| corvus | tonyb: i expect that i will propose a serious version of https://review.opendev.org/923084 (the change which used zuul to test out docker-compose-v2 with podman backend). i think i'd like zuul to end up using that, and nodepool too if it's still around. | 00:05 |
|---|---|---|
| corvus | tonyb: i don't mind if we merge a "--break" change that only affects the functional testing of nodepool/dib in the near term though. | 00:06 |
| corvus | one of our todos is to move that functional testing out of the nodepool repo into dib before we retire nodepool for good, so we'll have a chance to revisit. :) | 00:06 |
| corvus | (in other words, i have strong feelings we should never tell a user to do that, and don't care one way or the other if we just do that ourselves for functional testing for a short while) | 00:07 |
| corvus | oh, and i agree with clarkb that i don't think we should be doing that in general on opendev servers in system-config | 00:08 |
| tonyb | corvus: Okay thanks. That makes sense. | 00:10 |
| tonyb | I wonder if something like https://review.opendev.org/c/openstack/diskimage-builder/+/937836 would be better in openstack-base-jobs | 02:10 |
| Clark[m] | The zuul-jobs node information role was updated to cat /proc/cpuinfo contents. Maybe something like that which distills it into human readable form makes sense there? | 02:28 |
| fungi | for the "dead batteries" removed from stdlib in 3.13, most of them have at least been copied to pypi and are separately installable still | 03:35 |
| fungi | ubuntu-ports mirror update vos release completed, so i've started the script again to make sure it still runs clean, then i'll drop the flock | 03:38 |
| zigo | Clark[m]: frickler: fungi: Yes, we still use puppet, but we're in the process of moving to helm. Though this will probably take us a year or so to do the move. Anyways, we use the Debian packaged version, so we'll continue to do that, just Debian may use the fork (note: I'm part of the team that does the puppetserver packaging). | 08:22 |
| zigo | fungi: I wrote *many* patches to get rid of the dead-battery thing, and we also packaged a few "-legacy" packages in Debian to address the issue. For me, that's already an issue in the past ... :) | 08:23 |
| frickler | infra-root: fixed jammy kernel seems to be available in -updates pocket, I've unpause the build and will watch results once it succeeds | 08:33 |
| frickler | I also noticed nb04 has a full disk, not sure if the ports mirror is enough done already that it makes sense to clean that up, but should at least be a todo item later | 08:34 |
| opendevreview | Rafal Lewandowski proposed openstack/diskimage-builder master: Prevent from overwriting grub defaults if no variables are set https://review.opendev.org/c/openstack/diskimage-builder/+/937684 | 08:42 |
| tonyb | frickler: thanks for reestablishing the jammy builds | 08:44 |
| opendevreview | Rafal Lewandowski proposed openstack/diskimage-builder master: Set grub default values according to its project defaults https://review.opendev.org/c/openstack/diskimage-builder/+/937684 | 08:47 |
| opendevreview | Rafal Lewandowski proposed openstack/diskimage-builder master: Change grub variables for style and timeout https://review.opendev.org/c/openstack/diskimage-builder/+/937684 | 08:55 |
| frickler | build finished 15min ago, now waiting for uploads and then to check some results | 10:34 |
| tonyb | frickler: thanks | 10:36 |
| frickler | confirmed in the build log that linux-image-5.15.0-130-generic is being installed, so I don't expect any issues really | 10:37 |
| frickler | infra-root: I've cleaned up the remaining exim paniclogs to reduce mail spam, but noticed that I do not seem to be able to access keycloak01.opendev.org, maybe someone can check and just remove /var/log/exim4/paniclog there? | 10:47 |
| fungi | second ubuntu-ports mirror run finished about 8 hours ago, presumably because of how long the initial run took. i'll fire it again now since there should be a lot fewer additional packages to sync | 14:35 |
| fungi | frickler: i had to use an older ssh key to get into keycloak01 (we kept it around in case someone has time to copy the openinfraid client settings), but removed the exim paniclog just now | 14:39 |
| frickler | fungi: oh, good idea, didn't occur to me to try that earlier, but using my old key worked for me now, too, thx | 15:10 |
| clarkb | frickler: I assume new jammy is working happily if there haven't been complaints yet | 15:21 |
| frickler | clarkb: I think so, I didn't check systematically, but didn't see any major failures | 16:13 |
| clarkb | it has been long enough I would've expected at least some complaints if it was still broken | 16:15 |
| clarkb | also people did test the kernel before they released it according to the issue tracker so seems like it is good | 16:15 |
| fungi | ubuntu-ports third sync finished 20 minutes ago. started a fourth to see if it's essentially a no-op now | 16:19 |
| clarkb | I will start cleaning up /opt on nb04 in a screen there | 16:22 |
| clarkb | I stopped the buildercontainer then started a deletion of the dib_build and dib_image dirs in /opt/dib_tmp | 16:25 |
| clarkb | I believe the hourly nodepool job will restart the container but that should be fine. The deletions might take some time but also again fine to run in screen for a while. losetup -a reports 8 devices which should be under the limit too | 16:26 |
| clarkb | fungi: frickler we should probably clean up the old keycloak server? | 16:30 |
| clarkb | thats probably an after holidays task though | 16:30 |
| opendevreview | Matthieu Huin proposed zuul/zuul-jobs master: Fix the upload-logs-s3 test playbook https://review.opendev.org/c/zuul/zuul-jobs/+/927600 | 16:31 |
| opendevreview | Matthieu Huin proposed zuul/zuul-jobs master: Update the set-zuul-log-path-fact scheme to prevent huge url https://review.opendev.org/c/zuul/zuul-jobs/+/927582 | 16:31 |
| corvus | clarkb: would be good to save the configuration data for reference. i think there's a working openstackid app configured in there. but i'm sure it can be re-done. | 16:59 |
| corvus | (i think that's all encoded in the database) | 16:59 |
| clarkb | ack | 17:04 |
| clarkb | corvus: do you think the best way to do that is a db dump or is it more friendly to capture the web config somehow (so that it can be more easily reapplied that way) | 17:04 |
| corvus | clarkb: would be more friendly to do some web config, but i think it would still be recoverable for reference purposes if we just save a copy of the db, or a db dump. | 17:10 |
| fungi | last ubuntu-ports update completed in about 70 minutes, roughly 60 of which was the `reprepro update` step | 17:52 |
| fungi | i'll go ahead and drop the flock and consider this back to some semblance of normal | 17:52 |
| clarkb | fungi: thanks! | 18:08 |
| clarkb | checking on the gerrit cache h2 files things seem to be reasonably stable. I'm hopeful that the daily pruning of the databases rather than infrequent prune when we happen to restart will keep those files down to a reasonable size even without regular compaction | 18:36 |
| opendevreview | Clark Boylan proposed opendev/system-config master: Update gitea to v1.22.6 https://review.opendev.org/c/opendev/system-config/+/938144 | 18:59 |
| *** dhill is now known as Guest3662 | 21:51 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!