| @jim:acmegating.com | i think there's still a race in the waituntilsettled method in tests based on the output of https://5747b7c5054b9c0b2cf1-065293745cdcc6ecb6052ade0199e5ba.ssl.cf5.rackcdn.com/807406/6/gate/zuul-tox-py36/fe52c23/testr_results.html | 02:28 |
|---|---|---|
| @jim:acmegating.com | i'm not worried about that being a production issue; we should address it, but i think we can keep merging; i'll recheck it | 02:29 |
| -@gerrit:opendev.org- Zuul merged on behalf of James E. Blair https://matrix.to/#/@jim:acmegating.com: [zuul/zuul] Report nodepool resource stats gauges in scheduler https://review.opendev.org/c/zuul/zuul/+/807406 | 07:34 | |
| -@gerrit:opendev.org- Zuul merged on behalf of James E. Blair https://matrix.to/#/@jim:acmegating.com: [zuul/zuul] Add ZK session-aware elections https://review.opendev.org/c/zuul/zuul/+/807656 | 16:18 | |
| -@gerrit:opendev.org- Zuul merged on behalf of Felix Edel: [zuul/zuul] Don't use the AnsibleJob in the nodepool client https://review.opendev.org/c/zuul/zuul/+/807711 | 16:25 | |
| -@gerrit:opendev.org- Zuul merged on behalf of Simon Westphahl: [zuul/zuul] Make sure dynamic change queues are marked as such https://review.opendev.org/c/zuul/zuul/+/807121 | 17:58 | |
| @pabelanger:matrix.org | I was thinking more about an issue we are seeing with quay.io and timeouts. It got me thinking, if a buildset is running a dedicated buildset_registry, could we not proxy those quay.io images via the buildset registry? | 18:03 |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: [zuul/zuul] Fix test race with node allocation https://review.opendev.org/c/zuul/zuul/+/808551 | 18:07 | |
| @jim:acmegating.com | pabelanger: buildset registry isn't a proxy; it would be very complicated to make it a proxy. easier to wrap existing usage with retries. | 18:07 |
| @jim:acmegating.com | Clark, tobiash: 808551 should resolve the remaining test race we observed. it's a familiar one. ;) | 18:08 |
| @pabelanger:matrix.org | Yah, that is step one. because I can't find a way to retry container downloads when podman build is called, I'll podman pull in a pre-run playbook. But that got me thinking of our buildset registry | 18:09 |
| @jim:acmegating.com | pabelanger: yeah, doing the pull at the start is good. you could pull and then push to the buildset registry; that would be advantageous if you have more than one job using the buildset registry (and obvs using those same images). if it's only one job then there isn't an advantage above just pulling. | 18:10 |
| @pabelanger:matrix.org | yah, that is my thoughts too | 18:10 |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: [zuul/zuul] Handle de-serializing NodeRequests without requestor_data https://review.opendev.org/c/zuul/zuul/+/808553 | 18:44 | |
| @jim:acmegating.com | zuul-maint: ^ i restarted opendev with master, and observed exceptions in the logs related to that ^. it's not critical, but it's noisy logspam that we should correct soon. | 18:45 |
| -@gerrit:opendev.org- Paul Belanger proposed: [zuul/zuul-jobs] Expose build_opts for container_images https://review.opendev.org/c/zuul/zuul-jobs/+/808554 | 19:03 | |
| -@gerrit:opendev.org- Paul Belanger proposed: [zuul/zuul-jobs] Expose build_opts for container_images https://review.opendev.org/c/zuul/zuul-jobs/+/808554 | 19:15 | |
| -@gerrit:opendev.org- Paul Belanger proposed: [zuul/zuul-jobs] Expose build_opts for container_images https://review.opendev.org/c/zuul/zuul-jobs/+/808554 | 19:18 | |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: [zuul/zuul] Handle de-serializing NodeRequests without requestor_data https://review.opendev.org/c/zuul/zuul/+/808553 | 20:12 | |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: [zuul/nodepool] WIP static node slots https://review.opendev.org/c/zuul/nodepool/+/807806 | 21:54 | |
| @clarkb:matrix.org | corvus: I left a question on 808553 but +2'd as either way it should be an improvement over the current situation. The parent change lgtm too | 23:20 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!