Wednesday, 2026-06-24

opendevreviewMerged openstack/cinder master: Replace eventlet sleep calls in volume drivers  https://review.opendev.org/c/openstack/cinder/+/95069500:43
opendevreviewAmit Uniyal proposed openstack/cinder master: Adds urllib3 lib in requirements.txt  https://review.opendev.org/c/openstack/cinder/+/99361507:12
opendevreviewMaximilian Brandt proposed openstack/cinder master: RBD: server-side copy when uploading a volume to an image  https://review.opendev.org/c/openstack/cinder/+/99446707:49
opendevreviewRajat Dhasmana proposed openstack/cinder master: NetApp: Optimize ensure_export to skip LUN table refresh  https://review.opendev.org/c/openstack/cinder/+/99460007:51
opendevreviewRajat Dhasmana proposed openstack/cinder master: NetApp: Optimize ensure_export to skip LUN table refresh  https://review.opendev.org/c/openstack/cinder/+/99460008:21
opendevreviewFranciszek Przewoźny proposed openstack/cinder master: NetApp: Fix in-use boot and attach extension  https://review.opendev.org/c/openstack/cinder/+/95290010:15
opendevreviewFuli Qi proposed openstack/cinder-specs master: Add spec for retype volume including snapshots  https://review.opendev.org/c/openstack/cinder-specs/+/98932210:20
opendevreviewClaudiu Belu proposed openstack/cinder master: tests: Enforce autospec=True on mock patches  https://review.opendev.org/c/openstack/cinder/+/99415211:22
opendevreviewClaudiu Belu proposed openstack/cinder master: tests: Enforce autospec=True on mock patches  https://review.opendev.org/c/openstack/cinder/+/99415211:43
opendevreviewClaudiu Belu proposed openstack/cinder master: tests: Enforce autospec on mock objects and patches  https://review.opendev.org/c/openstack/cinder/+/99468611:43
hong-p[m]jbernard: please review this spec, Add spec for incremental backup with CBT optimization | https://review.opendev.org/c/openstack/cinder-specs/+/985089 already got +2, +112:37
eharneyi'll make another review pass on that one soon12:40
opendevreviewElod Illes proposed openstack/os-brick master: DNM: gate health test  https://review.opendev.org/c/openstack/os-brick/+/99472013:19
opendevreviewElod Illes proposed openstack/python-brick-cinderclient-ext master: DNM: gate health test  https://review.opendev.org/c/openstack/python-brick-cinderclient-ext/+/99472313:20
opendevreviewElod Illes proposed openstack/python-cinderclient master: DNM: gate health test  https://review.opendev.org/c/openstack/python-cinderclient/+/99472413:20
jbernard#startmeeting cinder14:02
opendevmeetMeeting started Wed Jun 24 14:02:56 2026 UTC and is due to finish in 60 minutes.  The chair is jbernard. Information about MeetBot at http://wiki.debian.org/MeetBot.14:02
opendevmeetUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.14:02
opendevmeetThe meeting name has been set to 'cinder'14:02
jbernard#link https://etherpad.opendev.org/p/cinder-hibiscus-meetings14:03
jbernardhmm, the courtesy ping nicks have disappeared14:03
rosmaitaprobably just as well, we can ask people to re-populate14:04
rosmaitaotherwise it's difficult to tell who doesn't care anymore14:04
* fungi is around for his topic again14:04
jbernardfixed14:04
jbernardoh, well, we could do that too14:04
jbernardfeel free to unfix ;)14:05
rosmaitafixed is good14:05
jbernardcourtesy ping: jungleboyj 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 kaisers aloke_dev Anoop_Shukla erlon 14:05
Anoop_ShuklaO/14:06
VolodymyrBoiko[m]Hi14:06
harsho/14:06
jbernardo/14:06
fungiprotip: i just add "#startmeeting whatever" to the channel highlights in my irc client, never miss a meeting i care about14:07
agalica_o/14:07
erlon\o14:07
fqio/14:07
rosmaitao/14:07
Jyotsnao/14:07
auniyalo/14:08
rosmaitafungi: that is a good tip!14:08
fungimostly helpful when you have a dozen or more different weekly irc meetings to keep track of all over the place14:09
jbernardok, let's get started14:09
jbernard#topic annoucements14:09
jbernardperiodic jobs are still looking good14:10
rosmaita\o/14:10
jbernardor, they're all false negatives and everything is broken14:10
jbernardeither way, green!14:10
jbernardthanks to whoever shortened that link14:11
jayaanandO/14:11
jbernardlast minute addition,it looks like we have a 3rd party CI that has become an issue14:11
jbernardthe "Ranjith" CI14:12
jbernarddoes anyone know who's running this?14:12
agalica_Why, Ranjith is!14:12
jbernardlol, but of course!14:12
erlonIs it floading in the patches?14:13
erlonow I see14:14
rosmaitahttps://review.opendev.org/q/cc:goldranjith@gmail.com14:15
jbernard^ anyone know who might be running that account?14:16
rosmaitai think there is only one patch where that CI actually did not fail: https://review.opendev.org/q/reviewer:goldranjith@gmail.com14:17
erlonOne of the posts have a link named vmstore...something14:17
erlonso, they must be trying to run their ci14:17
rosmaitanot very successfully14:18
erlonAnd I think the only handle we have on him would be his email14:18
funginote that the opendev sysadmins can disable a gerrit account if it's being a nuisance and the operator for it isn't responsive14:19
fungibut that's typically a last resort14:19
rosmaitai wonder whether this is the driver (it's not in-tree): https://github.com/Tintri/vmstore-cinder-driver14:19
jbernardhmm, i will send a query to that email and see if we get a response14:20
erlonthere is a "cinder/volume/drivers/vmstore/nfs.py' driver14:21
jbernardthey're welcome to run ci on proposed patches, but if the driver is out-of-tree it shouldn't be reporting on them14:21
erlonBut other than the uncommon (personal) name, I don't see that much difference from other CIs that fail most of the time14:21
rosmaitaerlon is correct, i was looking in an old branch14:22
agalica_so the CI is just failing and that's it, or is it spamming?14:23
rosmaitatwo issues: it is failing like 100%, and (2) shouldn't be using a user account14:23
jbernardand 3, testing an out-of-tree driver14:24
agalica_gotcha- yeah, the user account thing is what I complained about a few weeks ago.  Most everyone has fixed that now14:24
rosmaitajbernard: it is actually in-tree, erlon found it14:24
erlonhmm, got it, I remember theres a need to request a different account for cis14:24
jbernardahh, ok i was looking at the wrong thing, sorry14:24
*** agalica_ is now known as agalica14:25
jbernarddo we have a human vmstore contact?14:25
agalicaah, no wonder I didn't see the courtesy ping14:25
jbernardoh yeah, courtesy ping: agalica :)14:26
agalica:D14:26
rosmaitalooks like the lead developer is akhodos@ddn.com14:26
jbernardi pulled from a previous revision, i think i went too far back and missed some recent ones14:26
jbernardok14:26
jbernard#action jbernard to email vmstore (akhodos and goldranjith) about their CI, report in next meeting if not before14:27
auniyaljbernard on quring ranjit, may be on openstack-discuss mailing list instead of directly, so if Ranjit is not available to reply others from there team/org who knows the person might reply, may be in  other projects too14:27
erlonthis was the submitter: eivanov@ddn.com14:28
jbernardauniyal: good call14:28
auniyalor *@ddn14:28
erlonEugene Ivanov14:28
jbernarderlon: ill add that to the cc, thanks14:28
jbernard#topic gap bridging discussion (fungi)14:30
jbernardfungi: take it away14:30
fungibriefly revisiting this topic from last week now that folks have hopefully had time to digest it better, i've got a few more points that came out of a closer look into the data but also wanted to see if anyone had additional thoughts14:30
fungione thing i missed calling out, which is exciting, is that the typical time to response on reviews improved drastically in gazpacho compared to flamingo; twice as fast on average (mean) but the median was actually five times faster!14:30
fungii don't know if the team's reviewing habits changed in gazpacho, but that would be something to continue if so as it seems to have made a big impact, both the mean and median were significantly better than in any of the prior 5 cycles14:30
fungigranted, flamingo was anomalously bad in that regard, but gazpacho numbers were still easily twice as fast (on both median and mean) as any of the prior 5 cycles too, so something definitely improved dramatically14:30
fungii guess this could also be attributed to the small/medium patch triage dashboard that was mentioned in last week's discussion? conversely, multiple survey respondents did talk about larger patches not getting attention or being ignored14:30
fungiwe had takeaway suggestions, some similar to the previous round... contributors still aren't clear on where maintainers are tracking review priorities, so continuing to work on addressing that communication gap may help14:30
fungieven maintainers didn't necessarily agree on where and how review priorities were being tracked, which could be contributing to the confusion there14:31
fungialso the change and reviewer numbers are steadily growing while the active maintainer count is not, so if possible try to find some time to evaluate new maintainer candidates from the pool of established reviewers and code contributors14:31
fungiand on the topic of larger changes not getting as much review attention, maybe this is where the broader community can help by coming up with a set of recommendations to contributors for how to split up and make patches easier to review14:31
fungiif anybody has follow-up thoughts or questions on any of the above, as well as for the survey and metrics analysis from last week now that you've had time to mull it over, i'm happy to answer questions and entertain feedback here14:31
fungi(or in the #openstack-cinder irc channel after the meeting, or on the openstack-discuss mailing list, or even directly/in private for that matter)14:31
fungii wonder if i just broke everyone's irc clients, or merely left you all speechless14:33
jbernardsome of both, i suspect14:34
jbernard:)14:34
jbernardprocessing…14:34
harshstill her o/ :D 14:34
harshhere*14:34
fungi;)14:34
jbernardbig patches remain a challenge, there is room for improvement there certainly14:35
erlonI think that the delays are particularly bigger for driver maintainers, in new drivers submitters. For new drivers is really hard to split the code14:35
jbernardwe did just recently add 2 new cores, and have more coming up to speed14:35
fungiyes, i am getting the feeling that, as a community, maybe we should start seeing "proposing large patches" as the actual problem, rather than "failing to get around to reviewing large patches"14:35
agalicaI'm going to be doing reviews on friday, hopefully that helps14:36
agalica:D14:36
fungiexcellent news on the core team additions, congrats to you all!14:36
erlonBut I think that, if we collect the list of things that we usually require, and flag on new changes, we can delegate part of that (locally) to LLMs to assist to point those14:36
jbernardyeah, there is some /size/ where reviews take much more time and begin to collide with other commitments, it can be hard to balance/juggle everything14:36
fungiit seems like llms are making this problem worse too, based on what i'm hearing from various maintainers14:37
agalicai think that's a possibility too, but it might resolve down to people leaning on vibes rather than actually making sure the code it spits out is sane14:37
fungiand that people who insist on using an llm to create patches should be giving it explicit instructions to break the work up into logical reviewable chunks while avoiding making unrelated changes in the files they touch14:38
erlonit makes worse because people can generate much more code now :)14:38
auniyalsomewhat offtopic, I would like suggest to create a groups of interested people for particular topic , eventlet-removal, CBT if required separate syncup too - this will help fast feedback14:38
erlonso, if the reviewers don't get to new tools, we will be drawned14:38
auniyal*separate from upstream meet I meant14:38
fungisome project teams already do hold subteam meetings, yes14:39
fungiyou can see them if you skim the list at meetings.opendev.org14:39
fungiit's certainly a pattern that could be replicated14:39
jbernardfungi: for review response time improvements, do we know if it's across all reviewers, or mainly driven by one or two?14:41
erlonfungi: is there any discussions regarding to adding AGENTS.md, CLAUDE.md files into projects? That could steer the models into generating better and more concise code adhering to the projects guidelines14:41
fungii'm a little hesitant at the suggestions i keep seeing in various places to throw new-found ai tools at reviewing tasks, so far people who have done that have told me that while it helped them do more thorough reviews and maybe catch more bugs, it didn't improve their review bandwidth at all14:41
fungierlon: yes, get involved in the agentic workflows effort under the tact sig, they have a #openstack-agentic-workflows irc channel for those topics too14:41
fungibut they're collaborating on coming up with standard agents files and skills and whatnot14:42
jbernarderlon: https://review.opendev.org/c/openstack/cinder/+/99400714:42
jbernardfungi: for key takeaways for improvements,14:43
jbernard* better transparency on review priorities14:43
fungithe other concern i have with suggestions to just lean harder on ai is that so far all of the root problems we've found stemmed from insufficient communication between humans, so removing humans even farther is likely to make things worse rather than better14:43
jbernard* faster response times for larger patches14:43
fungior maybe having fewer large patches (which probably means openstack doing a better job of educating contributors)14:45
fungiand yeah, the third takeaway was growing the core team, which it sounds like is in progress for hibiscus already14:46
jbernardin progress14:46
erlonjbernard: I woke up with this in mind some time ago, may be a start: https://etherpad.opendev.org/p/cinder-priority-guidelines14:46
jbernardit's defintely a start, ordering them and being able to identify the different categories quickly would help immensely14:48
agalicaI usually dream about awesome things like the apocalypse.  It's depressing to me to know that erlon dreams of openstack14:49
fungii dream of the openstackapocalypse14:49
erlonhaha lol, I have a morning notepad, for those kind of goodies. they always come unexpectedly 14:50
agalicaI usually get mine in the shower where there isn't a place to write it down, haha14:50
fungiyou need a greaseboard in the shower where you can take notes14:51
Anoop_ShuklaA quick suggestion - Why don’t we have openstackathons?14:51
fungior one of those pads divers use14:51
erlonIt would be nice if gerrit had tags14:51
fungiAnoop_Shukla: like hackathons? we probably do and have different terms for them14:51
jbernardfungi: from your understanding, if you could me one, or two, recommendations to this group, do you think we've missed anything? priorities and reponse time; anything else?14:51
Anoop_ShuklaCan be a good way across openstack projects where contributers come together and solve problems..14:51
fungierlon: gerrit has hashtags, is that what you're looking for?14:51
jbernardfungi: s/me/make14:51
Anoop_ShuklaCool..would be good to know when they happen and participate14:52
fungijbernard: keep growing the core team if you can, because the incoming volume seems to be on a steady rise14:52
jbernardfungi: got it14:52
erlondoes it, I mean, you tag something like, 'low-prio', 'ci-blocker', then you can filter all of those in a list14:52
agalicayes14:53
jbernarderlon: i think that's how it works14:53
agalicalook at the "tags" section14:53
fungiAnoop_Shukla: some happen in the scope of local meetups for user groups, some happen at per-team midcycle events, some happen in ptg sessions (this is a lot of what the ptg was originally meant for)14:53
jbernardwe do that with eventlet-removal at lease14:53
agalicaon the left hand side around the checklist14:53
Anoop_ShuklaOkay good to know..thanks for the info @fungi14:54
agalicahmm, I'm not seeing them now, weird.  maybe it only shows up if there's actual tags14:54
erlonah, yes, topic14:54
fungiAnoop_Shukla: you could also consider cinder's semi-frequent "festival of reviews" a sort of recurring hackathon but for reviewing instead of coding14:54
eharneytopic is a different thing14:54
erlonBut that is hard to use for that kind of organization, since gerrit auto-create then, and I think you can only add one topic, and anyone can add14:55
eharneythe hashtags field is under "Show All"14:55
fungiright, we've been discouraging continued use of teh change "topic" field in gerrit for a while, because upstream gerrit has switched from treating it as informal metadata to having some configurable behaviors tied to same-topic changes14:55
agalicaahh, there it is, thanks eharney14:55
fungiwhat people previously relied on topics for is mostly better handled with gerrit hashtags now14:55
jayaanandI updated NetApp pending patches in festival-of-reviews etherpad. Need cors attention for these patches14:56
jbernard4 minutes remaining, 14:56
jbernard#topic open discussion14:56
fungii didn't mean to consume your entire meeting, sorry!14:56
fqiCould you please tell me the deadline of Cinder Spec for Hibiscus?14:57
agalicaconsumption with mushrooms usually goes the other way around14:57
JyotsnaNew HPE Alletra MP driver patch : https://review.opendev.org/c/openstack/cinder/+/969373  Thanks and appreciate consolidated feedback from <erlon>. We are in process of addressing the feedback. Soon will update the patch. We have 2 queries on patch feedback, whenever possible <erlon> can help answering them.  We have made an improvement in CI process, transitioned from zuul to Jenkins which made the CI runs smooth and easy to debug. 14:57
jbernardfungi: no worries, thanks so much for the data collection and summaries, much much appreciated14:57
harshI have added a couple of small bug fix patches for IBM driver. They have multiple +1s and reviews so would appreciate your attention in closing those small patches quickly. 14:58
fungiif cinder has a formal spec freeze you could consider including it in the release schedule like nova and manila do: https://releases.openstack.org/hibiscus/schedule.html14:59
erlonJyotsna: Ill address then later today14:59
JyotsnaThanks <erlon>15:00
erlonharsh: link?15:00
jbernardfqi: not formally defined, but soon, do you have something that needs attention?15:00
harsherlon : https://etherpad.opendev.org/p/cinder-hibiscus-reviews15:00
harshLast patches in the bugs section 15:01
erlonthanks15:03
jbernard#endmeeting15:04
opendevmeetMeeting ended Wed Jun 24 15:04:30 2026 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)15:04
opendevmeetMinutes:        https://meetings.opendev.org/meetings/cinder/2026/cinder.2026-06-24-14.02.html15:04
opendevmeetMinutes (text): https://meetings.opendev.org/meetings/cinder/2026/cinder.2026-06-24-14.02.txt15:04
opendevmeetLog:            https://meetings.opendev.org/meetings/cinder/2026/cinder.2026-06-24-14.02.log.html15:04
jbernardthanks everyone15:04
amoralejHi, I'm working in some watcher integration with cinder. I've found that while the volume_type in Volume cinder API is the volume_type name, in the notifications payload for volumes, volume_type is ID and name is not reflected. Is that expected and documented somewhere?15:04
fqijbernard: thank you. I would like to merge the spec retype volume including snapshots into Hibiscus.15:05
hemnahey sorry I missed the meeting.  busy work day.  I have a few patches I'd like to request for reviews.15:09
hemnahttps://review.opendev.org/c/openstack/cinder/+/99266715:09
hemnahttps://review.opendev.org/c/openstack/cinder/+/99187615:09
opendevreviewMerged openstack/cinder master: Fungible: use volume_utils to upload volume as image  https://review.opendev.org/c/openstack/cinder/+/98323715:24
opendevreviewMerged openstack/cinder master: Add correct mock for rados.Error  https://review.opendev.org/c/openstack/cinder/+/99186715:24
opendevreviewSaikumar Pulluri proposed openstack/cinder master: NetApp ONTAP: begin closing ZAPI-to-REST gaps  https://review.opendev.org/c/openstack/cinder/+/99336317:50
opendevreviewVinícius Fernando Rocha Lobo proposed openstack/cinder master: Fix service_uuid not set on cinder manage volumes  https://review.opendev.org/c/openstack/cinder/+/99481920:14
opendevreviewVinícius Fernando Rocha Lobo proposed openstack/cinder master: Fix service_uuid not set on cinder manage volumes  https://review.opendev.org/c/openstack/cinder/+/99481921:11

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