Thursday, 2026-09-24

opendevreviewMerged openstack/watcher master: zuul: Drop redundant override of cinder option  https://review.opendev.org/c/openstack/watcher/+/99947707:35
opendevreviewAlfredo Moralejo proposed openstack/watcher master: DNM test third-party jobs with s2i  https://review.opendev.org/c/openstack/watcher/+/100714412:23
dviroelhello everyone, watcher meeting will start in 8 minutes.. o/12:52
dviroelhere is the meeting agenda https://etherpad.opendev.org/p/openstack-watcher-irc-meeting#L2412:53
dviroel#startmeeting watcher13:00
opendevmeetMeeting started Thu Sep 24 13:00:57 2026 UTC and is due to finish in 60 minutes.  The chair is dviroel. Information about MeetBot at http://wiki.debian.org/MeetBot.13:00
opendevmeetUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.13:00
opendevmeetThe meeting name has been set to 'watcher'13:00
amoralejo/13:01
morenodo/13:01
rlandyo/13:01
dviroelcourtesy ping: 13:02
dviroel: )13:02
dviroelI think that joan is out too, so I didn't sent the ping13:02
dviroellets get started13:02
dviroellet's start with today's meeting agenda13:03
dviroel#link https://etherpad.opendev.org/p/openstack-watcher-irc-meeting#L30 (Meeting agenda)13:03
opendevreviewDavid proposed openstack/watcher-tempest-plugin master: Add test suite for audit scope  https://review.opendev.org/c/openstack/watcher-tempest-plugin/+/99579913:03
dviroelplease fill free to add topics to the agenda13:03
dviroellets start with13:04
dviroel#topic Announcements13:04
dviroeli added a few today13:04
dviroelfirst one13:04
dviroelHibiscus Schedule13:04
dviroeljust to track our release13:04
dviroelwe are at R-1 week, which means that Hibiscus release is officially next week13:04
dviroelwe didn't have any major fix to backport and request a new RC13:05
dviroelso we are good and on tracj13:05
dviroelnext announcement13:05
dviroelPTG - October 12-16, 202613:05
dviroelwe are just a few weeks/days from out PTG13:06
dviroelin case you didn't register, please take a minute and complete the registration at https://ptg.openinfra.org/13:06
dviroelour etherpad is the following13:06
dviroel#link https://etherpad.opendev.org/p/watcher-2027.1-ptg13:06
dviroeli see that you all started populating with topics, thanks for that13:07
dviroeli will be working on defining the slots for each one13:07
dviroeland groups the ones that are releted13:07
dviroelnote that, Cactus room is now booked for watcher on Wed and Thu13:08
dviroel#link https://ptg.opendev.org/ptg.html13:08
amoralejthanks, i need to check the performance discussions from previous ptg, but i think we have enough pending work for next release13:08
amoralejfrom the scalability work i mean13:08
dviroelack, it is good to revisit and check what is pending yet and plan some work items too13:09
dviroeli also plan to revistit the previous AI from past ptg13:09
dviroelwe can also sync later which other sessions may be interesting to join, and if conflict with ours13:10
amoralejyep, that's good13:10
dviroelat this point, i don't have a plan for a cross session with other teams13:10
dviroelnova/cyborg/telemetry13:11
dviroelbut we may have until the ptg, lets see13:11
dviroelbased on new spec/features proposals13:11
dviroelalright, feel free to include topics that may be missing there.13:11
dviroelnext announcement 13:11
dviroelOpenInfra Live in Oct 1st at 14:00 UTC 13:11
dviroelwe will have OpenInfra Live next thursday13:12
dviroel14:00 UTC in the OpenInfra youtube channel13:12
dviroelI will be presenting the watcher's most impactful features and cover what we are planning for 2027.113:13
dviroelit will be right after our weekly meeting13:13
dviroelany other announcement? 13:13
dviroelalright13:14
dviroel#topic Deprecation Proposals13:14
dviroelI want to bring these proposals here today13:14
dviroeland see what other think13:14
dviroelwe don't need to agree on them yet13:15
dviroelactually, the first one we already discussed at previous ptg13:15
dviroel1. Deprecate the Grafana datasource13:15
dviroel#link https://review.opendev.org/c/openstack/watcher/+/100677113:15
dviroelwe agree on proposing this during 2026.2 ptg13:15
amoralejwe even sent a mail about that to openstack-discuss, right?13:16
dviroeli don't remember for Grafana TBH13:16
dviroeli remember sending for MAAS (which we already removed)13:16
dviroelbut yes, the idea is to sent a email today about these deprecations13:17
dviroelthis is for the datasources13:17
dviroelthe next one is for strategies13:17
dviroel2. 1006323: Deprecate untested strategies and refresh status matrix13:17
dviroel#link https://review.opendev.org/c/openstack/watcher/+/100632313:18
dviroelnoisy neighbor was already deprecated in the past13:18
amoralejwrt graphana, i'm +2 13:18
dviroeli am proposing, and want to discuss with the team about the other ones13:18
dviroelwe may want to also wait until the ptg to agreed on that if needed13:19
dviroelfor the strategies ^13:19
amoralejwrt the untested strategies, i also agree although it deserves some discussion in pthg13:19
dviroeloutlet_temp_control13:19
dviroelsaving_energy13:19
dviroeluniform_airflow13:19
dviroelthese 3 lacks testing, and while working on a small Strategy refactoring, claude pointed to some bugs on them13:20
dviroeli opened some bugs that we can cover in a few, based on some of them which we can also discuss about13:21
amoralejwhat level of bugs? i mean, any is fundamentally broken as the noisy neighbour ?13:21
dviroelyeah, we may want to cover them in bugs topic, but the original proposal is really based on lack of testing and maintenance13:23
dviroelthere are strategies that create actions that are fundamentally broken13:23
dviroeland these strategies are already set as experimental for a couple of releases 13:24
dviroel2 releases at least13:24
dviroelbut I set the change to W-113:24
dviroelso we can discuss with time13:24
rlandynobody has inquired about these strategies?13:25
amoraleji like the idea of focusing in known good strategies proved solid13:25
amoralejand growing13:25
rlandyor logged any bugs?13:25
dviroelno, not recently at least13:26
amoralejbut there may be work worthy to maintain there or maybe even to think if it's better to follow a different approach, i.e. integrating functionalities in supported strategies13:26
dviroelso yeah, maybe we should discuss at PTG first, saving_energy is one that is related with the ironic work that jgilaber is working on right now13:27
dviroelso we may keep it or not based on the decision around ironic integration too13:27
amoralej+113:28
dviroelsince we are discussin about them, I think that makes sense to jump to Bugs topic, and get reviews after13:28
dviroel#topic Bugs13:29
dviroelstarting with the first one from jgilaber13:29
dviroel#link https://bugs.launchpad.net/watcher/+bug/216785213:29
dviroelCinder collector crashes on volumes in "creating" state with null host 13:29
amoralejseems valid issue, we need to skip those volumes from the collector13:30
dviroelok, so this one jgilaber wrote based on ci failures, not on specific investigation13:31
dviroelagree, seems valid13:32
dviroeland should be easy to fix13:32
dviroelaffects test_execute_zone_migration tests, which I saw faling before13:33
dviroelbut didn't investigate the failure at that time13:33
dviroelwhich priority should we set?13:34
amoraleji'd put medium13:34
dviroels/priority/importance13:34
dviroeli would go with medium too13:34
dviroelwe should fix this 13:35
dviroelok, next one13:35
dviroel#link https://bugs.launchpad.net/watcher/+bug/216787113:36
dviroeloutlet_temp_control and noisy_neighbor emit Migrate actions with node UUIDs13:36
dviroelthis is one of the bugs that is building wring Migrate actions 13:37
dviroelthis is a sideeffect of not using a common helper to build them, and something that I would like to fix in the future13:37
amoralejuh13:38
dviroelthe discussion would be around if makes sense to keep maintaining these strategies and proposing fixes13:38
dviroelnoisy_neighbohr we already deprecated, but didn't remove it yet13:38
amoralejyes, we should standardize how we call the migrate action anyway13:39
dviroelyeah, I plan to propose that in a spec for this cycle13:40
amoralejor fix the pre_condition if we want to keep supporting both13:40
amoralejbut, i'd be more in favor of just supporting one way13:40
dviroelthe action today only accepts hostname 13:41
amoralejah, sorry, i misunderstood, i though it was only a problem with the pre_condition13:42
amoralejso we'd need to fix the strategy to use hostnames13:42
dviroelyes, not uuids13:42
dviroelhttps://github.com/openstack/watcher/blob/master/watcher/applier/actions/migration.py#L3113:43
dviroel"The `source_node` and `destination_node` parameters are respectively the source and the destination compute hostname."13:43
amoralejack, then it's clear13:44
dviroelset an importance? 13:44
dviroelLow based on current support on these strategies?13:45
dviroelone is deprecated, other is experimental 13:45
dviroelsetting it to Low for now13:46
dviroelnext one13:46
dviroel#link https://bugs.launchpad.net/watcher/+bug/216787413:46
dviroelanother issue with noisy_neighbor, which is deprecated, but it is important to highlight that13:47
amoralejlow lgtm13:47
dviroelmultiple strategies check compute nodes availability and status13:47
dviroeldifferent implementation for similar checks13:48
dviroelanother piece that can be refactored 13:48
dviroel"noisy_neighbor selects offline and disabled compute nodes as migration destinations"13:48
dviroelnoisy neighbor don't check the state/status of a destination node13:48
dviroeli would still set to Low, noisy_neighbor is deprecated13:49
dviroeli would say that this one could even be Won't Fix, since it is deprecated for removal13:50
dviroelany other concern?13:51
dviroelany other Bug to highlight?13:51
amoralejwon't fix is good imo13:52
dviroelack13:52
amoralejnothing else from my side13:52
dviroel#topic Reviews13:52
dviroelthere is no specific review added in the etherpad13:52
dviroelso I just linked the review etherpad13:53
dviroelsome small fixes don't even get into the etherpad, but I plan to add the deprecation ones there13:53
dviroel#link https://etherpad.opendev.org/p/watcher-2027.1-status13:53
dviroelfor the records13:53
dviroelbtw amoralej can you check this one:13:54
dviroel#link https://review.opendev.org/c/openstack/watcher-specs/+/100443613:54
dviroelsee what you think13:54
dviroelthere are specs open, other ones may be proposed until the ptg 13:54
amoralejthat's just moving existing specs from thee previous ones, right?13:55
dviroelyes13:55
dviroelthe eventlet ones are merged13:55
dviroelthere is still one WSGI script removal open13:55
amoralejsorry, moving them to implemented13:55
dviroelwhich we are holding for a few more days13:55
dviroel#link https://review.opendev.org/c/openstack/watcher/+/1006113/13:56
dviroelamoralej: yes, not all, because not all were implemented :)13:56
dviroelany other review that requires attention?13:56
dviroelwe are almost out of time13:56
dviroel#topic ok13:57
amoraleji'll review13:57
dviroel#undo13:57
opendevmeetRemoving item from minutes: #topic ok13:57
dviroel#topic Volunteers to chair next week13:57
dviroeli  will be around13:57
dviroeli can chair, i  will just need to join the openinfra live at the final 10m of the meeting13:58
dviroel10m before starting the openinfra live13:58
dviroelwhich is fine13:58
amoraleji can chair13:58
amoralejso that you have no problem to join the openinfra live13:58
dviroelamoralej: that would be great13:58
dviroelthanks!13:58
amoralejno problem13:58
dviroel#topic Open Discussions13:59
dviroellast minute13:59
dviroelif there is nothing to highlight here, we can close13:59
dviroel...13:59
dviroelwe will meet again next week then13:59
dviroelthank you all for participating13:59
dviroel#endmeeting13:59
opendevmeetMeeting ended Thu Sep 24 13:59:47 2026 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)13:59
opendevmeetMinutes:        https://meetings.opendev.org/meetings/watcher/2026/watcher.2026-09-24-13.00.html13:59
opendevmeetMinutes (text): https://meetings.opendev.org/meetings/watcher/2026/watcher.2026-09-24-13.00.txt13:59
opendevmeetLog:            https://meetings.opendev.org/meetings/watcher/2026/watcher.2026-09-24-13.00.log.html13:59
*** haleyb is now known as haleyb|out23:39

Generated by irclog2html.py 4.1.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!