me | corvus well, I can't seem to get a good answer on what the api is - so I was thinking about an alternative route | 16:38 |
| odyssey4me | how about having an option per pool or perhaps per provider (not sure which makes more sense) to disable the quota check and blindly accept max-servers only? | 16:39 |
| corvus | SpamapS, mordred, tobiash, tristanC, jhesketh, clarkb: now that we've got big picture stuff out of the way, are there more details about containers we should discuss today? | 16:39 |
| SpamapS | I'd definitely be interested in talking about container based deployment tooling. | 16:40 |
| tobiash | corvus: my flight was rescheduled to today 12:30, happy to further participate if it doesn't conflict | 16:40 |
| corvus | tobiash: some openstack-infra folks are chatting with kata folks over lunch today, so i think we won't have any container discussions during then :) | 16:42 |
| tobiash | yay | 16:42 |
| corvus | SpamapS: cool -- mordred has some thoughts/wip on that too | 16:42 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Remove getPullBySha from FakeGithubConnection https://review.openstack.org/563416 | 16:43 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Status branch protection checking for github https://review.openstack.org/535680 | 16:43 |
| mordred | SpamapS: I'd love to chat with you | 16:43 |
| SpamapS | cool, I'm hiding in hotel at the moment, and heading to the airport after lunch | 16:43 |
| SpamapS | But I'll be at the conf in about 45. | 16:44 |
| corvus | odyssey4me: i think the quotas are per provider, so it may be a bit of work to extend that to essentially make a synthetic per-pool quota. but that sounds like a good solution to explore. | 16:44 |
| odyssey4me | corvus ok cool, happy that it sounds at least reasonable | 16:45 |
| corvus | SpamapS, mordred: i'm on the second floor at the back by the windows, watching the sea planes, of course. maybe we can meet up around here? | 16:46 |
| odyssey4me | I did go down a thought process of considering whether we should implement something in shade which understands quota types, etc... but I think that's a bit silly for what is most likely some sort of quota hack in the rax api. I think a quota bypass is simpler and more universal. | 16:46 |
| SpamapS | corvus: ACK, I'll make my way to that area. | 16:47 |
| *** sshnaidm has joined #zuul | 16:49 | |
| *** sshnaidm is now known as sshnaidm|off | 16:50 | |
| odyssey4me | corvus we had some fun trying to understand how request routing is done and what the rules are - I've got to spend more time following the actual code path because we hit a weird situation where a secondary pool across two providers was not getting routing requests distributed across both - only one pool (the first one in the config) was getting all the requests | 16:50 |
| odyssey4me | that's one which will be harder to replicate to confirm, but we'll try :) | 16:51 |
| odyssey4me | how keen would we be to have some docs on how nodepool actually does things internally (process flows, that sort of thing)? given I'll be mapping this stuff out, I figure it may be handy to push up to the docs | 16:52 |
| *** gouthamr has joined #zuul | 17:04 | |
| corvus | odyssey4me: i think that would be swell :) | 17:04 |
| corvus | odyssey4me: though i don't think a pool can span providers? i think a pool is only a "part" of one provider | 17:05 |
| odyssey4me | corvus oh yea, it didn't span... it was just a secondary pool in each provider which provided the same node names | 17:06 |
| odyssey4me | so think provider1 -> pool2 -> ubuntu-xenial, provider2 -> pool2 -> ubuntu-xenial | 17:06 |
| corvus | odyssey4me: ah thatk makes sense | 17:06 |
| odyssey4me | provider1 -> pool2 got all the requests, while provider2 -> pool2 was bored | 17:06 |
| odyssey4me | anyway, lemme try to replicate the issue and work my way through how those requests actually get routed - yesterday was a long day and that situation may have just been something silly I did | 17:08 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Don't process lock files as refs in tests https://review.openstack.org/570417 | 17:15 |
| *** jpena is now known as jpena|off | 17:28 | |
| Shrews | odyssey4me: requests aren't really "routed". rather, there is 1 thread per pool and each thread continuously goes through the request queue trying to grab any unhandled requests | 17:30 |
| Shrews | those threads can pause if they are at quota, but otherwise, first thread to grab the request wins | 17:31 |
| *** eumel8 has quit IRC | 17:36 | |
| *** D3VIATION has quit IRC | 17:37 | |
| *** gtema has quit IRC | 18:00 | |
| *** eumel8 has joined #zuul | 18:00 | |
| *** wei_liu has joined #zuul | 18:10 | |
| *** wei_liu has quit IRC | 18:13 | |
| *** eumel8 has quit IRC | 18:21 | |
| *** eumel8 has joined #zuul | 18:34 | |
| openstackgerrit | David Shrewsbury proposed openstack-infra/nodepool master: Add multilabel support to ZooKeeper Node model https://review.openstack.org/570227 | 18:39 |
| SpamapS | mordred: oops, missed you and corvus ;) | 18:42 |
| corvus | SpamapS: i'm still in the back for now, don't know where mordred is | 18:44 |
| *** eumel8 has quit IRC | 18:44 | |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Status branch protection checking for github https://review.openstack.org/535680 | 18:52 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Remove getRepoPermission fron FakeGithubConnection https://review.openstack.org/570428 | 18:52 |
| tobiash | adam_g: this adds the groundwork for counting the permission requests ^ | 18:57 |
| tobiash | adam_g: I think counting should be easy to be added in FakeGithubSession.get(...) | 18:58 |
| *** ChanServ has quit IRC | 19:07 | |
| *** ChanServ has joined #zuul | 19:12 | |
| *** barjavel.freenode.net sets mode: +o ChanServ | 19:12 | |
| *** ChanServ has quit IRC | 19:21 | |
| *** ChanServ has joined #zuul | 19:30 | |
| *** barjavel.freenode.net sets mode: +o ChanServ | 19:30 | |
| *** sc68cal is now known as Guest57137 | 19:33 | |
| *** jlvacation is now known as Guest65555 | 19:33 | |
| *** fungi is now known as Guest77695 | 19:34 | |
| *** sdake is now known as Guest14735 | 19:34 | |
| *** dmsimard is now known as Guest94448 | 19:34 | |
| *** trishnag is now known as Guest60589 | 19:34 | |
| *** kklimonda is now known as Guest19794 | 19:34 | |
| *** kmalloc is now known as Guest70437 | 19:34 | |
| *** tristanC is now known as Guest15248 | 19:34 | |
| *** Guest77695 is now known as fungi | 19:35 | |
| openstackgerrit | David Shrewsbury proposed openstack-infra/nodepool master: Add config read and node deletion notifications https://review.openstack.org/565882 | 19:37 |
| *** rlandy|rover is now known as rlandy | 19:42 | |
| *** harlowja has joined #zuul | 19:42 | |
| *** rlandy is now known as rlandy|rover | 19:43 | |
| *** Guest94448 has quit IRC | 19:51 | |
| *** dmsimard has joined #zuul | 19:52 | |
| *** eumel8 has joined #zuul | 19:52 | |
| *** eumel8 has quit IRC | 20:09 | |
| *** eumel8 has joined #zuul | 20:45 | |
| *** acozine1 has quit IRC | 20:52 | |
| *** pwhalen_ has joined #zuul | 20:54 | |
| *** pwhalen has quit IRC | 20:57 | |
| *** pwhalen_ is now known as pwhalen | 20:57 | |
| *** pwhalen has quit IRC | 20:58 | |
| *** pwhalen has joined #zuul | 20:58 | |
| *** Guest70437 is now known as kmalloc | 21:02 | |
| *** kmalloc has quit IRC | 21:02 | |
| *** kmalloc has joined #zuul | 21:02 | |
| *** kmalloc has quit IRC | 21:02 | |
| *** kmalloc has joined #zuul | 21:02 | |
| *** kmalloc is now known as morgan | 21:03 | |
| *** morgan is now known as kmalloc | 21:03 | |
| *** eumel8 has quit IRC | 21:04 | |
| *** gothicmindfood has quit IRC | 21:11 | |
| *** gothicmindfood has joined #zuul | 21:11 | |
| *** gothicmindfood has quit IRC | 21:14 | |
| *** gothicmindfood has joined #zuul | 21:15 | |
| *** D3VIATION has joined #zuul | 21:29 | |
| *** Guest65555 is now known as jlvillal | 21:37 | |
| *** D3VIATION1 has joined #zuul | 22:23 | |
| *** D3VIATION has quit IRC | 22:24 | |
| *** D3VIATION1 has quit IRC | 22:38 | |
| *** D3VIATION has joined #zuul | 22:54 | |
| *** D3VIATION has quit IRC | 23:03 | |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Status branch protection checking for github https://review.openstack.org/535680 | 23:20 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Remove getRepoPermission fron FakeGithubConnection https://review.openstack.org/570428 | 23:20 |
| *** rlandy|rover is now known as rlandy|rvr|bbl | 23:25 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!