| *** soniya29 is now known as soniya29|ruck | 04:39 | |
| *** rlandy__ is now known as rlandy|out | 05:08 | |
| *** ysandeep|out is now known as ysandeep | 05:22 | |
| *** jpena|off is now known as jpena | 08:02 | |
| *** ysandeep is now known as ysandeep|lunch | 08:29 | |
| *** ysandeep|lunch is now known as ysandeep | 08:52 | |
| opendevreview | daniel.pawlik proposed openstack/ci-log-processing master: Skip processing log for build that result is NODE_FAILURE https://review.opendev.org/c/openstack/ci-log-processing/+/832205 | 09:32 |
|---|---|---|
| *** arxcruz is now known as arxcruz|off | 11:04 | |
| *** rlandy is now known as rlandy|ruck | 11:13 | |
| *** dviroel|out is now known as dviroel | 11:23 | |
| ykarel | frickler, can u check https://review.opendev.org/c/openstack/project-config/+/831378 | 11:46 |
| fzzf[m] | Hello , I have problem when using static providers. all job logs is https://paste.opendev.org/show/bg4PUATe5WMuJvRirjVk/... (full message at https://matrix.org/_matrix/media/r0/download/matrix.org/obAseXxwazJcvaYKnFyhujmY) | 12:56 |
| fzzf[m] | * Hello , I have problem when using static providers. all job logs is https://paste.opendev.org/show/bg4PUATe5WMuJvRirjVk/... (full message at https://matrix.org/_matrix/media/r0/download/matrix.org/XiqPHXeLFtLkePZbGOLGLRaZ) | 12:56 |
| fungi | fzzf[m]: looks like the devstack playbook expects /opt/cache/files to exist on the node. you might ask the devstack maintainers in #openstack-qa whether that's intentional, but the reason we don't get that error in opendev is that we pre-populate /opt/cache/files when creating the images for our job nodes | 13:00 |
| fzzf[m] | fungi: okay, thank you :) | 13:05 |
| fungi | fzzf[m]: see the openstack/project-config matches at https://codesearch.opendev.org/?q=/opt/cache/files for examples of what we put in there | 13:07 |
| frickler | fzzf[m]: fungi: yes, I don't think we expect the devstack jobs to be working outside of the opendev environment. I think we would accept patches to fix that if things don't get too complicated, but maybe you'll rather want to set up your own variant of those jobs | 13:27 |
| fungi | a lot of third-party ci systems to just reuse the upstream devstack jobs, but i expect most of them are also doing ephemeral nodes and reusing much of opendev's image building config | 13:28 |
| fungi | er, do just reuse | 13:28 |
| frickler | ykarel: thanks for the reminder, I had already opened that patch this morning, but got distracted | 13:30 |
| ykarel | Thanks frickler , can u also check https://review.opendev.org/c/zuul/zuul-jobs/+/762650 | 13:32 |
| frickler | fungi: ah, right, missed the point about static providers. but then my point holds even more, because the devstack jobs don't care to clean up after them at all and likely would fail miserably when told to run twice on the same node | 13:32 |
| fungi | yes, i agree | 13:32 |
| ykarel | revised an old patch, needed for ovs-source based multinode jobs | 13:32 |
| fungi | and also just the more general point that we can only gate devstack changes on whether their testing works in opendev | 13:32 |
| fungi | so third-party ci maintainers will have to help keep it working in their environments where necessary | 13:33 |
| fzzf[m] | fungi: frickler : fine. I get it. thanks. I'll try other provider. | 13:37 |
| opendevreview | Merged openstack/project-config master: Add Some periodic jobs to Neutron Dashboard https://review.opendev.org/c/openstack/project-config/+/831378 | 13:40 |
| frickler | ykarel: ah, I saw that patch and it looks scary to me, maybe some other reviewers are more adventurous than me | 13:53 |
| ykarel | frickler, any idea who can help, i can reach out | 13:54 |
| opendevreview | daniel.pawlik proposed openstack/ci-log-processing master: Skip processing log for build that result is NODE_FAILURE https://review.opendev.org/c/openstack/ci-log-processing/+/832205 | 14:14 |
| frickler | ykarel: that's part of the issue, the people that developed that setup are no longer available. maybe fungi or mnaser want to have a look | 14:24 |
| ykarel | frickler, Thanks, fungi mnaser can you please help out with https://review.opendev.org/c/zuul/zuul-jobs/+/762650 | 14:26 |
| fungi | ykarel: why not install ovn before setting up the bridge? this seems similar to the fips-mode jobs, where fips setup and accompanying reboot needed to happen before multi-node setup | 14:40 |
| ykarel | fungi, actually we build ovn/ovs from source in that job as part of devstack install, so just want to reuse the same | 14:44 |
| ykarel | me not sure about fips jobs, those jobs also build ovs from source? | 14:44 |
| fungi | no, they adjust some system settings and then reboot the vm with different kernel command line parameters | 14:45 |
| ykarel | hmm so that's different | 14:45 |
| fungi | but the end result is that the multi-node setup gets undone by the reboot if it's been applied before rebooting, so those jobs move multi-node setup to happen after the reboot instead | 14:45 |
| ykarel | hmm similar happening for our case but during ovn setup | 14:50 |
| ykarel | with 762650 we extending the role to generate scripts, those can be reused with ovn setup to reconfigure bridges | 14:51 |
| fungi | is it useful to have the multi-node setup applied twice though? | 14:51 |
| fungi | seems inefficient | 14:52 |
| ykarel | hmm not needed twice just needed once but at correct place | 14:55 |
| *** dviroel is now known as dviroel|lunch | 14:56 | |
| *** soniya is now known as soniya|ruck | 14:56 | |
| ykarel | ok Thanks fungi /me will check again if it can be done in another way | 14:59 |
| ykarel | current idea was to reuse as much as possible | 14:59 |
| fungi | ykarel: another way to look at this: do you actually want to test the building of ovn? if not, then having the ovn build happen in the run phase seems like a poor fit | 15:01 |
| ykarel | fungi, yes we want to build to test master version | 15:02 |
| fungi | ykarel: so just to be clear, whether or not ovn is buildable from source isn't tested by the ovn project, and is openstack's responsibility? | 15:02 |
| ykarel | fungi, ovn build must be tested by ovn project, we testing if ovn master do not break neutron | 15:03 |
| ykarel | in experimental pipeline | 15:03 |
| fungi | yes, but do you need to test that ovn is buildable, or just that the result of building ovn works with neutron? | 15:03 |
| ykarel | just the result | 15:04 |
| fungi | because there could be an ensure-ovn role which gets called in pre-run, before multinode-setup, and then the devstack bits for building ovn could just notice it's already there and no-op | 15:04 |
| ykarel | hmm need to check that which may need more changes, first i will try out calling multinode-setup b/w devstack setup and tempest run | 15:06 |
| *** ysandeep is now known as ysandeep|dinner | 15:36 | |
| *** ysandeep|dinner is now known as ysandeep | 15:55 | |
| *** dviroel|lunch is now known as dviroel | 16:13 | |
| *** ysandeep is now known as ysandeep|out | 17:18 | |
| *** jpena is now known as jpena|off | 17:52 | |
| *** dviroel is now known as dviroel|out | 22:25 | |
| opendevreview | Merged openstack/project-config master: Add zuul load balancer dashboard https://review.opendev.org/c/openstack/project-config/+/832149 | 23:15 |
| *** dasm is now known as dasm|off | 23:36 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!