| *** jamesmcarthur has joined #zuul | 00:01 | |
| *** jamesmcarthur has quit IRC | 00:23 | |
| *** jamesmcarthur has joined #zuul | 00:47 | |
| *** jamesmcarthur has quit IRC | 01:00 | |
| *** jamesmcarthur has joined #zuul | 01:02 | |
| *** jamesmcarthur has quit IRC | 01:17 | |
| *** altlogbot_2 has quit IRC | 01:28 | |
| *** altlogbot_0 has joined #zuul | 01:30 | |
| *** jamesmcarthur has joined #zuul | 01:45 | |
| *** jamesmcarthur has quit IRC | 01:51 | |
| *** dmyrhorodskyi has quit IRC | 01:52 | |
| *** jamesmcarthur has joined #zuul | 02:05 | |
| *** rfolco has quit IRC | 02:09 | |
| *** bhavikdbavishi has joined #zuul | 02:25 | |
| *** bhavikdbavishi1 has joined #zuul | 02:28 | |
| *** bhavikdbavishi has quit IRC | 02:30 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 02:30 | |
| *** bhavikdbavishi has quit IRC | 02:59 | |
| *** jamesmcarthur has quit IRC | 03:25 | |
| *** bhavikdbavishi has joined #zuul | 04:10 | |
| *** SotK has quit IRC | 04:14 | |
| *** SotK has joined #zuul | 04:15 | |
| *** AJaeger_ is now known as AJaeger | 05:17 | |
| *** raukadah is now known as chandankumar | 05:51 | |
| *** pcaruana has joined #zuul | 06:12 | |
| *** sanjayu_ has joined #zuul | 06:26 | |
| *** themroc has joined #zuul | 06:57 | |
| *** sshnaidm|ruck is now known as sshnaidm|afk | 07:21 | |
| *** jangutter has joined #zuul | 07:31 | |
| *** gtema_ has joined #zuul | 07:46 | |
| *** hashar has joined #zuul | 07:52 | |
| *** openstackgerrit has joined #zuul | 08:06 | |
| openstackgerrit | Merged zuul/zuul master: Fix multi-tenant caching of extra config files https://review.opendev.org/669008 | 08:06 |
|---|---|---|
| *** brett-soric has joined #zuul | 08:11 | |
| *** tosky has joined #zuul | 08:11 | |
| *** AJaeger has quit IRC | 08:26 | |
| mhu | Hi there, we noticed that nodepool won't allocate a nodeset if all the nodes don't belong to the same pool: is this by design? | 08:33 |
| *** sshnaidm|afk is now known as sshnaidm|ruck | 08:38 | |
| tobiash | Yes, cross pool multi node is not possible and would require quite some thought | 08:47 |
| *** brett-soric has left #zuul | 08:56 | |
| *** AJaeger has joined #zuul | 09:40 | |
| *** gtema has joined #zuul | 09:46 | |
| *** gtema_ has quit IRC | 09:47 | |
| *** panda is now known as panda|bbl | 09:48 | |
| mhu | tobiash, sorry just seeing your answer - so this is deliberate IIUC? | 10:12 |
| *** gtema has quit IRC | 10:13 | |
| *** gtema has joined #zuul | 10:14 | |
| *** electrofelix has joined #zuul | 10:29 | |
| *** panda|bbl has quit IRC | 10:33 | |
| *** hashar has quit IRC | 10:33 | |
| *** panda has joined #zuul | 10:37 | |
| *** bhavikdbavishi has quit IRC | 10:38 | |
| *** hashar has joined #zuul | 10:58 | |
| *** hashar has quit IRC | 11:27 | |
| fungi | mhu: deliberate, correct | 11:31 |
| *** bjackman_ has quit IRC | 11:34 | |
| fungi | at a minimum we need continued support for avoiding cross-provider or inter-region nodesets because some users of nodepool rely on that | 11:35 |
| *** rfolco has joined #zuul | 11:38 | |
| *** bhavikdbavishi has joined #zuul | 11:43 | |
| *** bjackman_ has joined #zuul | 12:01 | |
| *** hashar has joined #zuul | 12:14 | |
| mhu | thanks fungi ! | 12:17 |
| *** rlandy has joined #zuul | 12:29 | |
| * Shrews waves with his one good paw | 12:43 | |
| pabelanger | mhu: there are also some inefficiencies around multi-node quota handling too. We see some cases where 1 of 2 nodes will come online, but 2nd node cannot launch due to quota limit. Then you have an idle locked ready node until a slot frees up. In our cases, that could be 90mins | 12:45 |
| pabelanger | would be nice to handle the nodeset request as single quota check | 12:45 |
| mhu | pabelanger, ah yeah makes sense | 12:45 |
| pabelanger | but yah, we've seen this issue too where a provider is filled up, but 3 others are empty | 12:46 |
| pabelanger | due to multinode jobs | 12:46 |
| *** bhavikdbavishi has quit IRC | 13:08 | |
| *** bhavikdbavishi has joined #zuul | 13:09 | |
| *** yolanda has quit IRC | 13:14 | |
| *** gtema has quit IRC | 13:31 | |
| *** electrofelix has quit IRC | 13:45 | |
| *** jamesmcarthur has joined #zuul | 13:47 | |
| *** jamesmcarthur has quit IRC | 13:52 | |
| *** michael-beaver has joined #zuul | 13:53 | |
| *** sgw has quit IRC | 13:57 | |
| *** dmellado has quit IRC | 14:09 | |
| *** dmellado has joined #zuul | 14:11 | |
| *** yolanda has joined #zuul | 14:15 | |
| openstackgerrit | James E. Blair proposed zuul/zuul-jobs master: Don't use -latest nodesets in integration jobs https://review.opendev.org/669675 | 14:18 |
| openstackgerrit | James E. Blair proposed zuul/zuul-jobs master: Change the repo used in some base-roles tests https://review.opendev.org/669676 | 14:18 |
| pabelanger | looking at statsd, I can see ansible-test-network-integration-ios-python37.SUCCESS.sum (or job.SUCCESS.sum) which I think is the time it look for the job to run. I can then add that into graphite or grafana to see the average run times over a week. However, I am having a little trouble trying to see what the average time for all job runs in a week is (not sure which function i should be using is). | 14:29 |
| pabelanger | I'm trying to calculate compute hours basically | 14:30 |
| *** dmellado has quit IRC | 14:30 | |
| *** dmellado has joined #zuul | 14:32 | |
| pabelanger | sumSeries, looks to give me the data in 1 day itervals | 14:32 |
| pabelanger | but, I need to zoom out to 6 months in grafana | 14:33 |
| *** chandankumar is now known as raukadah | 14:35 | |
| clarkb | pabelanger: https://review.opendev.org/#/c/616306/ the data reported by that change may be more useful for compute hours? | 14:39 |
| pabelanger | clarkb: yah, that is a good start. I was looking to break it down further by job | 14:43 |
| openstackgerrit | James E. Blair proposed zuul/zuul-jobs master: Don't use -latest nodesets in integration jobs https://review.opendev.org/669675 | 14:45 |
| openstackgerrit | James E. Blair proposed zuul/zuul-jobs master: Change the repo used in some base-roles tests https://review.opendev.org/669676 | 14:45 |
| clarkb | pabelanger: the script in the zuul/tools dir called node_usage.py (I think) can do it down to job level | 14:46 |
| clarkb | but requires access to logs | 14:46 |
| pabelanger | ack | 14:47 |
| *** gouthamr has quit IRC | 14:47 | |
| pabelanger | I'm looking at grafyaml dashboards in openstack now, to see if this is already done | 14:47 |
| clarkb | it isn't | 14:49 |
| openstackgerrit | James E. Blair proposed zuul/zuul-jobs master: Add test job for ensure-sphinx https://review.opendev.org/668958 | 14:49 |
| openstackgerrit | James E. Blair proposed zuul/zuul-jobs master: Update the sphinx python package sets https://review.opendev.org/668954 | 14:49 |
| *** dmellado has quit IRC | 14:49 | |
| *** dmellado has joined #zuul | 14:53 | |
| *** jamesmcarthur has joined #zuul | 14:54 | |
| *** gouthamr has joined #zuul | 14:56 | |
| *** jeliu_ has joined #zuul | 14:57 | |
| *** sgw has joined #zuul | 15:15 | |
| *** hashar has quit IRC | 15:30 | |
| *** rlandy has quit IRC | 15:31 | |
| *** rlandy_ has joined #zuul | 15:36 | |
| *** rlandy_ is now known as rlandy | 15:37 | |
| *** rlandy has quit IRC | 15:38 | |
| corvus | mordred: i left a comment on 659180 which i suspect is at the level of minor typographical error, but could be more and is worth double checking. otherwise, i dig it. :) | 15:38 |
| *** rlandy has joined #zuul | 15:44 | |
| *** rlandy is now known as rlandy_ | 15:45 | |
| *** rlandy has joined #zuul | 15:45 | |
| *** jeliu_ has quit IRC | 15:57 | |
| Shrews | corvus: wow, mucho changes with nodepool testing, i see. | 15:58 |
| corvus | Shrews: yeah, we landed the rework where we stop running nodepool as a devstack plugin, and just use devstack to get an openstack. glean, sdk, dib, etc inherit from the new jobs, but they pass in most of their configuration, so that stuff isn't in the nodepool repo anymore. i think there's still a little bit more we could move out of the nodepool repo and into the "consumers" but the heavy lifting is done; | 16:00 |
| corvus | we can tweak that as we need to later | 16:00 |
| corvus | Shrews: so i think zuul and nodepool repos are both ready to move to the zuul tenant; i'm finishing up zuul-jobs now... then i think we might be ready to make the move? i'll have to check. | 16:01 |
| *** gtema has joined #zuul | 16:01 | |
| Shrews | neat | 16:01 |
| corvus | Shrews: last i checked we hadn't merged the admin web api yet, so i haven't pushed on your autohold stack. there have been lots of people taking time off, so the atmosphere has been... "relaxed"? :) | 16:03 |
| Shrews | corvus: ok. i noticed my stack is all in merge conflict anyway. i suppose it can just remain that way until we are ready to consider it | 16:05 |
| corvus | Shrews: ack; i don't imagine the conflicts are major | 16:06 |
| Shrews | fyi, i'm not expecting to be super productive for a couple more weeks. a fairly bulky sling prevents me from using that arm for typing | 16:07 |
| Shrews | but alas, my vacation days are all used up, so i"ll do what i can | 16:07 |
| corvus | Shrews: it's nice to see your voice again :) | 16:08 |
| Shrews | lol | 16:08 |
| *** mattw4 has joined #zuul | 16:09 | |
| corvus | a general question for folks: i'm looking at adding a feature to make files matchers match when a job's configuration is changed (right now, the best we can do is match "^\.zuul.yaml", which, if the repo has a lot of jobs, could mean running a bunch of unecessary changes; the only antidote to that is to put each job in its own file, and that's tedious). so, if i add this feature, should a files matcher | 16:12 |
| corvus | just always automatically match if the job config itself is updated? or should we make it explicit, perhaps with something like matching a fake file which can't exist ("^/ZUUL_JOB$")? | 16:12 |
| corvus | clarkb, AJaeger, tobiash, SpamapS, pabelanger: ^ you may have thoughts, but i'd love to hear from anyone | 16:13 |
| clarkb | I think most users will want the default behavior of testing the job if its config changes. Maybe instead of an explicit matcher we have some way to disable a default of testing the job if its config updates? | 16:14 |
| clarkb | That could be files-matcher-opts dict ? | 16:15 |
| corvus | that could work | 16:15 |
| *** bhavikdbavishi has quit IRC | 16:17 | |
| *** hwangbo has joined #zuul | 16:23 | |
| tobiash | corvus: that sounds like a great idea | 16:25 |
| tobiash | ++ for explicit | 16:25 |
| *** jamesmcarthur has quit IRC | 16:27 | |
| fungi | i do like the idea of always running a job if that job's configuration changes | 16:30 |
| *** jamesmcarthur has joined #zuul | 16:30 | |
| fungi | extending that to also trigger the job if any of its in-repo playbooks or roles are altered would probably also be desirable, though that seems like it would be nontrivial to determine | 16:31 |
| AJaeger | corvus: yeah, I was thinking about that problem when writing https://review.opendev.org/#/c/669148 - how to best self test a change. So, I'm all in for any change to a job triggers it - and agree with clarkb, we need a way to disable that... | 16:34 |
| AJaeger | corvus: and what about if a parent changes that is in the same repo? | 16:35 |
| *** jamesmcarthur has quit IRC | 16:35 | |
| * AJaeger would be fine if .zuul.yaml (or smiliar) is updated, all jobs defined in that file are run | 16:36 | |
| *** igordc has joined #zuul | 16:37 | |
| corvus | AJaeger: you can do that today -- the problem is that if we did that in either the zuul-jobs or the system-config repo, a single change could run enough unecessary jobs to exhaust a significant portion of our quota | 16:38 |
| corvus | basically, those two repos have a lot of "unrelated" jobs -- system-config will eventually have a job for every server we run, and zuul-jobs will eventually have a job for every role it has | 16:39 |
| AJaeger | corvus: so, some more fine grained control would indeed be better... | 16:39 |
| corvus | i'm pretty sure i can get "change C has modifed job J, therefore run job J". i'm not sure if calculating "change C depends on change B which modifies job J, so run job J on change C" is practical... it's possible, but it may involve quite a lot of computation. so i'm going to aim for the simpler case first; we can expand it later. | 16:43 |
| *** themroc has quit IRC | 16:48 | |
| fungi | yeah, i think anything at all there is of some benefit | 16:52 |
| fungi | and the immediate change detection is probably the 90% case anyway | 16:52 |
| openstackgerrit | James E. Blair proposed zuul/zuul master: WIP: Run jobs when their own config changes https://review.opendev.org/669722 | 16:53 |
| openstackgerrit | Merged zuul/zuul-jobs master: Add a script to make platform-specific versions of jobs https://review.opendev.org/668955 | 16:54 |
| *** jamesmcarthur has joined #zuul | 16:56 | |
| AJaeger | corvus: sure, start simple | 16:59 |
| SpamapS | corvus:I've wanted that smart file matcher for a long time. :) | 17:02 |
| openstackgerrit | Andreas Jaeger proposed zuul/zuul-jobs master: Add Gentoo integration tests https://review.opendev.org/669147 | 17:06 |
| *** igordc has quit IRC | 17:10 | |
| openstackgerrit | Merged zuul/zuul-jobs master: Add base role integration jobs https://review.opendev.org/668061 | 17:11 |
| pabelanger | corvus: wfm | 17:15 |
| openstackgerrit | Merged zuul/zuul-jobs master: Add multi-node integration jobs https://review.opendev.org/668767 | 17:15 |
| openstackgerrit | Merged zuul/zuul-jobs master: Don't use -latest nodesets in integration jobs https://review.opendev.org/669675 | 17:15 |
| corvus | SpamapS, pabelanger: do you have an opinion on whether it should be opt-in or opt-out? | 17:15 |
| pabelanger | maybe slight preference to opt-out? | 17:16 |
| pabelanger | I mean, if enabled by default wouldn't be a big issue for us I think | 17:17 |
| AJaeger | corvus: can we now remove testing from openstack-zuul-jobs? | 17:20 |
| corvus | AJaeger: iirc, there's some stuff that's only in ozj which may need to stay, but i think we can remove most of that; i'll work on that. | 17:22 |
| AJaeger | thanks, corvus ! | 17:23 |
| *** gtema has quit IRC | 17:24 | |
| *** jeliu_ has joined #zuul | 17:24 | |
| *** jamesmcarthur has quit IRC | 17:34 | |
| *** jamesmcarthur has joined #zuul | 17:37 | |
| SpamapS | corvus:opt-out would feel more helpful. I could see some people not liking that | 17:41 |
| *** igordc has joined #zuul | 17:41 | |
| openstackgerrit | Merged zuul/zuul-jobs master: Change the repo used in some base-roles tests https://review.opendev.org/669676 | 17:42 |
| openstackgerrit | Merged zuul/zuul-jobs master: Add Gentoo integration tests https://review.opendev.org/669147 | 17:42 |
| *** igordc has quit IRC | 17:43 | |
| *** SpamapS has quit IRC | 17:46 | |
| *** igordc has joined #zuul | 17:46 | |
| *** sanjayu_ has quit IRC | 17:47 | |
| *** jamesmcarthur has quit IRC | 17:52 | |
| *** SpamapS has joined #zuul | 17:55 | |
| *** jeliu_ has quit IRC | 17:57 | |
| *** jeliu_ has joined #zuul | 17:58 | |
| *** jamesmcarthur has joined #zuul | 17:58 | |
| *** jamesmcarthur has quit IRC | 17:59 | |
| *** jamesmcarthur has joined #zuul | 18:14 | |
| *** jamesmcarthur has quit IRC | 18:14 | |
| *** jamesmcarthur has joined #zuul | 18:15 | |
| openstackgerrit | Merged zuul/zuul-jobs master: Add test job for ensure-sphinx https://review.opendev.org/668958 | 18:30 |
| openstackgerrit | Merged zuul/zuul-jobs master: Update the sphinx python package sets https://review.opendev.org/668954 | 18:34 |
| *** sshnaidm|ruck is now known as sshnaidm|afk | 18:46 | |
| *** themroc has joined #zuul | 18:54 | |
| *** themroc has quit IRC | 18:56 | |
| openstackgerrit | Jeff Liu proposed zuul/zuul-operator master: Create tasks that spin up kubernetes and run the operator https://review.opendev.org/668029 | 19:12 |
| fungi | tristanC: i'm curious, what was the impetus for having zuul-operator depend on postgres? is that a more common rdbms choice in the kubernetes ecosystem? | 19:25 |
| corvus | fungi: mordred suggests that we replace that in the current revision of the spec | 19:28 |
| fungi | ahh | 19:29 |
| fungi | seems reasonable | 19:29 |
| fungi | i'll resist the urge to submit a typo fix to the readme for it if that is likely to go away anyway ;) | 19:30 |
| corvus | fungi: yeah, my view of the situation is that we're designing the system in the spec, and once we know where that's heading, we'll update tristanC's work to match; in the mean time, jeliu_ is working on getting jobs working so when we're ready to move forward, we'll have tests ready to go. i'm not too concerned with the actual content of the repo right now (other than it's something that jeliu_ can load | 19:32 |
| corvus | into k8s and execute); i don't think it reflects any final decisions | 19:32 |
| *** jeliu_ has quit IRC | 19:32 | |
| *** jeliu_ has joined #zuul | 19:33 | |
| corvus | though maybe we should merge a readme update that says "this is not ready" :) | 19:33 |
| fungi | thanks for the detail! | 19:33 |
| openstackgerrit | Jeff Liu proposed zuul/zuul-operator master: Create tasks that spin up kubernetes and run the operator https://review.opendev.org/668029 | 19:33 |
| openstackgerrit | Jeff Liu proposed zuul/zuul-operator master: Create tasks that spin up kubernetes and run the operator https://review.opendev.org/668029 | 20:11 |
| mordred | corvus fungi: I agree with all of those words | 20:24 |
| mordred | corvus: additionally, I love the idea of the feature to have file matchers match on job config changes - and I lean towards opt-out | 20:35 |
| openstackgerrit | Jeff Liu proposed zuul/zuul-operator master: Create tasks that spin up kubernetes and run the operator https://review.opendev.org/668029 | 20:35 |
| mordred | corvus: I think it's the behavior people want to have happen most of the time | 20:35 |
| mordred | even if they haven't thought about it super specifically | 20:36 |
| mordred | and _not_ doing it is a great way to foot-gun | 20:36 |
| *** rlandy has quit IRC | 20:39 | |
| *** rlandy_ is now known as rlandy | 20:39 | |
| *** ianychoi has joined #zuul | 20:41 | |
| openstackgerrit | James E. Blair proposed zuul/zuul master: WIP: Run jobs when their own config changes https://review.opendev.org/669752 | 20:46 |
| *** pcaruana has quit IRC | 20:50 | |
| corvus | i'm feeling more love for "out-out" so i'm going with that for now; if folks want to -1 that in review it won't be too hard to change. i will now spend 2 hours attempting to name the option to disable it. | 20:51 |
| fungi | opt-out sgtm too, though it does mean a (probably subtle) change in behavior | 20:54 |
| corvus | yeah. hopefully a welcome one. :) | 20:55 |
| *** mattw4 has quit IRC | 21:02 | |
| *** mattw4 has joined #zuul | 21:03 | |
| mordred | corvus: what if you name the option "untest-my-jobs-kthxbai" | 21:06 |
| fungi | plz-to-be-breaking-my-jobs | 21:08 |
| *** jeliu_ has quit IRC | 21:16 | |
| *** jamesmcarthur has quit IRC | 21:17 | |
| *** mattw4 has quit IRC | 21:26 | |
| *** mattw4 has joined #zuul | 21:27 | |
| openstackgerrit | James E. Blair proposed zuul/zuul master: Run jobs when their own config changes https://review.opendev.org/669752 | 21:39 |
| corvus | those are both way better than what i went with. | 21:39 |
| mordred | corvus: yours is less bad for different reasons | 21:40 |
| *** jeliu_ has joined #zuul | 21:44 | |
| *** tjgresha has joined #zuul | 22:10 | |
| *** jeliu_ has quit IRC | 22:23 | |
| *** jamesmcarthur has joined #zuul | 22:34 | |
| openstackgerrit | James E. Blair proposed zuul/zuul master: Run jobs when their own config changes https://review.opendev.org/669752 | 22:47 |
| *** jamesmcarthur has quit IRC | 22:53 | |
| *** jamesmcarthur has joined #zuul | 23:00 | |
| *** tosky has quit IRC | 23:06 | |
| tristanC | fungi: corvus: it's just that the available postgresql operator seemed easier to use | 23:14 |
| fungi | got it, thanks tristanC! | 23:14 |
| *** andreaf has quit IRC | 23:15 | |
| *** andreaf has joined #zuul | 23:15 | |
| *** rlandy has quit IRC | 23:36 | |
| *** jamesmcarthur has quit IRC | 23:37 | |
| *** hwangbo has quit IRC | 23:42 | |
| *** mattw4 has quit IRC | 23:45 | |
| *** michael-beaver has quit IRC | 23:53 | |
| *** jamesmcarthur has joined #zuul | 23:57 | |
| *** hwangbo has joined #zuul | 23:58 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!