Monday, 2026-01-26

*** mhen_ is now known as mhen02:41
opendevreviewIvan Anfimov proposed openstack/freezer-specs master: Remove outdated comment in setup.py  https://review.opendev.org/c/openstack/freezer-specs/+/97468309:26
opendevreviewIvan Anfimov proposed openstack/freezer-specs master: Remove outdated comment in setup.py  https://review.opendev.org/c/openstack/freezer-specs/+/97468309:27
opendevreviewMerged openstack/freezer master: docs: fix few mistakes  https://review.opendev.org/c/openstack/freezer/+/97439911:15
opendevreviewMerged openstack/freezer-api master: Remove LegacyEngineFacade  https://review.opendev.org/c/openstack/freezer-api/+/97444713:36
opendevreviewDmitriy Rabotyagov proposed openstack/freezer-api master: Ensure all_projects are defined for get_search_query  https://review.opendev.org/c/openstack/freezer-api/+/97388614:07
opendevreviewMerged openstack/freezer-api master: Use SSL and authentication for elasticsearch  https://review.opendev.org/c/openstack/freezer-api/+/97441114:50
noonedeadpunk#startmeeting freezer15:00
opendevmeetMeeting started Mon Jan 26 15:00:18 2026 UTC and is due to finish in 60 minutes.  The chair is noonedeadpunk. Information about MeetBot at http://wiki.debian.org/MeetBot.15:00
opendevmeetUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.15:00
opendevmeetThe meeting name has been set to 'freezer'15:00
noonedeadpunk#topic rollcall15:00
noonedeadpunko/15:00
noonedeadpunkhey there!15:00
mheno/15:00
noonedeadpunkwe have quite an agenda today btw15:01
noonedeadpunk#topic CI status15:03
noonedeadpunkSo, CI for the project should be restored now, with https://review.opendev.org/c/openstack/freezer-api/+/97444715:03
mhengood job!15:04
noonedeadpunkIt was caused by oslo.db finally getting rid of legacy engine15:04
noonedeadpunkSo now I'm re-checking things and will continue reviewing existing patches15:04
noonedeadpunkWith that I kinda got stuck a little bit with elasticsearch engine on devstack15:05
mhenmy patches pass tests again, yay \o/15:05
noonedeadpunk(I hope I did not break anything more then it's broken with this change)15:05
noonedeadpunk#topic Elasticsearch DB15:05
noonedeadpunkI have spent quite some time into elasticsearch in devstack, once it's version was bumped15:06
noonedeadpunkwhile devstack does actually deploy now, it's completely borked15:06
noonedeadpunkRight now U-C has elactisearch library on 2.4.1: https://opendev.org/openstack/requirements/src/branch/master/upper-constraints.txt#L33215:07
noonedeadpunkWith ELK 8/9 we need to update it to 9 basically, according to https://github.com/elastic/elasticsearch-py/tree/v9.2.1?tab=readme-ov-file#compatibility15:07
noonedeadpunkBut then there are plenty of huge changes, so source should be refactored pretty much significantly as well15:08
noonedeadpunkNot impossible to do15:08
noonedeadpunkBUT. I wonder if we really wanna have that option...15:08
mhenwhat is the benefit for Freezer to use ES?15:08
noonedeadpunkI checked through the codesearch, and except kolla/helm (which have it for infra reasons), only freezer and venus having it as engine15:10
noonedeadpunkI think intention is to have faster alternative to relational DB15:11
mhenbut is Freezer really under that heavy load with managing the clients and jobs to justify that?15:11
mhenI think some other OpenStack core components are much busier15:12
noonedeadpunkthat is really a great question15:12
noonedeadpunkI'd say - it not15:13
mhenmy suggestion would be to remove it as it's broken anyway15:13
noonedeadpunkI am inclining to that15:14
noonedeadpunkI think we should have some deprecation period, but right after 2026.1 we should be able to drop it15:14
mhenif the need arises because somebody identifies a bottleneck, it can be addressed by (re)introducing an alternative backend when the time comes15:14
noonedeadpunkand now we can just switch default devstack logic to use mariadb15:14
mhen+115:14
noonedeadpunkI think the most likely issue would be to drop historical records from DB15:15
noonedeadpunkbut indeed - let's come to bridge before crossing itr15:15
noonedeadpunk#agreed deprecate elasticsearch engine and switch devstack defaults to use sqlalchemy15:16
noonedeadpunk#topic horizon plugin state15:16
noonedeadpunkI _think_ I have mostly fixed the horizon plugin15:17
noonedeadpunkwith that I also was reviewing the patch when realized it's probably better to go slightly further15:17
noonedeadpunk#link https://review.opendev.org/c/openstack/freezer-web-ui/+/95858415:17
noonedeadpunkand under further I mean getting rid of own dashboard and top-level menu, with getting group under Projects'15:18
noonedeadpunkSorry, previous link was wrong, I meant https://review.opendev.org/c/openstack/freezer-web-ui/+/96131415:18
noonedeadpunkSo I came up with the following one15:19
noonedeadpunk#link https://review.opendev.org/c/openstack/freezer-web-ui/+/97394915:19
noonedeadpunkIt would be really nice to have some external "verification" of it15:19
mhenwhat do you mean with "external verification"?15:20
noonedeadpunkif it makes sense to somebody except of me, and if works :D15:21
mhenI could test it on my devstack and have a look at it if it helps15:21
mhenhowever, not before next week15:21
* mhen is in a different context this week and cannot work on Freezer until next week15:21
noonedeadpunkok, I see15:22
noonedeadpunkI think there's no rush with this change15:22
noonedeadpunkand I'll also try to ask somebody from my org to test it in the meanwhile15:22
noonedeadpunk#topic container names passed to cinder-backup with invalid separators15:23
noonedeadpunkWe've discussed that during previous week, and I see you've proposed a patch15:23
noonedeadpunk#link https://review.opendev.org/c/openstack/freezer/+/973627/12/freezer/openstack/backup.py#b9115:23
noonedeadpunkI have not reviewed it yet, so will post an update in the patch, if it's fine with you15:24
mhensure15:24
mhenit was a minor change so I added it to my Cinder fixup patchset15:24
noonedeadpunk++ sure15:25
noonedeadpunk#topic restore-from-date behavior is different from documented one15:25
noonedeadpunkthere's also a related bug report to this topic15:25
noonedeadpunk#link https://bugs.launchpad.net/freezer/+bug/213888315:25
mhenyea I thought keeping it as a bug is better to document the observed behavior15:26
noonedeadpunkyeah, I think I also saw some patch related to this topic?15:26
mhenyou did?15:27
noonedeadpunknah, maybe I am thinking about smth completely different now15:27
noonedeadpunkI can take this as an action point during this week15:27
noonedeadpunk#action noonedeadpunk to triage the bug report and propose a fix15:28
mhenwe need to decide whether we want to have it behave like it says or if we simply adjust the documentation to reflect its actual behavior15:28
noonedeadpunkit could be also that we might want to rename the option to be more explicit about what it's doing...15:29
mhenfrom a user perspective I would lean towards the current documented behavior making a bit more sense and that the implementation is wrong15:30
mhenre: rename - yes that would also help15:30
mhenbecause currently it suggests that it will restore that exact timestamp down to the second when in reality it actually has a fuzziness to it15:31
mhenmaybe we can discuss in the bug15:31
noonedeadpunkyes, sounds good15:31
mhenthat would then also serve as some kind of decision record15:31
noonedeadpunkI am not sure so far what is most reasonable/expected behavior would be15:32
* noonedeadpunk needs to sleep over with this15:32
noonedeadpunk#topic consistent OpenStack resource naming15:33
noonedeadpunkthere's a patch which suggest to align naming of created by Freezer resources15:34
noonedeadpunk#link https://bugs.launchpad.net/freezer/+bug/213888315:34
mhen#link https://review.opendev.org/c/openstack/freezer/+/974483/1/releasenotes/notes/consistent-openstack-resource-naming-fa8f3d72c24faefa.yaml15:34
mhenthe release notes explain in detail how the change works15:34
noonedeadpunkoops I sent a wrong link again15:34
noonedeadpunk`~` - that is really a smart one15:35
mhenRFC 3986 said its safe15:35
mhenand it is less used by users and other components15:35
noonedeadpunkyeah, it must be. It did not come to my mind15:35
noonedeadpunkyup, really like it15:36
mhenI found the resulting user experience to be tremendously better when observing all the resources during backup/restore15:36
noonedeadpunkthere seems to be some related pep8 issues15:37
noonedeadpunk#link https://zuul.opendev.org/t/openstack/build/ec4b099489ab48fda74b87eb8dc91ffa15:37
mhenI know15:37
mhenwill address them15:37
noonedeadpunkotherwise it looks quite fair to me now15:38
noonedeadpunkthis indeed should improve quality of life for users15:38
opendevreviewMerged openstack/freezer master: Fix issues in Python code  https://review.opendev.org/c/openstack/freezer/+/97362015:38
noonedeadpunkok, I guess this should be pretty much it from my side at least15:39
noonedeadpunk#topic open discussion15:39
mhenon another note I noticed that the backup timestamp of a cinder volume backup (via glance image) communicated by freezer-agent's output mismatches the resulting timestamp on disk in the backup directory15:39
noonedeadpunkit's off by couple of seconds?15:39
mhen#link https://github.com/openstack/freezer/blob/master/freezer/openstack/backup.py#L6615:39
noonedeadpunkyeah, I think we should be taking timestamp from the resource....15:40
mhenyes, because of this part in the code where a new timestamp is generated after the initial one from the job's start15:40
noonedeadpunkwhich is really confusing, agree15:41
mhenjust a minor thing I noticed while I tested my timestamp appendix patch15:41
noonedeadpunkcan take a look during the week15:41
mhengreat15:41
mhenas stated before, I sadly cannot contribute this week15:42
noonedeadpunkthis is fine, there're quite some outstanding patches from you to review/merge15:42
noonedeadpunkok, cool then, thanks for joining!15:44
noonedeadpunk#endmeeting15:45
opendevmeetMeeting ended Mon Jan 26 15:45:40 2026 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)15:45
opendevmeetMinutes:        https://meetings.opendev.org/meetings/freezer/2026/freezer.2026-01-26-15.00.html15:45
opendevmeetMinutes (text): https://meetings.opendev.org/meetings/freezer/2026/freezer.2026-01-26-15.00.txt15:45
opendevmeetLog:            https://meetings.opendev.org/meetings/freezer/2026/freezer.2026-01-26-15.00.log.html15:45
opendevreviewMerged openstack/freezer-api master: Ensure all_projects are defined for get_search_query  https://review.opendev.org/c/openstack/freezer-api/+/97388615:46
opendevreviewMerged openstack/freezer master: Fix Glance integration  https://review.opendev.org/c/openstack/freezer/+/97362316:45
opendevreviewMarkus Hentsch proposed openstack/freezer master: Fix Nova integration and change clients to SDK  https://review.opendev.org/c/openstack/freezer/+/97362516:58
opendevreviewDmitriy Rabotyagov proposed openstack/freezer-api master: Remove redundant wsgi file  https://review.opendev.org/c/openstack/freezer-api/+/97475418:03
opendevreviewMerged openstack/freezer master: Fix Nova integration and change clients to SDK  https://review.opendev.org/c/openstack/freezer/+/97362518:11
opendevreviewDmitriy Rabotyagov proposed openstack/freezer-api master: Update contributor guide  https://review.opendev.org/c/openstack/freezer-api/+/97475518:12
opendevreviewMerged openstack/freezer master: Remove installation guide for openSUSE/SLES  https://review.opendev.org/c/openstack/freezer/+/97465918:19
opendevreviewDmitriy Rabotyagov proposed openstack/freezer master: Add [os_brick]/lock_path config option  https://review.opendev.org/c/openstack/freezer/+/97477220:07
opendevreviewMerged openstack/freezer master: Replace cinderclient with SDK, fix cinder engines  https://review.opendev.org/c/openstack/freezer/+/97362720:49

Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!