Monday, 2026-02-02

opendevreviewDmitriy Rabotyagov proposed openstack/freezer master: Use image timestamp for storage package naming  https://review.opendev.org/c/openstack/freezer/+/97538011:06
opendevreviewDmitriy Rabotyagov proposed openstack/freezer master: Use image timestamp for storage package naming  https://review.opendev.org/c/openstack/freezer/+/97538014:05
noonedeadpunk#startmeeting freezer15:00
opendevmeetMeeting started Mon Feb  2 15:00:25 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
Luzio/15:01
mheno/15:01
noonedeadpunk#topic elections/project governance15:04
noonedeadpunkSo election time is coming, and we need to decide if we want to continue governing the project as DPL model (distributed leadership)15:04
noonedeadpunkI'd think it might make sense now still, to spread to load15:05
mhenexcuse my ignorance, what does DPL imply?15:05
noonedeadpunk#link https://governance.openstack.org/tc/reference/distributed-project-leadership.html15:05
noonedeadpunkin short - instead of Project Team Leader which is responsible for everything, and they decide on delegating some aspects to volunteers15:06
mhenI see, thanks for the summary15:06
noonedeadpunkDPL where ppl self-sign under some responsibilities, so it's less administrative load to a single person15:09
Luzithis sounds like the better way right now instead of a ptl15:09
noonedeadpunkeventually, if we wanna continue with DPL, khyr0n should vote negatively on the patch to governance15:09
noonedeadpunk#link https://review.opendev.org/c/openstack/governance/+/97493115:10
noonedeadpunkbut we also can extend/add liasons if there are volunteers :)15:10
noonedeadpunkthe deadline is Feb 415:10
noonedeadpunkok, going next15:13
noonedeadpunk#topic image timestamp missmatch15:13
noonedeadpunkI took that as an action last week and has proposed a patch15:13
noonedeadpunk#link https://review.opendev.org/c/openstack/freezer/+/97538015:13
noonedeadpunkit's quite trivial and should be good to go15:15
noonedeadpunk#topic fixing Horizon plugin15:15
noonedeadpunkSo there's a renaming patch which I worked on a while back15:15
noonedeadpunk#link https://review.opendev.org/c/openstack/freezer-web-ui/+/97394915:15
noonedeadpunkthe problem with it, is that I messed up somewhere, and pointed templates to non-existing location15:16
noonedeadpunkI will try to push the update later today15:16
noonedeadpunk#topic elasticsearch deprecation15:16
noonedeadpunklast meeting we have agreed to deprecate elasticsearch storage engine as alternative to sqlalchemy15:17
noonedeadpunkI have proposed initial patches for this cycle to execute on it15:17
noonedeadpunk#link https://review.opendev.org/c/openstack/freezer-api/+/97484215:17
noonedeadpunk#link https://review.opendev.org/c/openstack/freezer-tempest-plugin/+/97484415:17
noonedeadpunkgiven this is a SLURP release, we should be able to drop code early in 2026.2 15:18
noonedeadpunk#topic restore-from-date has an unexpected behavior15:19
noonedeadpunkthis is regarding the bug report15:19
noonedeadpunk#link https://bugs.launchpad.net/freezer/+bug/213888315:19
noonedeadpunkso, thinking about this one, I'm having an opinion, that either option would be kida confusing15:20
noonedeadpunkso we need to change an approach to this option, and make it more "intended"15:20
mhenagreed, the restore-from-date would suggest an exact timestamp (e.g. from a list retrieved beforehand) and not suggest any fuzziness15:21
noonedeadpunkI think we should introduce to API some kind of "margin" when backup is selected, and fail the API request if the timeframe supplied by the user is "arguable" (or within the margin)15:22
noonedeadpunkwell, or that...15:22
mhenwe also need to be careful what interface we are talking about15:24
noonedeadpunkI was thinking, for instance, if there're like 2 backups, and user supplied somewhere in the middle, to have a conf option controllable by administrator, of percentage, within which reqiuuest will fail. So if difference in time is less then 20% - request will fail and have available options in response15:24
noonedeadpunkah, right, that's an agent thing15:25
mhenfreezer-agent does not really expose a list of backups, only api does15:25
mhenit discovers them on-the-fly when asked to do things15:25
noonedeadpunkso indeed, maybe we should do only exact timestamp for agent, and then attempt to do smth smarter for API?15:26
mhensounds reasonable15:26
noonedeadpunkyeah, ok , makes sense to me as well15:28
noonedeadpunkI think we might need to have a spec for this though, if we are to implement a new API call15:28
mhenregarding the margin approach: do you know any other backup products and how they handle this?15:28
mhenjust curious15:28
noonedeadpunkI frankly was not thinking about that. I'd say for all backups products I worked with - there was a database of existing backups, you could see right away, and it was all in UI15:29
noonedeadpunkand then I can't recall dealing with API part at all15:30
mhenbackups in freezer-agent don't have an ID iirc, do they have one in freezer-api?15:31
noonedeadpunkYes, they do15:31
noonedeadpunkat least according to what I see in schema15:32
mhenI was just thinking about putting the ID from the 'GET /backups' list into a job for freezer-agent but I guess freezer-agent does not know about the IDs15:33
noonedeadpunkwell, it kinda makes sense. I think it also depends if we want to make API part as a required component15:34
mhenthen you could offer the "select backup from list" approach as usual15:35
noonedeadpunkit's *kinda* useful to be able to use freezer-agent in standalone mode15:35
noonedeadpunkwhen your cloud provider does not want to spawn API, you still are able to backup things...15:35
noonedeadpunkbut it adds quite some unnecessary complexity for sure15:36
mhenas an agent within the guest VM?15:36
noonedeadpunkyeah15:36
mhenimho sounds more like a stopgap measure in contrast to other OpenStack services offered as an integral part of the cloud15:38
mhenI like the central scheduler approach more ;)15:39
noonedeadpunkIt would be ideal to have both somehow....15:39
noonedeadpunkas then we might put pressure on providers to adopt API to help users :D15:39
noonedeadpunkstruggling with local setups on VMs15:40
noonedeadpunkbut yeah15:40
noonedeadpunkthus question - where in priority do we see this restore-from-date refactoring?15:40
noonedeadpunkI'd say that we shouldn't chase it for 2026.115:41
noonedeadpunkMaybe we can update the doc to reflect actual behavior now15:41
mhenfor the time being, fixing the parameter's description would at least fix the interpretation issue15:41
noonedeadpunkand scoper refactoring for 2026.215:41
noonedeadpunkok, agree then15:42
mhenwe shouldn't rush it, the way to specify the restoration of backups is absolutely crucial15:42
noonedeadpunkI've recorded the decision in https://etherpad.opendev.org/p/freezer_meetings15:43
mhensounds good15:44
noonedeadpunkok, great15:44
noonedeadpunk#topic open discussion15:44
noonedeadpunkI'm done with agenda for today - is there anything else to raise/discuss this time?15:44
mhenthanks for merging all my patchsets btw ;)15:44
noonedeadpunkthanks for proposing them :)15:45
mhenI'll see if I can return the favor and test/review some of yours this week15:45
noonedeadpunkI'd really love some level of review for my patches. 15:45
noonedeadpunkyeah15:45
noonedeadpunkfor now I was like - if some issue is still present - we can patch it quickly. At the same time attempting to grow CI coverage15:46
noonedeadpunkI was also thinking to reach back some users who attempted to test freezer lately and failed due to clients being outdated and some other things15:47
noonedeadpunkBut probably I'd want to ensure that Horizon plugin is fixed first15:47
noonedeadpunkok, cool thanks for taking time and joining the meeting15:48
noonedeadpunksee you folks around!15:48
noonedeadpunk#endmeeting15:48
opendevmeetMeeting ended Mon Feb  2 15:48:35 2026 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)15:48
opendevmeetMinutes:        https://meetings.opendev.org/meetings/freezer/2026/freezer.2026-02-02-15.00.html15:48
opendevmeetMinutes (text): https://meetings.opendev.org/meetings/freezer/2026/freezer.2026-02-02-15.00.txt15:48
opendevmeetLog:            https://meetings.opendev.org/meetings/freezer/2026/freezer.2026-02-02-15.00.log.html15:48
opendevreviewDmitriy Rabotyagov proposed openstack/freezer-web-ui master: Move freezer panel under Projects dashboard  https://review.opendev.org/c/openstack/freezer-web-ui/+/97394915:58
opendevreviewDmitriy Rabotyagov proposed openstack/freezer-web-ui master: Register a panel for API url  https://review.opendev.org/c/openstack/freezer-web-ui/+/97397216:00
opendevreviewMerged openstack/freezer-web-ui master: Remove information about removed option "--remove-older-then"  https://review.opendev.org/c/openstack/freezer-web-ui/+/96131116:14
opendevreviewDmitriy Rabotyagov proposed openstack/freezer-web-ui master: Do not raise KeyError when sessions/jobs do not exist  https://review.opendev.org/c/openstack/freezer-web-ui/+/97542717:02
opendevreviewDmitriy Rabotyagov proposed openstack/freezer master: Fix restore-from-date behavior  https://review.opendev.org/c/openstack/freezer/+/97543717:54
opendevreviewDmitriy Rabotyagov proposed openstack/freezer master: Fix restore-from-date behavior  https://review.opendev.org/c/openstack/freezer/+/97543721:59

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