| jeblair | Shrews, clarkb: i think 424863 is good enough (logs are in an extra subdirectory, but at least we have them) | 00:07 |
|---|---|---|
| openstackgerrit | Merged openstack-infra/zuul: Rename nodepool integration script https://review.openstack.org/424863 | 00:11 |
| *** ianw is now known as ianw_pto | 04:34 | |
| *** Guest58531 has quit IRC | 05:04 | |
| *** mgagne has joined #zuul | 05:07 | |
| *** mgagne is now known as Guest33539 | 05:07 | |
| *** abregman has joined #zuul | 05:22 | |
| *** saneax has joined #zuul | 06:34 | |
| *** saneax is now known as saneax-_-|AFK | 06:46 | |
| *** isaacb has joined #zuul | 07:33 | |
| *** hashar has joined #zuul | 08:03 | |
| *** abregman is now known as abregman|mtg | 08:27 | |
| *** abregman|mtg is now known as abregman | 09:04 | |
| *** isaacb has quit IRC | 10:01 | |
| *** isaacb has joined #zuul | 10:03 | |
| *** abregman is now known as abregman|afk | 10:19 | |
| *** yolanda has quit IRC | 10:23 | |
| *** abregman|afk is now known as abregman | 10:25 | |
| *** abregman is now known as abregman|mtg | 10:25 | |
| *** yolanda has joined #zuul | 10:26 | |
| *** abregman|mtg is now known as abregman | 10:40 | |
| *** isaacb has quit IRC | 10:52 | |
| *** isaacb has joined #zuul | 10:53 | |
| *** isaacb has quit IRC | 11:03 | |
| *** openstackgerrit has quit IRC | 11:17 | |
| *** Zara has quit IRC | 11:30 | |
| *** Zara has joined #zuul | 11:35 | |
| *** isaacb has joined #zuul | 11:51 | |
| *** bhavik1 has joined #zuul | 12:15 | |
| *** bhavik1 has quit IRC | 12:53 | |
| *** openstackgerrit has joined #zuul | 13:23 | |
| openstackgerrit | David Shrewsbury proposed openstack-infra/zuul: Add zuul->np test for invalid node type https://review.openstack.org/425686 | 13:23 |
| openstackgerrit | David Shrewsbury proposed openstack-infra/zuul: Add zuul->np test for invalid node type https://review.openstack.org/425686 | 13:29 |
| Shrews | jeblair: w00t! our first integration test pass: http://logs.openstack.org/86/425686/2/experimental/gate-zuul-nodepool/a83475e/testr_results.html.gz | 13:39 |
| mordred | Shrews: that's exciting! | 13:40 |
| Shrews | mordred: i'm so excited | 13:40 |
| Shrews | and i just can't hide it | 13:41 |
| *** isaacb has quit IRC | 13:43 | |
| pabelanger | that is awesome | 13:44 |
| dmsimard | Integration tests !? | 13:52 |
| dmsimard | Are those for Nodepool or Zuul ? or both ? Sort of hard to tell from the perspective of a noob. | 13:53 |
| Shrews | dmsimard: both. it tests the v3 zuul->nodepool communication/interaction | 13:54 |
| dmsimard | Neat. The lack of integration tests has always scared me away from really trying to contribute anything meaningful to Zuul :) | 13:54 |
| dmsimard | Sending things straight to production is not something I usually do but maybe I'm old school :p | 13:55 |
| *** isaacb has joined #zuul | 13:57 | |
| mordred | dmsimard: yah - I think we're all going to be much more pleased with the integration tests being there | 13:59 |
| *** saneax-_-|AFK is now known as saneax | 14:11 | |
| *** hashar is now known as hasharAway | 15:15 | |
| jeblair | dmsimard: zuul's tests are actually fairly reliable. nearly all of the tests are functional with all of the systems it interacts with comprehensively faked. so if something is tested in zuul, i'm confident about sending it into production. | 15:37 |
| dmsimard | jeblair: yeah but for example, in this old patch (that I'll revive one day) https://review.openstack.org/#/c/330874/ I was hoping to see the results of that patch in the gate jobs | 15:38 |
| jeblair | dmsimard: this zuul-nodepool integration test is very small, just covering the interaction between the two systems. and it's mainly useful because neither system *exists* yet, so they necessarily need to evolve together. even so, zuul has nodepool fakes still, and nodepool has a zuul fake, and both will continue to test things in their unit/functional tests. | 15:38 |
| jeblair | dmsimard: oh, that's the v2.5 ansible launcher. it's an exception to what i've said because it was a dead-end bit of code. | 15:39 |
| jeblair | dmsimard: the v3 equivalent of that code has tests now, and even runs ansible and can now examine the results of ansible to verify correctness | 15:39 |
| dmsimard | cool, I'll have to check it out | 15:39 |
| jeblair | dmsimard: so if you proposed that to v3, even now, it would be fully testable | 15:40 |
| mordred | \o/ for fully testable | 15:45 |
| *** openstackgerrit has quit IRC | 15:48 | |
| *** abregman has quit IRC | 16:19 | |
| *** hasharAway has quit IRC | 16:20 | |
| *** saneax is now known as saneax-_-|AFK | 16:52 | |
| *** hashar has joined #zuul | 16:57 | |
| *** isaacb has quit IRC | 16:59 | |
| *** saneax-_-|AFK is now known as saneax | 17:00 | |
| *** persia has quit IRC | 17:16 | |
| *** openstackgerrit has joined #zuul | 17:18 | |
| openstackgerrit | James E. Blair proposed openstack-infra/nodepool: Merge branch 'master' into feature/zuulv3 https://review.openstack.org/424974 | 17:18 |
| *** persia has joined #zuul | 17:20 | |
| *** hashar is now known as hasharAway | 17:30 | |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul: Handle multiple playbook extensions https://review.openstack.org/425450 | 17:56 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul: Add playbooks for all test jobs https://review.openstack.org/425810 | 17:56 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul: Add playbooks for all test jobs https://review.openstack.org/425810 | 18:16 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul: Handle multiple playbook extensions https://review.openstack.org/425450 | 18:16 |
| Shrews | jeblair: for a single node request that wants multiple nodes, do we want to spin up separate threads for each node that needs launched to make fulfillment quicker? | 18:58 |
| jeblair | Shrews: i think so. | 18:59 |
| jeblair | (otherwise, things could be very slow) | 18:59 |
| jeblair | Shrews: and some of those threads will fail. so probably the main worker thread will need to keep track of those, check in on them, and see if the request is done, or if threads need to be re-launched... | 19:00 |
| *** harlowja has quit IRC | 19:02 | |
| Shrews | yeah | 19:02 |
| *** yolanda has quit IRC | 19:17 | |
| Shrews | jeblair: any limit on the number of re-launches we try? | 19:18 |
| Shrews | i'd think we'd want to just give up after a while if there are provider issues | 19:18 |
| Shrews | i don't see that the current code does that. one and done | 19:20 |
| clarkb | pre zk nodepool will continue to ask cloud for resources until it gets them, but it doesn't allocate resources to specific requests so that is a bit easier to sort out | 19:22 |
| Shrews | clarkb: hmm. i'd think in the new world, we'd want to give up and decline the request so that it might be fulfilled by another provider | 19:24 |
| Shrews | (assuming another provider has the requested images) | 19:24 |
| clarkb | ya I think with things bucketed into specific requests it would be good to be a little smarter about it and give up and let others attempt it | 19:25 |
| Shrews | i suppose i could make the retry configurable, and support "never give up, never give in" mode | 19:26 |
| *** saneax is now known as saneax-_-|AFK | 19:31 | |
| *** yolanda has joined #zuul | 19:34 | |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul: Use playbooks defined in repos https://review.openstack.org/423550 | 19:41 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul: Put Zuul vars in an ansible vars file https://review.openstack.org/424367 | 19:41 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul: Add playbooks for all test jobs https://review.openstack.org/425810 | 19:41 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul: Handle multiple playbook extensions https://review.openstack.org/425450 | 19:41 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul: Fix nodeset return on skipped jobs https://review.openstack.org/425865 | 19:41 |
| jeblair | goonies never say die. but nodepool does? | 19:42 |
| jeblair | Shrews: yeah, i think configurable makes sense for this. 3x default maybe? | 19:42 |
| Shrews | sounds reasonable | 19:43 |
| jeblair | those patches ^ are a restack on top of a fix (425865) which will hopefully help with our recent test instability | 19:43 |
| *** Guest33539 is now known as mgagne | 20:01 | |
| *** mgagne has quit IRC | 20:01 | |
| *** mgagne has joined #zuul | 20:01 | |
| *** harlowja has joined #zuul | 20:09 | |
| *** hasharAway is now known as hashar | 20:10 | |
| mordred | jeblair, Shrews: ++ 3x retry with configurable | 20:27 |
| *** saneax-_-|AFK is now known as saneax | 23:02 | |
| *** hashar has quit IRC | 23:03 | |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul: Move job running into own class https://review.openstack.org/385964 | 23:25 |
| jeblair | jhesketh: ^ i dusted off that change because i'm looking at implementing the pre/post playbook stuff and your commit message is correct. :) | 23:26 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!