| corvus | https://pypi.org/project/zuul/3.1.0/ exists | 00:07 |
|---|---|---|
| tristanC | corvus: nice, thanks! | 00:13 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul master: Add ansible 2.5 release note https://review.openstack.org/575879 | 00:21 |
| corvus | clarkb, fungi, mordred, tobiash: oops, ^ we should probably land that asap | 00:21 |
| fungi | oh!!! | 00:22 |
| *** myoung|bbl is now known as myoung | 00:26 | |
| *** rlandy has quit IRC | 00:39 | |
| openstackgerrit | Merged openstack-infra/zuul master: Add ansible 2.5 release note https://review.openstack.org/575879 | 01:07 |
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/nodepool master: zk: use kazoo retry facilities https://review.openstack.org/535537 | 02:01 |
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/nodepool master: Implement a Runc driver https://review.openstack.org/535556 | 02:10 |
| *** yolanda_ has joined #zuul | 02:22 | |
| *** yolanda has quit IRC | 02:22 | |
| *** yolanda__ has joined #zuul | 04:08 | |
| *** yolanda__ has quit IRC | 04:10 | |
| *** yolanda__ has joined #zuul | 04:10 | |
| *** yolanda_ has quit IRC | 04:11 | |
| *** toabctl has joined #zuul | 05:17 | |
| tobiash | corvus: fix for #1 is already in review | 06:46 |
| tobiash | corvus: fix for #2 is my plan for next week as this hits us on pretty much every new pr due to the different working model with github | 06:48 |
| tobiash | corvus: fix for #1 is 574086 | 06:50 |
| tobiash | These useless tenant reconfigurations are becoming a serious scaling issue for us | 06:52 |
| tobiash | My idea about that is to cache the project branches | 06:53 |
| tobiash | And push events can then be checked against that | 06:56 |
| tobiash | Maybe that check can be done even before that caching | 06:57 |
| tobiash | The cost of asking github are probably smaller than the costs of a tenant reconfig | 06:58 |
| tobiash | Further a secend check that checks if we exclude all config from that repo could be useful | 06:59 |
| *** gtema has joined #zuul | 07:56 | |
| *** rbergeron has quit IRC | 10:13 | |
| *** rbergeron has joined #zuul | 10:14 | |
| *** gtema has quit IRC | 10:19 | |
| openstackgerrit | Merged openstack-infra/zuul-website master: Trivial: update url to new url https://review.openstack.org/573079 | 12:44 |
| corvus | tobiash: great! 574086 lgtm, approved. | 13:05 |
| corvus | tobiash: re #2, if there's any checking that we need to do, it should happen in the github driver before it sends the event to the scheduler. it's okay for it to take time to perform a query. then the event given to the scheduler should have all the information needed for the scheduler to make the decision needed. the main thing is we don't want to slow down the scheduler loop | 13:11 |
| openstackgerrit | Merged openstack-infra/zuul master: Improve logging around tenant reconfiguration https://review.openstack.org/574077 | 14:16 |
| openstackgerrit | Merged openstack-infra/zuul master: Also cache empty unparsed config https://review.openstack.org/574086 | 14:23 |
| tobiash | corvus: ok, so we may not react on a push event in the scheduler to enqueue a TenantReconfig into the scheduler loop but maybe let the driver to decide that and emit that event directly | 15:13 |
| corvus | tobiash: well, if you look in the scheduler where it checks branch_changed, it already checks to see if it touches config files | 15:17 |
| corvus | tobiash: so a push event that doesn't touch config files already doesn't trigger a reconfig -- i think the thing we probably want to add is also have it check to see if that branch is ignored because of exclude_unprotected_branches | 15:18 |
| corvus | tobiash: so the check for that should still happen in the scheduler. and we have the branch list cached on the tenant already. so that's probably just a simple change to the executor. | 15:20 |
| corvus | s/executor/scheduler/ | 15:20 |
| corvus | tobiash: if the branch create and delete events need extra info from a query though (like, does the newly created branch have branch protections), that query should happen in the github driver, and extra info added to the event. | 15:21 |
| corvus | tobiash: but in all cases, the event should still go to the scheduler (it might be used for something else) | 15:21 |
| tobiash | corvus: ah you want to piggy back the 'branch-ignores' information to the push event | 15:37 |
| tobiash | corvus: that's a probably more light weight change than adding a config-branch-changed event which I initially thought | 15:38 |
| tobiash | will try that next week then | 15:39 |
| *** sshnaidm|off has quit IRC | 16:47 | |
| *** sshnaidm|off has joined #zuul | 16:48 | |
| *** sshnaidm|off has quit IRC | 17:30 | |
| *** sshnaidm|off has joined #zuul | 17:34 | |
| openstackgerrit | Fabien Boucher proposed openstack-infra/zuul master: Add tenant yaml validation option to scheduler https://review.openstack.org/574265 | 17:37 |
| *** sshnaidm|off has quit IRC | 18:03 | |
| *** sshnaidm|off has joined #zuul | 18:09 | |
| *** snapiri- has joined #zuul | 21:05 | |
| *** snapiri has quit IRC | 21:08 | |
| *** snapiri has joined #zuul | 21:33 | |
| *** snapiri- has quit IRC | 21:34 | |
| *** tranzemc has quit IRC | 21:56 | |
| *** tranzemc has joined #zuul | 21:59 | |
| *** yolanda_ has joined #zuul | 22:50 | |
| *** yolanda__ has quit IRC | 22:53 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!