| openstackgerrit | Merged openstack-infra/zuul master: gerrit: cast change and patchset numbers to str https://review.openstack.org/577154 | 00:09 |
|---|---|---|
| tristanC | dmsimard: could you please have a look to https://review.openstack.org/577676, it's about using ara_report_path | 00:16 |
| *** yolanda_ has joined #zuul | 01:05 | |
| openstackgerrit | Merged openstack-infra/zuul-jobs master: ara-report: use a different fact name for ara_report_path https://review.openstack.org/577676 | 01:08 |
| *** yolanda has quit IRC | 01:08 | |
| *** yolanda has joined #zuul | 01:09 | |
| *** yolanda_ has quit IRC | 01:11 | |
| *** yolanda_ has joined #zuul | 01:13 | |
| *** yolanda has quit IRC | 01:16 | |
| *** yolanda__ has joined #zuul | 01:19 | |
| *** yolanda_ has quit IRC | 01:22 | |
| *** rlandy has quit IRC | 01:33 | |
| pabelanger | clarkb: yes, different pools for different max-servers, and to consolidate resources into a single cloud | 01:54 |
| pabelanger | (project) | 01:54 |
| *** Shrews has quit IRC | 02:00 | |
| dmsimard | tristanC: good catch | 02:44 |
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/nodepool master: builder: do not configure provider that doesn't manage images https://review.openstack.org/578642 | 03:35 |
| *** nguyenhai has left #zuul | 04:32 | |
| *** nchakrab has joined #zuul | 05:20 | |
| *** nchakrab has quit IRC | 05:23 | |
| *** openstackgerrit has quit IRC | 05:34 | |
| *** openstackstatus has quit IRC | 05:51 | |
| *** openstackstatus has joined #zuul | 05:54 | |
| *** ChanServ sets mode: +v openstackstatus | 05:54 | |
| *** yolanda__ is now known as yolanda | 05:59 | |
| *** nchakrab has joined #zuul | 06:03 | |
| *** gtema has joined #zuul | 06:11 | |
| *** aspiers[m] has quit IRC | 06:40 | |
| *** aspiers[m] has joined #zuul | 06:41 | |
| *** openstackgerrit has joined #zuul | 07:08 | |
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/zuul master: executor: add executing job INFO log https://review.openstack.org/578704 | 07:08 |
| *** pcaruana has joined #zuul | 07:10 | |
| *** yolanda_ has joined #zuul | 07:35 | |
| *** yolanda is now known as Guest63448 | 07:37 | |
| *** yolanda_ is now known as yolanda | 07:37 | |
| *** Guest63448 has quit IRC | 07:38 | |
| *** jpena|off is now known as jpena | 07:44 | |
| *** hashar has joined #zuul | 07:44 | |
| *** nchakrab_ has joined #zuul | 07:54 | |
| *** nchakrab has quit IRC | 07:56 | |
| *** threestrands has joined #zuul | 08:06 | |
| *** sshnaidm is now known as sshnaidm|afk | 08:09 | |
| *** threestrands has quit IRC | 08:13 | |
| *** electrofelix has joined #zuul | 08:15 | |
| *** nchakrab_ has quit IRC | 08:30 | |
| *** nchakrab has joined #zuul | 09:00 | |
| *** nchakrab has quit IRC | 09:04 | |
| *** nchakrab has joined #zuul | 09:07 | |
| *** nchakrab has quit IRC | 09:30 | |
| *** nchakrab has joined #zuul | 10:39 | |
| *** sshnaidm|afk is now known as sshnaidm | 10:58 | |
| *** jpena is now known as jpena|lunch | 10:59 | |
| *** nchakrab has quit IRC | 11:05 | |
| *** elyezer has quit IRC | 11:06 | |
| *** nchakrab has joined #zuul | 11:12 | |
| *** elyezer has joined #zuul | 11:18 | |
| *** nchakrab has quit IRC | 11:27 | |
| *** nchakrab has joined #zuul | 11:28 | |
| *** nchakrab has quit IRC | 11:36 | |
| openstackgerrit | Fabien Boucher proposed openstack-infra/zuul master: Add tenant yaml validation option to scheduler https://review.openstack.org/574265 | 11:39 |
| *** pcaruana has quit IRC | 11:56 | |
| *** jpena|lunch is now known as jpena | 11:56 | |
| *** rlandy has joined #zuul | 12:13 | |
| *** rlandy is now known as rlandy|afk | 12:40 | |
| *** pcaruana has joined #zuul | 12:44 | |
| goern | hmm.. how do I include the default openstack zuul jobs in my zuul instance?! :) | 13:02 |
| *** myoung|pto has quit IRC | 13:03 | |
| *** _ari_ has quit IRC | 13:03 | |
| *** _ari_ has joined #zuul | 13:04 | |
| *** myoung has joined #zuul | 13:04 | |
| fungi | goern: https://zuul-ci.org/docs/zuul-jobs/install.html | 13:27 |
| fungi | is that what you're looking for? | 13:27 |
| fungi | goern: oh, you're asking about the content of the openstack-zuul-jobs repository rather the zuul-jobs repository? basically the same except you need git.openstack.org and the openstack-infra/openstack-zuul-jobs project | 13:29 |
| *** Shrews has joined #zuul | 13:34 | |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Remove credentials from urls in merger log https://review.openstack.org/578803 | 13:45 |
| tobiash | fungi, corvus: what do you think about having credentials in service logs? | 13:46 |
| tobiash | we noticed that the merger logs the github access tokens used in the git remote url which is pretty bad in our case as we have a check config job that runs zuul on all repos | 13:46 |
| *** rlandy|afk is now known as rlandy | 13:50 | |
| mordred | tobiash: I think we should not have credentials in service logs - also, I think we've been conservative about making service logs more available because we're not sure that we aren't leaking anything in to them | 14:05 |
| tobiash | mordred: we're also not making them available except in the check-config job where we need to ;) | 14:09 |
| mordred | :) | 14:11 |
| Shrews | tristanC: i think i see the cause of the problem you encountered | 14:17 |
| goern | fungi, and that is added to zuul and will make the openstack jobs generally available?! | 14:18 |
| fungi | goern: yes, that would make the jobs, playbooks and roles defined in the openstack-infra/openstack-zuul-jobs repository available. it would not include those defined in openstack-infra/project-config however | 14:19 |
| fungi | nor would it include jobs defined in the repositories of individual projects | 14:20 |
| goern | fungi, got that, thx | 14:20 |
| fungi | so, for example, if you wanted devstack jobs you would need to track the openstack-dev/devstack configuration | 14:20 |
| fungi | as that is where devstack-based jobs are defined | 14:21 |
| *** weshay|ruck has quit IRC | 14:27 | |
| *** _ari_ has quit IRC | 14:27 | |
| *** _ari_ has joined #zuul | 14:28 | |
| *** weshay has joined #zuul | 14:28 | |
| *** CrayZee has joined #zuul | 14:33 | |
| *** weshay has quit IRC | 14:37 | |
| *** myoung has quit IRC | 14:37 | |
| *** _ari_ has quit IRC | 14:37 | |
| *** pabelanger has quit IRC | 14:37 | |
| *** weshay has joined #zuul | 14:38 | |
| *** pabelanger has joined #zuul | 14:38 | |
| *** _ari_ has joined #zuul | 14:38 | |
| *** myoung has joined #zuul | 14:44 | |
| openstackgerrit | Merged openstack-infra/zuul master: Remove credentials from urls in merger log https://review.openstack.org/578803 | 14:44 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: Consolidate navigation functions https://review.openstack.org/578385 | 14:46 |
| mordred | tristanC: ^^ I'm curious to know if that works with your suburl deployment too | 14:47 |
| openstackgerrit | David Shrewsbury proposed openstack-infra/nodepool master: WIP: Fix for pools with diff labels https://review.openstack.org/578831 | 14:49 |
| Shrews | tristanC: ^^ above should correct the issue you see (please verify for me), but I need to figure out how to test that (which is a bit tricky)> | 14:50 |
| Shrews | Is anyone else having IRC issues the past couple of days? I am continually getting disconnected | 14:57 |
| fungi | not me | 14:58 |
| fungi | well, not with freenode anyway | 14:58 |
| Shrews | yeah, freenode is what i meant | 14:58 |
| fungi | i've been getting a lot of disconnects with oftc but freenode's been fine | 14:58 |
| Shrews | maybe i should rebuild my bouncer | 14:59 |
| *** CrayZee is now known as snapiri- | 15:00 | |
| *** Shrews has quit IRC | 15:20 | |
| *** Shrews has joined #zuul | 15:26 | |
| Shrews | fungi: do you connect to chat.freenode.net? or do you connect to a specific server? half of my pings to chat.f.n result in packet loss, i guess depending on which server it resolves to | 15:34 |
| fungi | freenode.addresses = "chat.freenode.net/6697" | 15:35 |
| fungi | over ipv6 | 15:35 |
| fungi | from a shell server in rax-iad | 15:35 |
| Shrews | i see it happen on my vexxhost bouncer and my local laptop | 15:36 |
| Shrews | frustrating | 15:36 |
| fungi | looks like i'm currently hitting wilhelm.freenode.net | 15:36 |
| fungi | and you're on moon.freenode.net | 15:36 |
| Shrews | yeah. i dunno what's going on | 15:39 |
| openstackgerrit | Merged openstack-infra/zuul master: Don't add implied branch matchers to project-pipeline variants https://review.openstack.org/577881 | 15:39 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul master: Fix typos in docs https://review.openstack.org/578854 | 15:44 |
| corvus | i'm on orwell over ipv4 | 15:44 |
| *** Shrews has quit IRC | 15:45 | |
| *** Shrews has joined #zuul | 15:47 | |
| Shrews | *sigh* | 15:48 |
| pabelanger | are we on track for zuul 3.1.1 today / tomorrow? Or looking more like next week? Wouldn't mind a release to pick up new HDD sensors for executor | 15:49 |
| corvus | Shrews: i'm on orwell over ipv4 | 15:50 |
| clarkb | Shrews: if connectes via ipv6 maybe try ipv4 or vice versa? | 15:50 |
| corvus | pabelanger: i'd like to let the js stuff land again, then restart all of openstack, then release. could happen tomorrow. maybe next week. | 15:51 |
| corvus | (but as of now, we haven't restarted the executors in a while) | 15:52 |
| *** hashar is now known as hasharAway | 15:52 | |
| pabelanger | corvus: wfm mostly curious what else we wanted to do for 3.1.1, thanks! | 15:52 |
| clarkb | corvus: mordred is the js stuff ready for review again? | 15:52 |
| corvus | Shrews: regarding containers spec -- yes, i was thinking the same thing about the nodepool bits being handwavey. do you think we should expand on that a bit more in the spec now, or do you think we should say the spec sufficiently articulates our shared intention, and try out options in changes/code review? | 15:54 |
| mordred | clarkb: I thought so - but just found another failure in local test | 15:54 |
| Shrews | corvus: well, i don't think we need to go too deep, but maybe similarly as deep as the zuul parts you've outlined. | 15:58 |
| Shrews | corvus: "nodepool would need to create the requested container" ... do you have a particular ansible module in mind? | 15:59 |
| Shrews | or some other method? | 15:59 |
| Shrews | same for k8s service | 15:59 |
| corvus | Shrews: i'm assuming we'd use the oc (openshift) library -- it seems simple enough and is what the ansible modules use, and works with openshift and k8s. but i don't have a strong opinion on that (whatever gets the job done). i can add that in there. | 16:02 |
| Shrews | zuul seems to depend on kubectl quite a bit in the spec, so that's at least some detail that nodepool doesn't have similarly mentioned | 16:02 |
| Shrews | (if those words make sense) | 16:02 |
| corvus | yep | 16:02 |
| corvus | Shrews: the thing i'm least sure about is what it should look like to create a k8s namespace (our first non-node thing). should we just use a label for that, or should we create an entirely new concept? | 16:03 |
| corvus | Shrews: i'm inclined to say "lets start with just using labels and see how far we get with that" | 16:03 |
| Shrews | corvus: yeah, i can't answer that since i'm pretty k8s ignorant. i guess i should spend some time learning it | 16:03 |
| Shrews | which i guess is the main reason why i'd like to see a bit more detail so i'd know where to start | 16:04 |
| corvus | ok. i'll add a little more detail to flesh out these areas. thanks. | 16:04 |
| corvus | i should be able to do that today/tomorrow and maybe we can shift into really thinking about this next week. or maybe just have beer and hotdogs. | 16:05 |
| Shrews | the 2nd thing | 16:05 |
| mordred | in case anyone was wondering - in javascript, [''].join('/') returns '' not '/' | 16:05 |
| corvus | ya | 16:05 |
| Shrews | mordred: get outta muh head! | 16:06 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: Consolidate navigation functions https://review.openstack.org/578385 | 16:07 |
| corvus | in python, '/'.join(['']) returns '' | 16:07 |
| mordred | corvus: yah | 16:07 |
| mordred | it's a reasonable thing to return | 16:07 |
| corvus | it's still backwards land, but at least it doesn't return 'True' :) | 16:08 |
| mordred | it's just that return path.split('/').slice(0, -1).join('/') doesn't actually return / if you had a short url | 16:08 |
| corvus | [''].join('/') -> OK | 16:08 |
| mordred | return path.split('/').slice(0, -1).join('/') || '/' ... on the other hand, does | 16:08 |
| mordred | clarkb, corvus: ok - https://review.openstack.org/578385 seems to work for me running it in apache without a ZUUL_API_URL setting set | 16:09 |
| mordred | and also in the two test scenarios we do in the gate | 16:10 |
| mordred | so review welcome now | 16:10 |
| mordred | if we're happy with the stack, we can squash it into the revert revert | 16:10 |
| clarkb | mordred: ok I will take a look | 16:11 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: web: fix status page flickering https://review.openstack.org/578226 | 16:11 |
| mordred | tristanC: local testing seems to show that ^^ working with your suburl oranization too | 16:19 |
| mordred | sorry, https://review.openstack.org/578385 | 16:20 |
| *** Shrews has quit IRC | 16:21 | |
| *** Shrews has joined #zuul | 16:23 | |
| clarkb | mordred: cool that does confirm we needed the path too (which was my concern with doing a large scale refactoring myself | 16:28 |
| *** Shrews has quit IRC | 16:29 | |
| mordred | clarkb: awesome. so the difference makes at least some amount of sense then? | 16:29 |
| clarkb | mordred: yes, basically we need both the root path an dthe root url | 16:30 |
| clarkb | and we need them in separate places | 16:30 |
| clarkb | for the git log disection it might be better to not squash all these fixes and instead imrpove them at the same time as the revert revert? | 16:32 |
| corvus | i'm ambivalent on that point. i +2d them all, but would be happy to re +2 if we decide to squash | 16:33 |
| clarkb | heh improve == approve | 16:34 |
| clarkb | corvus: mordred I'm up for approving the stack now and watching it deploy | 16:39 |
| *** Shrews has joined #zuul | 16:44 | |
| mordred | clarkb, corvus: kk. I'm good with that | 16:54 |
| mordred | clarkb, corvus: I clicked the button - will track/watch | 16:55 |
| *** Shrews has quit IRC | 17:00 | |
| *** jpena is now known as jpena|off | 17:01 | |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: Translate zuulStartStream into typescript https://review.openstack.org/558618 | 17:02 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: Shift log streaming code into StreamComponent https://review.openstack.org/558619 | 17:02 |
| *** Shrews has joined #zuul | 17:03 | |
| mordred | clarkb, corvus: if you still have javascript things paged in ^^ those are essentially just code organization | 17:04 |
| mordred | corvus: also - did we ever get anywhere on fontaweseome vs. glyphicons vs. something else? I feel like there was a good discussion about that but I forget where it got | 17:08 |
| *** Shrews has quit IRC | 17:09 | |
| *** electrofelix has quit IRC | 17:10 | |
| *** Shrews has joined #zuul | 17:11 | |
| clarkb | mordred: I'm going to assume `(${params['logfile']})` is just as magic in ts as it was in js (I also have no idea what those operators mean but have inferred that it gives us the file suffix) | 17:15 |
| *** Shrews has quit IRC | 17:16 | |
| *** Shrews has joined #zuul | 17:17 | |
| mordred | clarkb: ya - the `` quotes is an ES6 feature (also in typescript) that lets you do variable interpolation from the surrounding context | 17:18 |
| mordred | so instead of '(' + params['logfile'] + ')' | 17:18 |
| clarkb | mordred: there is something weird to me about makking a mutable state object const (the websocket) | 17:28 |
| openstackgerrit | Merged openstack-infra/zuul master: Revert "Revert "Upgrade from angularjs (v1) to angular (v6)"" https://review.openstack.org/578384 | 17:33 |
| openstackgerrit | Merged openstack-infra/zuul master: Use window.location.href for the href not pathname https://review.openstack.org/578243 | 17:33 |
| mordred | clarkb: yah. const in typescript is a bit different than const in c++ - it seems to mean "this isn't going to get re-assigned" not "the object assigned to this variable is immutable" | 17:35 |
| mordred | so I guess it's most like const * than const & | 17:35 |
| clarkb | for https://review.openstack.org/#/c/558619/13 and beyond it might be good to see that the above changes that are merging are good first before we add to potential revery backlog | 17:40 |
| clarkb | I'm actually going to remove my +W on the parent of that for this reason too | 17:41 |
| clarkb | but if angular v2/5/6 is happy then we can start pushing those in next | 17:41 |
| mordred | https://www.typescriptlang.org/docs/handbook/variable-declarations.html is a good read, fwiw | 17:42 |
| mordred | clarkb: ++ | 17:42 |
| *** Shrews has quit IRC | 17:45 | |
| clarkb | mordred: is my change sufficient for having a working dashboard? or should I disable ansible puppet on zuul01 until we get thorugh the consolidation change? | 17:45 |
| corvus | they're gonna land really close to each other anyway, right? if it breaks between ansible cycles, we can just manually reinstall it | 17:46 |
| clarkb | yes, looking at the zuul node request backlog we should see the consolidation change take the second spot in the new supercedent post queu | 17:47 |
| clarkb | and then run soon after the revert revert post queue jobs | 17:47 |
| *** snapiri- has quit IRC | 17:47 | |
| *** Shrews has joined #zuul | 17:58 | |
| *** pcaruana has quit IRC | 18:14 | |
| corvus | mordred: hrm, i'm in the same headspace as you with icons -- i know we started with me saying "maybe we can use fontawesome since that's what we use on the web site" but i don't know where we went after that. i'll try to dig up logs | 18:17 |
| corvus | http://eavesdrop.openstack.org/irclogs/%23zuul/%23zuul.2018-06-08.log.html#t2018-06-08T14:26:17 | 18:19 |
| corvus | mordred: my read of that is that it was inconclusive, but i think the least-bad thing is to use fontawesome for consistency | 18:21 |
| *** Shrews has quit IRC | 18:23 | |
| corvus | fungi: https://review.openstack.org/578854 | 18:24 |
| *** yolanda_ has joined #zuul | 18:25 | |
| *** yolanda has quit IRC | 18:27 | |
| openstackgerrit | Merged openstack-infra/zuul master: Consolidate navigation functions https://review.openstack.org/578385 | 18:29 |
| *** Shrews has joined #zuul | 18:29 | |
| *** Shrews_ has joined #zuul | 18:32 | |
| fungi | corvus: thanks for the followup! | 18:33 |
| *** Shrews_ has quit IRC | 18:34 | |
| *** Shrews has quit IRC | 18:35 | |
| mordred | corvus: gotcha | 18:35 |
| *** Shrews has joined #zuul | 18:37 | |
| mnaser | mordred: did you get a chance to review our patches that got reverted? :( | 18:49 |
| Shrews | corvus: re: nodepool tag... i'm fairly confident that https://review.openstack.org/578831 will fix the issue tristanC is seeing, but I'm having a devil of a time coming up with a test for it. If we are in a hurry to tag, I can unWIP and figure out the test later. | 18:49 |
| Shrews | I am also open to suggestions for the test | 18:50 |
| mnaser | today is a good candidate for the filters.. my browser is getting destroyed watching zuul.o.o | 18:50 |
| mordred | mnaser: they just got re-added | 18:52 |
| mnaser | mordred: oh yay | 18:53 |
| mnaser | clarkb: if infra is switching things up soon, we can troubleshoot live quickly i guess | 18:53 |
| clarkb | mnaser: we will redeploy it as soon as zuul post jobs run then our puppet runs | 18:53 |
| * mnaser should be around most of the day | 18:54 | |
| corvus | Shrews: i'm not in favor of being in a hurry :) | 18:57 |
| fungi | hurry is too much like hassle | 18:58 |
| fungi | or... work | 18:58 |
| *** gtema has quit IRC | 18:59 | |
| corvus | Shrews: though, honestly, i could skip the test on this one. the system worked, it just spewed an extra error. | 19:00 |
| Shrews | corvus: then let's let tristanC verify the fix for us and merge it based on that. We don't have any mechanism in place for tests to get the reason WHY we declined a request | 19:01 |
| corvus | wfm | 19:02 |
| Shrews | or to coordinate pool threads processing the request list :/ | 19:03 |
| openstackgerrit | David Shrewsbury proposed openstack-infra/nodepool master: Fix for pools with different labels https://review.openstack.org/578831 | 19:05 |
| openstackgerrit | Merged openstack-infra/zuul master: Fix typos in docs https://review.openstack.org/578854 | 19:09 |
| Shrews | hrm, 30+ minutes without an freenode disconnect after turning off SASL authentication. | 19:12 |
| Shrews | and now my bouncer instance in vexxhost is not starting up >:( | 19:18 |
| corvus | NO CARRIER | 19:19 |
| Shrews | mnaser: no vexxhost issue currently, are there? | 19:19 |
| Shrews | i might just give up on this day | 19:19 |
| mnaser | Shrews: hm no, instance id? | 19:19 |
| mnaser | or you can pm if you'd liek | 19:19 |
| mnaser | s/liek/like/ | 19:19 |
| Shrews | mnaser: 6745215a-a79d-4a2e-b847-5e6a1738d1e4 | 19:19 |
| mnaser | nf_conntrack: table full, dropping packet | 19:22 |
| mnaser | oh well | 19:22 |
| Shrews | O.o | 19:22 |
| mnaser | i guess somehow something wild happened | 19:23 |
| mnaser | in this hypervisor | 19:23 |
| Shrews | neat | 19:23 |
| mnaser | one second | 19:23 |
| mnaser | (it really isnt your day i guess) | 19:23 |
| *** Shrews has quit IRC | 19:33 | |
| *** Shrews has joined #zuul | 19:34 | |
| *** Shrews has joined #zuul | 19:35 | |
| fungi | Shrews: good news is that at least likely explains all your earlier disconnects | 19:41 |
| *** Shrews has quit IRC | 19:44 | |
| *** Shrews has joined #zuul | 19:49 | |
| Shrews | fungi: nope, only freenode connection is affected. and it's happening locally on my laptop too. and it just happened 2 more times | 19:50 |
| Shrews | so i quit | 19:50 |
| Shrews | like, literally. i'm going off irc to avoid connection flap spam in the channels. i'll monitor eavesdrop and just try again tomorrow. | 19:51 |
| *** Shrews has quit IRC | 19:51 | |
| tobiash | pabelanger: I've added a question on https://review.openstack.org/#/c/578181 | 19:58 |
| pabelanger | tobiash: no, you are right now. That function just results a list, so we remove the loop | 20:01 |
| pabelanger | fixing | 20:01 |
| openstackgerrit | Paul Belanger proposed openstack-infra/zuul master: Add zuul.child_jobs in ansible inventory file https://review.openstack.org/578181 | 20:09 |
| pabelanger | tobiash: ^should be better | 20:09 |
| tobiash | pabelanger: lgtm | 20:12 |
| pabelanger | tobiash: thanks! | 20:13 |
| pabelanger | is https://docs.openstack.org/infra/nodepool/operation.html#removing-a-provider still valid, specifically max-servers: -1 | 20:17 |
| pabelanger | it seems max-server: -1 might not delete online nodes any more | 20:18 |
| pabelanger | trying to confirm | 20:18 |
| pabelanger | yah, looks like a bug. nodepool doesn't seem to delete a ready node if you set max-servers: -1 or labels: [] for that provider | 20:40 |
| pabelanger | I'll see about adding a unit test to confirm | 20:40 |
| openstackgerrit | Clark Boylan proposed openstack-infra/zuul-jobs master: Dynamically determine overlay network mtu https://review.openstack.org/578153 | 20:48 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul master: WIP: Match tag items against containing branches https://review.openstack.org/578557 | 21:07 |
| pabelanger | clarkb: corvus: mordred : with https://review.openstack.org/578230/ is my first attempt at using zuul_return zuul.child_jobs to skip children. However, right now zuul still does a -1 on result because of the SKIPPED jobs. So, I am think either we agree SKIPPED job in that case are +1 or maybe a new result to indicate to use that zuul_return dynamically skipped those child jobs (but struggling to come up with | 21:12 |
| pabelanger | new flag). | 21:12 |
| clarkb | pabelanger: maybe just update SKIPPED to be +/-0 | 21:13 |
| pabelanger | clarkb: and not affect the exiting +1 if all other jobs passed? | 21:15 |
| corvus | that sounds like it should be okay. presumably if you want to require the children to run, you'd have the parent job voting. | 21:15 |
| corvus | (this would be a modification to didAllJobsSucceed() | 21:15 |
| clarkb | pabelanger: ya, basically if a job that did ran failed then it is still a -1 but if all the jobs that did run are +1 call it good | 21:16 |
| pabelanger | if a child job was voting and it was skipped, that would still be +/-0 right? | 21:16 |
| clarkb | pabelanger: I think so as long as we are accounting the parent vote properly now (which we should be with the ability to use zuul return for this) | 21:16 |
| pabelanger | okay, that should work for this use case | 21:17 |
| pabelanger | I'll look at didAllJobsSucceed() shortly | 21:17 |
| corvus | i think as a special case, we would need didAllJobsSucceed to return False if all jobs are skipped. there are some places which assume that. | 21:19 |
| pabelanger | +1 | 21:20 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul master: WIP: Report to gerrit over HTTP https://review.openstack.org/577027 | 21:25 |
| *** hasharAway has quit IRC | 21:32 | |
| openstackgerrit | Paul Belanger proposed openstack-infra/zuul master: Support skip_child_jobs via zuul_return https://review.openstack.org/578230 | 22:05 |
| *** pwhalen has quit IRC | 22:07 | |
| *** rlandy has quit IRC | 22:12 | |
| corvus | pabelanger: +2 on the first change, -1 on the second (also some doc typos) | 22:16 |
| pabelanger | thanks, will look shortly | 22:16 |
| pabelanger | corvus: so rather then self.addBuild() I should be using self.setResult(), did I understand correct? | 22:38 |
| corvus | pabelanger: i mean that the stuff you have in findJobsToRun (creating fakebuilds with skipped results) should be in setResult | 22:39 |
| corvus | pabelanger: that's where the current code for doing the same thing for failing parent jobs happens | 22:39 |
| corvus | findJobsToRun should be 'read-only' | 22:39 |
| pabelanger | okay, let me move and testing | 22:43 |
| pabelanger | test* | 22:43 |
| *** Shrews has joined #zuul | 22:48 | |
| Shrews | pabelanger: https://docs.openstack.org/infra/nodepool/operation.html#removing-a-provider doesn't say ready nodes will be removed when max-servers is set to -1. It says they'll go through the normal lifecycle, but no more will be created. | 22:49 |
| Shrews | "If urgency is required you can delete the nodes directly instead of waiting for them to go through their normal lifecycle but the effect is the same." | 22:50 |
| pabelanger | Shrews: thanks, I confused it with nodepoolv2 where it would delete ready nodes with -1 | 22:53 |
| pabelanger | but I think there is an issue setting it to max-servers: -1, I think we had an exception on clean up handler | 22:53 |
| pabelanger | need to test some more | 22:53 |
| Shrews | pabelanger: i'm not certain v2 did that either, but i'm not that familiar with that version | 22:54 |
| *** threestrands has joined #zuul | 23:05 | |
| *** threestrands has quit IRC | 23:05 | |
| *** threestrands has joined #zuul | 23:05 | |
| *** yolanda__ has joined #zuul | 23:07 | |
| *** yolanda_ has quit IRC | 23:11 | |
| tristanC | Shrews: thanks for the prompt update, I'll pick 578831 shorty | 23:11 |
| tristanC | Shrews: and how about we used invalidNodeTypes here http://git.openstack.org/cgit/openstack-infra/nodepool/tree/nodepool/launcher.py#n177 ? | 23:14 |
| tristanC | Shrews: wouldn't that help keeping the declined_by list short and relevant? | 23:16 |
| *** yolanda__ has quit IRC | 23:18 | |
| openstackgerrit | Paul Belanger proposed openstack-infra/zuul master: Support skip_child_jobs via zuul_return https://review.openstack.org/578230 | 23:26 |
| pabelanger | corvus: okay, I believe I updated setResult() per your comments^. There is likey an optimization there I am missing | 23:27 |
| pabelanger | also, is there any way to have stestr for zuul dynamically skip tests if database isn't setup? | 23:27 |
| pabelanger | at least on local environment | 23:28 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul-jobs master: WIP: Add a role to return file comments https://review.openstack.org/579033 | 23:28 |
| clarkb | pabelanger: you can provide a regex to exclude or only include the tests you want | 23:28 |
| clarkb | pabelanger: I tend to do something like tox -epy36 -- test_scheduler | 23:29 |
| pabelanger | clarkb: yah, I do that too but usually run everthing once before pushing up the change. Not a big issues, I can restart database, just haven't yet | 23:30 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul master: WIP: Support line comments in Gerrit https://review.openstack.org/577035 | 23:31 |
| corvus | tobiash, mordred: https://review.openstack.org/579033 might get us the correct line numbers for reporting on syntax errors after speculative merges | 23:31 |
| corvus | at least, the process works in my very simple contrived example. i haven't tested that code yet. | 23:32 |
| *** threestrands has quit IRC | 23:38 | |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul-jobs master: WIP: Add a role to return file comments https://review.openstack.org/579033 | 23:38 |
| tristanC | mordred: i think we should create a selenium test to validate the JS code actually work as expected. I meant to rewrite our webui test in zuulv3 but maybe i can start a new one from scratch for zuul? | 23:43 |
| corvus | i think that would be swell | 23:43 |
| tristanC | is there a role/playbook we can use to setup a working zuul-web? | 23:44 |
| clarkb | tristanC: there is the tox target | 23:44 |
| clarkb | I would probably run it out of the test framework that way | 23:45 |
| tristanC | then it shouldn't be difficult to install apache vhost with rewrite rules for the different configuration, and start writting simple js validation using selenium | 23:45 |
| clarkb | do you need apache? avoiding that would make it easier to run locally with just `tox` | 23:45 |
| tristanC | clarkb: i think we need apache to reproduce zuul.openstack.org white-label deployement, e.g.: https://zuul-ci.org/docs/zuul/admin/installation.html#white-labeled-tenant | 23:47 |
| tristanC | clarkb: what's the "tox target" ? | 23:47 |
| corvus | if you run zuul-web out of the test framework, there's already a proxy fixture | 23:48 |
| clarkb | tristanC: any of the py3* targets, they run nodeenv installation and then yarn install and npm run build:dev | 23:48 |
| clarkb | basically the python unittest framework already has a built web installed into it and there are fixtures to interact with it | 23:49 |
| corvus | http://git.zuul-ci.org/cgit/zuul/tree/tests/unit/test_web_urls.py#n74 | 23:49 |
| corvus | that one runs zuul-web behind a proxy in openstack's configuration | 23:49 |
| tristanC | corvus: clarkb: oh i see, then we could extend that TestWebURLs with something like https://selenium-python.readthedocs.io/getting-started.html#simple-usage | 23:52 |
| corvus | yeah, that looks great | 23:55 |
| tristanC | Would you mind having firefox as a bindep test requirement, or should i rather propose a new zuul-jobs to prepare the node correctly? | 23:57 |
| clarkb | if its test only bindep I think its fine | 23:58 |
| tristanC | we used to also setup a Xnest so that we could record the session with ffmpeg and store the video in the jobs artifacts... | 23:58 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!