Wednesday, 2026-03-25

opendevreviewBrian Rosmaita proposed openstack/python-brick-cinderclient-ext master: Add periodic jobs for brick-cinderclient-ext  https://review.opendev.org/c/openstack/python-brick-cinderclient-ext/+/98203402:13
opendevreviewBrian Rosmaita proposed openstack/python-brick-cinderclient-ext master: DNM: test proposed periodic-weekly jobs  https://review.opendev.org/c/openstack/python-brick-cinderclient-ext/+/98203502:16
*** zseguin_ is now known as zseguin06:11
opendevreviewVictor Chembaev proposed openstack/os-brick master: Multipath hangs on detach  https://review.opendev.org/c/openstack/os-brick/+/98206710:24
opendevreviewArtem Goncharov proposed openstack/cinder master: Fix jsonschema parameter  https://review.opendev.org/c/openstack/cinder/+/98207912:33
opendevreviewSanketh Shetty proposed openstack/cinder master: Add support for AutomatedFailOver policy with consistent replication  https://review.opendev.org/c/openstack/cinder/+/97588012:59
rosmaitaCinder meeting in #openstack-cinder at 1400 UTC14:01
rosmaitajungleboyj rosmaita smcginnis tosky whoami-rajat m5z e0ne geguileo eharney jbernard hemna fabiooliveira yuval tobias-urdin adiare happystacker dosaboy hillpd msaravan sp-bmilanov Luzi sfernand simondodsley  zaubea nileshthathagar flelain wizardbit agalica lutimura kaisers14:01
rosmaita#startmeeting cinder14:01
opendevmeetMeeting started Wed Mar 25 14:01:55 2026 UTC and is due to finish in 60 minutes.  The chair is rosmaita. Information about MeetBot at http://wiki.debian.org/MeetBot.14:01
opendevmeetUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.14:01
opendevmeetThe meeting name has been set to 'cinder'14:01
rosmaita#topic roll call14:02
jbernardo/14:02
raghavendrathi14:02
toskyo/14:02
jbernard#link https://etherpad.opendev.org/p/cinder-gazpacho-meetings14:02
rosmaita#chair jbernard14:02
opendevmeetCurrent chairs: jbernard rosmaita14:02
jbernardrosmaita: thanks14:02
hvlcchao1o/14:02
whoami-rajathi14:03
VolodymyrBoiko[m]o/14:04
jbernardwelcome everyone14:05
jbernardthis is the final gazpacho meeting, next week will be the opening of hibiscus14:05
jbernard#link https://etherpad.opendev.org/p/cinder-hibiscus-meetings14:05
jbernardwe are preparing the final RC14:06
jbernardit is due Thursday14:06
jbernardrosmaita: did you add the powerflex notes?14:07
rosmaitayes14:07
jbernardtake it away14:07
rosmaitai think we can't include these in the RC because os-brick is frozen14:07
rosmaitaand i think the cinder patch requires the os-brick change14:07
jbernardcan we roll them both into rc2?14:08
rosmaitano, can't release a new os-brick14:08
jbernardor post release backport?14:08
rosmaitayes, we can backport them after release14:09
rosmaitathose were the only changes proposed for stable/2026.114:09
rosmaitaso unless someone speaks up now, looks like we will use RC1 for the release14:09
rosmaitaunless we need to revert a change14:10
rosmaitawhich we will discuss later14:10
chembervinthi, I'm for the first time on the meeting in this channel. I've made a small but important patch to os-brick, which is needed to fix cinder-backup with mutipath devices. when it could be reviewd? thanks you in advance!14:12
chembervinthttps://review.opendev.org/c/openstack/os-brick/+/98206714:12
jbernardso to be clear, are there any other bugs that you might be working on that haven't been proposed to stable 2026.1 yet?14:12
rosmaita"you" meaning everyone present14:13
jbernardyup14:13
jbernardwe can start with topics14:13
jbernard#topic cinderclient gates14:14
jbernard#link https://review.opendev.org/q/topic:%22periodic-cinder%2214:14
rosmaitayou all may remember some unpleasantness just before cinderclient release this cycle14:14
rosmaitato avert that next time, we decided to add periodic jobs to make sure the gates are not blocked14:15
rosmaitaso, those are the patches14:15
rosmaitai just need some quick reviews and approval14:15
rosmaitabecause we will need to revise the zuul.yaml in all cinder projects for the hibiscus supported pythons soon-ish14:16
rosmaitathat's all14:16
jbernardwill review14:17
erlon\o14:17
rosmaitathanks14:17
Anoop_Shukla\o14:18
kaiserso/14:18
jbernard#topic unit test job failures due to cinder.tests.unit.volume.drivers.hpe.test_hpe3par.TestHPE3PARDriverBase.test_thread_local_sessions_are_isolated (rosmaita)14:18
rosmaitayeah, so this is what i was talking about when i said maybe we need to revert a change from gazpacho14:19
rosmaitait's an intermittent failure14:19
raghavendrathi14:19
jbernardhow intermittent?14:19
rosmaitai've seen it mostly on python 3.1314:20
jbernarddo you have a sense?14:20
rosmaitawell, it's happened on a few patches i was reviewing14:20
eharneyi saw it on one of the eventlet removal patches too14:20
rosmaitabut it also sometimes passes14:20
jbernardi lean towards reverting14:21
rosmaitaeharney: do you remember if it was py313 or another python?14:21
eharneynot sure, i can probably look14:21
raghavendratrevert meaning: reverting just that single UT function ?14:21
rosmaitaraghavendrat: i'm not sure14:21
rosmaitait's not clear whether it's a test artefact, or whether the functionality is hosed14:22
rosmaita"hosed" == "unreliable"14:22
raghavendrati just wish to elaborate14:22
rosmaitasure, please go ahead14:22
raghavendratthat UT function was added in: https://review.opendev.org/c/openstack/cinder/+/97659614:23
raghavendratthis patch was merged on March 09th14:23
raghavendrator March 10th early morning14:24
raghavendratthis patch is important for HPE; because it addresses hot customer fix.14:24
erlondoenst having a driver creating threads breaks the cinder driver model?14:25
raghavendratmy humble request would be: in case revert has to be done; then is it possible to submit a patch that removes just that UT function please14:25
erlonI haven't reviewed that patch, but I haven't seen that kind of implementation in drivers elsewhere14:26
eharneyerlon: the driver is actually creating greenthreads IIRC14:26
erlonCan those continue running after the function call returns to the volume manager? 14:27
eharneysorry, it isn't creating them, it's using the ones provided by the cinder, but is a bit different in that it's using (green)thread-local storage14:29
raghavendratwe have performed extensive testing over a period of 1 month; and the functionality is working fine14:30
rosmaitaraghavendrat: so you are inclined to think it is a test-only problem?14:31
jbernardwhere is the flaw in the test?14:32
raghavendratyes rosmaita:14:32
erlonraghavendrat: our concern here is that this model might be breaking something else on Cinder. It might be just unit tests though, but it might be related to this threads handling that's not usually used in drivers14:32
eharneydo we know if the failure looks the same each time?14:33
rosmaitaanecdotally, i have seen the mismatch error for line 722814:34
rosmaitaa few times14:34
eharneyi see a few in the build history that look exactly like https://zuul.opendev.org/t/openstack/build/057f9f567ccb4a15b4715d66e973720c14:34
rosmaitabut i don't remember whether it's always the same mismatch values or not14:34
erlonraghavendrat: if you spin a unit test, then the unit tests calls a function and this function creates a thread, the test might finish before the function returns14:35
erlonsee:14:38
erlonTraceback (most recent call last):14:38
erlon  File "/usr/lib/python3.13/threading.py", line 1385, in __del__14:38
erlonTypeError: 'NoneType' object does not support the context manager protocol14:38
erlonthe threading termination code is ending after the unit tests finishes, hence the intermitent behaviour.14:38
erlonI don't think this (having a thread) would interfere with the rest of the cinder workings, but is a practice that should be avoided. I believe that removing only the test (but leaving the code untested?) should be ok14:40
rosmaitaok, so raghavendrat can you file a bug for this issue14:41
rosmaitaand then put up a patch to skip test citing the bug14:41
raghavendratok, i will submit a launchpad bug14:41
rosmaitaand then tell us in cinder channel14:41
rosmaitaand then once it merges to master, please backport to stable/2026.114:42
rosmaitaand then we will merge it into gazpacho and cut a new RC-214:42
raghavendratsure, i will inform14:42
rosmaitaok, great14:42
rosmaitathanks erlon, eharney, and raghavendrat14:42
raghavendratthank you all14:43
jbernard#topic ogging in drivers (rosmaita)14:43
rosmaitayeah, i just can't keep my mouth shut14:43
jbernard:)14:43
rosmaitathis is something i noticed in doing some reviews14:43
rosmaitajust want to remind driver maintainers14:43
rosmaitawe understand that you want the debug logs to be informative14:44
rosmaitabut remember that people will file security bugs if passwords show up in the logs14:44
erlondo you have the links14:44
Anoop_ShuklaAre there any specific examples?14:44
rosmaitasure, on the agenda14:45
rosmaita#link https://review.opendev.org/c/openstack/os-brick/+/97744714:45
Anoop_ShuklaDo we have any guidelines on logging for Cinder core and drivers today?14:46
rosmaitakey thing is if you are logging a dict of info, use the oslo_utils.strutils.mask_dict_password() function14:46
rosmaitaAnoop_Shukla: probably in the driver checklist14:46
rosmaitabut Anoop_Shukla that is is good point, if you don't see it somewhere obvious, let us know or put up a docs patch14:47
Anoop_Shuklahttps://docs.openstack.org/cinder/latest/contributor/new_driver_checklist.html14:47
jbernardyeah, we should have a permalink to point to for that14:48
Anoop_ShuklaMay be we can improve the checklist and talk about logging do’s and donts.14:48
erlonwouldnt be nice if LOG.debug() already did that for all logs? There will be always cases like that that is hard to tell what's inside some data structures14:48
rosmaitaAnoop_Shukla: that would be a good topic for a PTG session, to do a review and brainstorm additions, if you want to propose it14:49
Anoop_ShuklaI can take it up. Thanks for the suggestion @rosmaita14:49
rosmaitaAnoop_Shukla: ++14:50
Anoop_ShuklaOn that topic..do we have the PTG etherpad already available?14:50
erlonIt shouldn't be hard to wrap around it, check if the parameters are dictionaries, and mask them. But that is a oslo thing I guess14:50
jbernardhttps://etherpad.opendev.org/p/cinder-hibiscus-ptg14:51
jbernardAnoop_Shukla: ^14:51
Anoop_ShuklaGreat thanks!14:51
jbernard#topic open discussion14:51
rosmaitaerlon: might be worth looking into, not sure why it didn't get integrated directly into oslo.logging14:52
Anoop_ShuklaOne topic that came up sometime back was about guidelines on how backup vendors use Cinder backup workflows..14:53
erlonrosmaita, what would be the teams to bring it? I mean specifically for the ptg discussions, we might want to take that to a broader public14:53
rosmaitai think there should be some kind of oslo ptg session14:54
rosmaitabut i don't know for sure14:54
Anoop_ShuklaToday there are no standard guidelines for applications like commvault, Veeam, Rubrik, Veritas14:54
Anoop_ShuklaIn general these applications use Cinder workflows and Nova to perform their backups14:55
Anoop_ShuklaFor Cinder drivers there is no way of knowing if its a Cinder Snapshot or a backup workflows14:55
Anoop_ShuklaDoes oslo have ptg session?14:56
rosmaitawell, there's not an oslo-specific etherpad here: https://ptg.opendev.org/etherpads.html14:57
raghavendratwish to ask for review: we have addressed review comments on below patch.14:58
raghavendratEarlier there was one +2, but it went away after code was updated to address review comment.14:58
Anoop_ShuklaMay be I will put it as a topic to brainstorm from Cinder perspective and we can take it from there.14:58
raghavendratRequest to please re-visit & approve (if all is good). Thanks.14:58
raghavendratHPE: Improved handling of QOS for Alletra MP14:59
raghavendrathttps://review.opendev.org/c/openstack/cinder/+/97911514:59
rosmaitamight be worth asking on the mailing list if oslo is going to have a ptg presence14:59
rosmaitaraghavendrat: i will re-review15:00
raghavendratthank you15:00
Anoop_ShuklaReview request for the patch: https://review.opendev.org/c/openstack/cinder/+/977911 NetApp CI is successful. I have 1 +1 from agalica - thank you for that :)15:01
Anoop_ShuklaI had another question to folks here - Does anyone have any experience migrating from ZUUL to Jenkins for CI?15:03
jbernardi know there is knowledge within the team, but i cannot recall who mentioned it last15:04
Anoop_ShuklaYes..someone did mention in mid-cycle about there experiences..15:04
Anoop_ShuklaPlease PM me if anyone can help :)15:04
jbernardi recall jenkins being reported as not-so-bad to setup15:04
Anoop_ShuklaWe are looking to upgrade ZUUL on our CI, but also weighing in the cost to move to Jenkins and its pros/cons15:05
jbernardlast call before we wrap up15:08
jbernardthanks everyone15:10
jbernard#endmeeting15:10
opendevmeetMeeting ended Wed Mar 25 15:10:40 2026 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)15:10
opendevmeetMinutes:        https://meetings.opendev.org/meetings/cinder/2026/cinder.2026-03-25-14.01.html15:10
opendevmeetMinutes (text): https://meetings.opendev.org/meetings/cinder/2026/cinder.2026-03-25-14.01.txt15:10
opendevmeetLog:            https://meetings.opendev.org/meetings/cinder/2026/cinder.2026-03-25-14.01.log.html15:10
opendevreviewEric Harney proposed openstack/cinder master: Enable threads mode in cinder-volume  https://review.opendev.org/c/openstack/cinder/+/95465015:16
opendevreviewEric Harney proposed openstack/cinder master: Replace _add_to_threadpool with native threads  https://review.opendev.org/c/openstack/cinder/+/95938315:16
opendevreviewEric Harney proposed openstack/cinder master: cinder/monkey_patch.py for threading mode  https://review.opendev.org/c/openstack/cinder/+/97033215:16
opendevreviewRaghavendra Tilay proposed openstack/cinder master: HPE: Remove UT that fails intermittently  https://review.opendev.org/c/openstack/cinder/+/98212916:14
raghavendrathi jbernard: rosmaita: as discussed in weekly meeting, created launchpad bug16:16
raghavendrathttps://bugs.launchpad.net/cinder/+bug/214633916:16
raghavendratThe UT test_thread_local_sessions_are_isolated fails intermittently & causes zuul to fail16:16
raghavendratand submitted below patch to remove UT function:16:16
raghavendrathttps://review.opendev.org/c/openstack/cinder/+/98212916:17
cardoehello. just wanted to ask about https://review.opendev.org/c/openstack/cinder/+/962085 it was previously on the review etherpad.18:02
jbernardcardoe: i missed circling back on that one, my appologies, looks okay just needs another +218:11
opendevreviewMerged openstack/cinder master: HPE: Improved handling of QOS for Alletra MP  https://review.opendev.org/c/openstack/cinder/+/97911518:43
opendevreviewJon Bernard proposed openstack/os-brick master: Remove eventlet from tests and requirements  https://review.opendev.org/c/openstack/os-brick/+/97318418:53
jbernarderlon: did you Niklas' last comment on https://review.opendev.org/c/openstack/cinder/+/484729 ?19:01
jbernarderlon: curious if you have any followup19:01
opendevreviewJon Bernard proposed openstack/cinder master: Add a non-voting job to test eventlet removal  https://review.opendev.org/c/openstack/cinder/+/98214419:12
opendevreviewEric Harney proposed openstack/cinder master: Release notes: move hitachi-vsp-one note  https://review.opendev.org/c/openstack/cinder/+/98214819:20
eharneyjbernard: fyi we landed a releasenote in the wrong dir ^19:20
jbernardeharney: i was just looking at that, i dont see it in my tree... i am losing my mind19:23
eharneyjbernard: it's hard to see because it has a leading space19:23
eharneyhttps://opendev.org/openstack/cinder/src/branch/master/%20releasenotes/notes19:23
jbernardooooh19:25
jbernardwow, i missed that19:25
jbernardthanks19:25
eharneyi looked at the original review in gerrit and it's almost invisible there. lol19:25
opendevreviewEric Harney proposed openstack/cinder master: Enable threads mode in cinder-volume  https://review.opendev.org/c/openstack/cinder/+/95465019:36
opendevreviewEric Harney proposed openstack/cinder master: Replace _add_to_threadpool with native threads  https://review.opendev.org/c/openstack/cinder/+/95938319:36
opendevreviewEric Harney proposed openstack/cinder master: cinder/monkey_patch.py for threading mode  https://review.opendev.org/c/openstack/cinder/+/97033219:36
opendevreviewAnthony Galica proposed openstack/cinder stable/2024.2: Hitachi: Add support for Hitachi VSP One B20  https://review.opendev.org/c/openstack/cinder/+/98216420:43
agalicajbernard: oof, I guess that's my error.  Need me to do anything about that?21:13
agalica(other than mind my extra spaces next time ~boing~21:14
agalicaoh, looks like Eric already created a patch set21:14
agalicaand you already workflowed it21:15
agalicanever mind me21:15
opendevreviewMerged openstack/cinder master: Release notes: move hitachi-vsp-one note  https://review.opendev.org/c/openstack/cinder/+/98214822:41

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