| openstackgerrit | Jesse Keating proposed openstack-infra/zuul feature/zuulv3: Allow using webapp from connections https://review.openstack.org/439831 | 00:00 |
|---|---|---|
| openstackgerrit | Jesse Keating proposed openstack-infra/zuul feature/zuulv3: Support GitHub PR webhooks https://review.openstack.org/439834 | 00:05 |
| jlk | auggy: thanks for reminding me to link back to a storyboard story :D | 00:06 |
| auggy | jlk: no prob, just don't want anyone else to pick it up and cross the streams | 00:22 |
| rbergeron | spamaps: sorry i couldnt join, had a dentist appointment that i already cancelled once | 00:39 |
| *** jamielennox is now known as jamielennox|away | 01:16 | |
| jeblair | jlk: yes, that is our webapp tenant isolation story to date. it's entirely possible we may want to put real access control on that in the future. :) however, anticpating what you're working on -- connections/drivers are just about the only things that actually span tenants. so i would expect a github driver to have an endpoint which did not include a tenant name. so maybe the hooks should go under "hostname/github/<connectionname>". | 01:23 |
| jeblair | the use of the word 'hostname' in that is confusing. let's just pretend i said "/github/<connectionname>". | 01:25 |
| openstackgerrit | Clint 'SpamapS' Byrum proposed openstack-infra/zuul feature/zuulv3: Re-enable test_timer https://review.openstack.org/430968 | 04:16 |
| openstackgerrit | Clint 'SpamapS' Byrum proposed openstack-infra/zuul feature/zuulv3: Re-enable test_timer_smtp https://review.openstack.org/430983 | 04:16 |
| openstackgerrit | Clint 'SpamapS' Byrum proposed openstack-infra/zuul feature/zuulv3: Re-enable test_idle https://review.openstack.org/430568 | 04:16 |
| openstackgerrit | Clint 'SpamapS' Byrum proposed openstack-infra/zuul feature/zuulv3: Re-enable test_dependent_behind_dequeue https://review.openstack.org/437076 | 04:16 |
| SpamapS | jeblair: ^ addressed comment on 430568 | 04:16 |
| *** yolanda has quit IRC | 05:35 | |
| *** Cibo_ has joined #zuul | 06:19 | |
| *** saneax-_-|AFK is now known as saneax | 06:26 | |
| *** saneax is now known as saneax-_-|AFK | 06:52 | |
| *** saneax-_-|AFK is now known as saneax | 07:02 | |
| *** Cibo_ has quit IRC | 07:20 | |
| *** isaacb has joined #zuul | 07:22 | |
| *** saneax is now known as saneax-_-|AFK | 07:24 | |
| *** Cibo_ has joined #zuul | 07:28 | |
| *** saneax-_-|AFK is now known as saneax | 07:32 | |
| *** Cibo_ has quit IRC | 07:34 | |
| *** isaacb has quit IRC | 07:56 | |
| *** isaacb has joined #zuul | 07:59 | |
| *** hashar has joined #zuul | 09:11 | |
| *** isaacb has quit IRC | 10:19 | |
| *** isaacb has joined #zuul | 11:10 | |
| *** Cibo_ has joined #zuul | 11:36 | |
| *** wznoinsk has quit IRC | 12:39 | |
| *** Shrews has quit IRC | 13:11 | |
| *** Shrews has joined #zuul | 13:15 | |
| Shrews | oh neat. we can now access zuulv3 logs \o/ | 13:22 |
| *** Cibo has joined #zuul | 13:34 | |
| *** Cibo_ has quit IRC | 13:34 | |
| *** Cibo has quit IRC | 13:46 | |
| *** saneax is now known as saneax-_-|AFK | 13:58 | |
| *** isaacb has quit IRC | 14:58 | |
| pabelanger | Shrews: yar! | 15:05 |
| pabelanger | wouldn't mind some eyes on https://review.openstack.org/#/c/439637/ this morning, currently jobs are capped at 6min run times | 15:06 |
| mordred | look, it's a Shrews ! | 15:08 |
| mordred | Shrews: while you were gone, we rewrote nodepool in haskell - hope that's ok | 15:09 |
| *** Cibo has joined #zuul | 15:17 | |
| Shrews | mordred: totes | 15:26 |
| *** isaacb has joined #zuul | 15:55 | |
| *** isaacb has quit IRC | 16:04 | |
| *** yolanda has joined #zuul | 16:05 | |
| *** hashar has quit IRC | 16:20 | |
| Shrews | jeblair: I'd like to clarify this part of your algorithm: "If request < quota and request > available nodes (due to current usage), begin satisfying the request and do not process further requests until satisfied" | 16:43 |
| Shrews | jeblair: Does that mean "go ahead and exceed max_servers to satisify this one request, ignoring other requests until this one is done"? | 16:44 |
| Shrews | jeblair: Or does it mean "spin on this request until the needed number of nodes appear"? | 16:44 |
| Shrews | i'm guessing the 2nd, but want to verify | 16:45 |
| Shrews | well, "spin + ignore others" | 16:48 |
| *** Cibo has quit IRC | 16:48 | |
| jeblair | Shrews: yes, the second. | 16:50 |
| Shrews | jeblair: ok. will have to come up with a clever way to flag such a situation. thanks | 16:51 |
| Shrews | ugh, this might actually be a bit complicated. yay for returning to work! | 16:57 |
| Shrews | ah, i think moving the quota and availability checks up from NodeRequestHandler to the parent ProviderWorker makes this easy | 17:10 |
| pabelanger | jeblair: nice email. I like the trusted-projects / untrusted-projects changes to yaml | 17:50 |
| jeblair | Shrews: that sounds reasonable | 17:53 |
| jeblair | pabelanger: ah thanks. also, folks in channel, i sent an email to openstack-infra@lists about some proposed changes to clarify confusion around having projects from multiple sources in the same pipelines and depending on each other and such. | 17:54 |
| Shrews | i concur re: the trusted-projects/untrusted-projects label change | 18:02 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Report layout config errors for config repos https://review.openstack.org/440760 | 18:52 |
| jeblair | when we land that change and its parent, it should be safe to throw whatever garbage we want at our running zuulv3 and it will throw it right back at us! | 18:52 |
| jeblair | (in case it's not clear: that is a desirable thing) | 18:53 |
| *** Cibo has joined #zuul | 18:57 | |
| pabelanger | cool 439637 would be nice to get too | 19:09 |
| *** hashar has joined #zuul | 19:38 | |
| *** hashar has quit IRC | 19:42 | |
| *** hashar has joined #zuul | 19:42 | |
| openstackgerrit | David Shrewsbury proposed openstack-infra/nodepool feature/zuulv3: Correct the quota/availability algorithm https://review.openstack.org/440780 | 19:43 |
| jlk | Oh boy, somehow I got tests to non-stop be executing git commands in popen, really really fast. No real errors visible. GOOD TIMES | 19:46 |
| jeblair | jlk: i should document my tips for running zuul tests on a tmpfs... | 19:48 |
| jlk | during tox tests does zuul services log anywhere ? | 19:50 |
| clarkb | jlk: yes to a logging fixture which injects the logs into the subunit. On successes you won't see it, but on failures its all printed to stdout | 19:51 |
| jlk | I'm just seeing InterruptedError() occasionally from RPCListner | 19:51 |
| clarkb | jlk: to find the successes look in .testrepository/$number | 19:51 |
| clarkb | or commands like testr last --subunit iirc | 19:52 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Report layout config errors for config repos https://review.openstack.org/440760 | 19:57 |
| jeblair | jlk: InterruptedError is normal; we should trap it and ignore it. it's part of the shutdown. | 19:59 |
| jlk | oh. | 19:59 |
| jlk | then I don't know why tests are doing this. | 19:59 |
| openstackgerrit | Jesse Keating proposed openstack-infra/zuul feature/zuulv3: Support GitHub PR webhooks https://review.openstack.org/439834 | 20:00 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Catch gear.InterruptedError https://review.openstack.org/440789 | 20:02 |
| *** jamielennox|away is now known as jamielennox | 20:02 | |
| jeblair | jlk: ^ i had developed a blindness to that. thank you for reminding me it's annoying. :) | 20:02 |
| openstackgerrit | Cullen Taylor proposed openstack-infra/zuul feature/zuulv3: Re-enable test_tags https://review.openstack.org/439858 | 20:48 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Catch gear.InterruptedError https://review.openstack.org/440789 | 20:50 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Report layout config errors for config repos https://review.openstack.org/440760 | 20:50 |
| rcarrillocruz | pst pst | 20:51 |
| * rcarrillocruz uber excited that just saw Zuul v3 mentioned in an Ansible 2.4 roadmap | 20:51 | |
| pabelanger | rcarrillocruz: woah, URL? | 20:53 |
| rcarrillocruz | not public :P | 20:54 |
| rcarrillocruz | it appears is been pushed for 3rd party CI on vendor modules | 20:54 |
| rcarrillocruz | mordred, Shrews ^ | 20:54 |
| rcarrillocruz | i'm told you're pushing for it | 20:54 |
| rcarrillocruz | and just volunteered to get stuff assigned | 20:55 |
| rcarrillocruz | i mean, I volunteers | 20:56 |
| rcarrillocruz | d | 20:56 |
| rcarrillocruz | i can't even type | 20:56 |
| Shrews | i'm not aware of timelines around that. it may be speculation | 21:00 |
| clarkb | when is 2.3 happening so wr can have sync retries that work? | 21:01 |
| * clarkb is only slightly impatient :) | 21:01 | |
| rcarrillocruz | hah | 21:04 |
| pabelanger | clarkb: Shrews: mind a review on 439637 | 21:18 |
| clarkb | pabelanger: is that just pulled over from v2.5 | 21:28 |
| clarkb | ? | 21:28 |
| pabelanger | clarkb: ya, built on from v2.5, but some changes | 21:29 |
| pabelanger | to watchdog | 21:29 |
| pabelanger | basically, we support jobs that run forever now | 21:29 |
| pabelanger | and remove 5m buffer to abort jobs | 21:30 |
| clarkb | because ansible is responsible for running job and copying logs ya? | 21:31 |
| pabelanger | right, we better control that in post-run | 21:33 |
| clarkb | pabelanger: I approved but also left the tiniest of nitpicks for possible followup | 21:39 |
| pabelanger | clarkb: ack | 21:40 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Add timeout variable to jobs https://review.openstack.org/439637 | 21:47 |
| pabelanger | jeblair: Shrews: clarkb: objections to restarting zuulv3-dev to pick up^? | 21:49 |
| clarkb | none from me | 21:49 |
| Shrews | Nope | 21:49 |
| pabelanger | danke | 21:51 |
| jeblair | pabelanger, clarkb, mordred: can you review 439887 and 440760 so we can have config error reports? :) | 21:52 |
| rbergeron | there are no errors! everythign is perfect | 21:53 |
| *** anteaya has quit IRC | 21:54 | |
| pabelanger | jeblair: I can shortly | 21:55 |
| *** bstinson has quit IRC | 21:55 | |
| jeblair | rbergeron: zuul will go into an infinite loop telling itself that right now :) | 21:57 |
| *** bstinson has joined #zuul | 21:59 | |
| clarkb | jeblair: maybe I don't have enough sugar today but where is the error in the yaml at https://review.openstack.org/#/c/439887/3/tests/unit/test_v3.py ? | 22:07 |
| *** anteaya has joined #zuul | 22:07 | |
| jeblair | clarkb: it's where it says "foo: error" | 22:08 |
| * jeblair hands clarkb a chocolate bar | 22:08 | |
| clarkb | because foo isn't a valid key? | 22:08 |
| jeblair | ya | 22:09 |
| jeblair | maybe i should have said "error: foo" | 22:09 |
| jeblair | clarkb: i copied the full error report into the gerrit comments, so you can see what it look like | 22:09 |
| *** hashar has quit IRC | 22:12 | |
| clarkb | jeblair: thanks. Also is it worth having a test where the yaml itself is invalid? | 22:13 |
| jeblair | clarkb: probably so, i'll add one in a followup | 22:13 |
| openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Add generic tox job https://review.openstack.org/438281 | 22:21 |
| pabelanger | jeblair: I think I'm ready to report a feature request! With^ we now have duplicate playbooks, because job names have to point to a playbook. Maybe we can also add a job.playbook key to be used, if defined, otherwise fall back to job.name for the playbook? | 22:24 |
| jeblair | pabelanger: it's called 'job.run' | 22:29 |
| pabelanger | Yay! | 22:29 |
| jeblair | pabelanger: but what are the duplicate playbooks? | 22:30 |
| pabelanger | tox-cover, tox-linters, tox-py27 right now | 22:30 |
| pabelanger | because of the way they are setup | 22:30 |
| pabelanger | however, moving forward they might grow specific actions | 22:31 |
| jeblair | pabelanger: why not 3 playbooks and put the argument to the tox role in the playbook rather than the job? | 22:32 |
| pabelanger | jeblair: so, that is how I initially did it. But, now that we have job.vars, I could abstract that out more to a single playbook | 22:33 |
| pabelanger | jeblair: PS9 did that | 22:33 |
| openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Add generic tox job https://review.openstack.org/438281 | 22:34 |
| pabelanger | ^would be another way. | 22:34 |
| pabelanger | don't have a preference at this point | 22:35 |
| pabelanger | aside from flexing parts of zuulv3 | 22:35 |
| mordred | yah - seems like playing with both approaches to learn which thing 'feels' better is worthwhile at this point | 22:37 |
| pabelanger | the upside I think for 3 playbooks, with vars included, makes it easier for other zuulv3 to consume. Which, is something we want | 22:39 |
| jeblair | true. i think when it comes down to it, i'm going to have a preference that our fundamental jobs be really clear and easy to understand without too much ansible magic. i'd like us to have something to point to where a simple job is as easy to understand as jjb. | 22:39 |
| jeblair | it's the complicated jobs (ilke devstack) where the extra magic makes it *easier* to understand than jjb :) | 22:39 |
| pabelanger | ++ | 22:39 |
| mordred | yah | 22:39 |
| pabelanger | let me revert to PS9 | 22:40 |
| openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Add generic tox job https://review.openstack.org/438281 | 22:41 |
| pabelanger | okay, have to run errands with family. Plan to do some more hacking this evening | 22:42 |
| jeblair | pabelanger, mordred: i think it might be nice to have one or more competing changes for how to set up those tox jobs, and to try to get each approach looking the best. i think that might be the best way to do what mordred describes and achieve our goal. | 22:58 |
| jeblair | pabelanger: so start with a change for the PS9 approach, and another for PS10, and iterate on both of them for a bit | 22:58 |
| jeblair | pabelanger: okay, i left a bunch of comments on that change with things that may apply to either of those approaches. i think if you can incorporate those and make 2 new changes, we'll be off to the races. | 23:03 |
| *** saneax-_-|AFK is now known as saneax | 23:20 | |
| jlk | wooo, SO MANY LINES. This log has too many lines to open it in the browswer. | 23:32 |
| jlk | is there a easy RTFM for running a subset of tests with tox? | 23:33 |
| *** jamielennox is now known as jamielennox|away | 23:37 | |
| jeblair | jlk: the TESTING file has suggestions on how to run one test in the foreground. i think you can give a pattern match to testr to run a set. | 23:38 |
| jlk | oh hey look at that, I didn't read far enough. Derp. | 23:39 |
| pabelanger | jeblair: agree, I'll iterate on the comments | 23:44 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!