15:00:41 <priteau> #startmeeting blazar
15:00:41 <opendevmeet> Meeting started Thu May  2 15:00:41 2024 UTC and is due to finish in 60 minutes.  The chair is priteau. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:41 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:00:41 <opendevmeet> The meeting name has been set to 'blazar'
15:00:45 <priteau> #topic Roll call
15:00:48 <priteau> o/
15:00:48 <mattcrees> o/
15:01:19 <mppowers> o/
15:02:41 <priteau> #topic Dalmatian-1 milestone
15:03:12 <priteau> We are working towards the Dalmatian-1 milestone
15:03:29 <priteau> This is scheduled for May 16: https://releases.openstack.org/dalmatian/schedule.html
15:04:31 <priteau> We have the following patches in progress
15:05:02 <priteau> 907907: Disabled field in computehost | https://review.opendev.org/c/openstack/blazar/+/907907
15:05:30 <priteau> There were comments provided by Anish since the last round of reviews
15:06:12 <priteau> mppowers: Anish is unlikely to work on these any further, right?
15:06:35 <mppowers> Correct
15:07:47 <priteau> We will do another round of review, I hope we can resolve all the remaining issues soon.
15:08:09 <priteau> Next we have a series of patches from johnthetubaguy to start integrating flavor-based reservation: https://review.opendev.org/q/topic:%22flavor_plugin%22
15:08:32 <priteau> I started reviewing but I imagine they will require more iterations
15:08:45 <priteau> We will let you know when they are ready for more general review.
15:08:57 <priteau> Of course if you are curious feel free to take a look already :)
15:10:14 <priteau> Any other patches to focus on at the moment?
15:11:19 <mattcrees> There's the various unmaintained/zed patches raised, are we ready to merge these yet?
15:12:27 <priteau> I guess that's a separate topic
15:12:33 <priteau> #topic Stable branches
15:13:00 <priteau> This is the list of the currently open patches on stable branches: https://review.opendev.org/q/project:openstack/barbican+-branch:master+status:open
15:13:13 <priteau> Sorry, wrong link
15:13:26 <priteau> https://review.opendev.org/q/project:openstack/blazar+-branch:master+status:open
15:13:37 <priteau> https://review.opendev.org/q/project:openstack/python-blazarclient+-branch:master+status:open
15:13:44 <priteau> https://review.opendev.org/q/project:openstack/blazar-nova+-branch:master+status:open
15:14:32 <priteau> https://review.opendev.org/q/project:openstack/blazar-dashboard+-branch:master+status:open
15:15:26 <priteau> I have noticed that they all failed to pass Zuul checks
15:15:33 <priteau> But I haven't investigated why
15:16:22 <priteau> The zed branch would be good to fix. However, do we want to spend the effort of fixing branches older than yoga?
15:16:43 <priteau> We have the option to retire them early
15:16:55 <priteau> mppowers: Would you be affected if we EOL xena?
15:17:46 <mppowers> No, I think that would be fine. We are working on upgrading to Antelope soon, and we can backport something ourselves if needed.
15:18:10 <priteau> Can I propose to EOL xena, wallaby and victoria?
15:18:37 <priteau> Sorry, be back in a few minutes
15:19:07 <mppowers> Yes, I agree with that proposal
15:19:36 <priteau> Back
15:19:59 <priteau> We can try to keep yoga and zed alive if possible, but if it's too much effort we would have to EOL them too
15:21:34 <mattcrees> Sounds good
15:22:58 <priteau> The blazar zed devstack failure doesn't seem related to us
15:23:09 <priteau> ImportError: cannot import name 'security_groups_rules_belongs_to_default_sg' from 'neutron_lib.api.definitions' (/usr/local/lib/python3.8/dist-packages/neutron_lib/api/definitions/__init__.py)
15:25:58 <priteau> Maybe we just need to wait a few days
15:26:25 <priteau> #topic AOB
15:26:30 <priteau> Anything else to share today?
15:27:05 <mattcrees> Nothing else from me
15:28:06 <mppowers> We had a bit of an interesting bug that came up, where in blazar-nova aggregates were showing as out of date. It turned out this was not because of ironic, but we added threading to the host on start event, which triggered a race condition in nova.
15:28:41 <mppowers> This was the commit, which was just in our fork for now https://github.com/ChameleonCloud/blazar/commit/a855147cbbcd9c1b9ae0eafa3a7ff83692d45594
15:28:45 <priteau> Right, I have been following this thread :)
15:30:28 <mppowers> We have a workaround for now
15:31:18 <priteau> Well done to Michael for tracking this down
15:31:55 <mppowers> Yes, race conditions are the worst
15:31:59 <priteau> Is the workaround your request filter?
15:32:55 <mppowers> Right now, we have blazar-nova filtering code moved to "prefilter" stage in nova https://github.com/ChameleonCloud/nova/commit/7f9058bb7000fbd9f3d226009f388cc7ce315902
15:34:05 <mppowers> but this doesn't include support for everything blazar-nova does, only filtering based on passed in scheduler hint
15:34:42 <priteau> We discussed request filters with the Nova team a long time ago. I would be interested in seeing if we do without the blazar-nova scheduler filter thanks to them.
15:36:45 <mppowers> Yes, it seems nova has a very similar pre-filter based on aggregates, but we opted to copy it into our own for extra debugging + clear error messages
15:39:50 <priteau> Thanks for bringing this up. So we need some solution upstream if we ever want to parallelise aggregate ops
15:40:49 <mppowers> I believe this is the bug https://bugs.launchpad.net/nova/+bug/1542491
15:42:22 <priteau> Any immediate action to take upstream regarding this bug?
15:42:34 <priteau> Except for Nova to investigate how to fix it of course
15:43:17 <mppowers> I don't think so, we can monitor the situation as you say, if we ever want to parallelize this in blazar
15:43:53 <priteau> Anything else?
15:44:20 <mppowers> nothing else from me
15:44:39 <priteau> Thanks all for joining!
15:44:48 <priteau> Good bye
15:44:50 <priteau> #endmeeting