Friday, 2022-02-18

*** hemna5 is now known as hemna01:45
*** _pewp___ is now known as _pewp_03:17
*** hemna5 is now known as hemna03:21
opendevreviewRajat Dhasmana proposed openstack/os-brick master: Remove unused host driver
opendevreviewTushar Trambak Gite proposed openstack/cinder master: Reset state robustification for volume os-reset_status
*** dviroel|out is now known as dviroel10:54
opendevreviewTushar Trambak Gite proposed openstack/cinder master: Reset state robustification for group os-reset_status
opendevreviewAtsushi Kawai proposed openstack/cinder master: Hitachi: add GAD volume support
opendevreviewTushar Trambak Gite proposed openstack/cinder master: Reset state robustification for group-snapshot os-reset_status
raghavendrathi rosmaita: this is regarding
raghavendrati updated "Milestone target" to "zed-2"13:26
raghavendratand "Series goal" as "zed"13:26
rosmaitaoh, ok, i thought you were going to ask why I did that, and i could not remember moving that BP!13:26
raghavendratsomehow, it is not seen under
rosmaitaok, that's because the "Series goal" says "Proposed for zed"13:28
rosmaitatake a look now13:28
raghavendratnow visible. Thank you for quick help13:29
rosmaitai still don't understand how Launchpad handles blueprints, so always ask if you see something weird13:29
TusharTgiterosmaita: zuul test passed take a look
rosmaitawhat did you change?13:50
TusharTgiterosmaita: i added "ing " condition checks and i was checking same state twice in the code like available to available13:52
TusharTgitethats what cause error13:52
rosmaitaok, good job13:54
rosmaitaTusharTgite: left a comment, i think you may have a typo in one of your messages, but otherwise looks good13:57
rosmaitathough i need to look at the tests carefully on my next review13:57
rosmaitabut, looking good!13:57
TusharTgiterosmaita: is there anything else left cause it almost patchset35 i'll push a final patchset then13:58
rosmaitawell, like i said, i want to look at the UTs to check what situations you are covering, so no promises!13:59
rosmaitathat's all i saw on this run-through13:59
TusharTgiterosmaita: ok i'll wait for that and will push patch tommorow then14:01
rosmaitareminder: Cinder Festival of Reviews today, 1400-1600 UTCjoin us in meetpad: go directly to the etherpad: rosmaita smcginnis tosky whoami-rajat m5z e0ne geguileo eharney walshh_ jbernard sfernand enriquetaso hemna fabiooliveira14:01
rosmaitawell, that was a formatting fail14:01
* enriquetaso having bandwidth problems 14:03
rosmaitafestival etherpad:
jungleboyjWill join as I am able today.14:08
whoami-rajatrosmaita, I don't remember the reason but why we don't do this in Bluejeans ? :)14:12
rosmaitabecause you can't see as many people at once14:12
enriquetasoI'm not really familiar with availability zones but I thought that backups could have their own. That's why I'm not sure what to do about . If anyone is familiar with this please add a comment on the patch. 14:15
opendevreviewMerged openstack/cinder master: DOC: Add cinder modules on initial configuration
TusharTgitei need final call on this one shall i really abandon this it's a good clean up i've covered entire cinder repo.14:26
enriquetasoTusharTgite, think geguileo has a good point regarding 799631 14:32
TusharTgiteenriquetaso: ok i'll abandon it then14:37
enriquetaso:( 14:40
dswebbapologies, totally spaced on this meeting.  Is it too late to try and get this reviewed? 14:46
rosmaitadswebb: that is kind of not an XS patch, so maybe later14:48
dswebbah ok, no worries, will remove it from the meeting notes then14:48
rosmaitayou can leave it, maybe someone will follow up after the meeting14:48
rosmaitawe do want to try to power through as many XS reviews as possible14:49
jungleboyjrosmaita: Has the meeting started already?  Thought it started in 10 minutes.14:51
rosmaita1400 utc!!!14:52
jungleboyjUgh.  I logged onto the meetpad and didn't see anyone.14:53
rosmaitadswebb: "extra small" -- <=10 lines14:53
rosmaitajungleboyj: that's weird, there are 8 of us there14:53
jungleboyjLet me try again.14:54
rosmaitameetpad link above is incorrect due to formatting ... here is where we are:
rosmaita(sorry about that)14:57
jungleboyjWe hear you.14:59
* enriquetaso lost connection 15:10
opendevreviewMerged openstack/cinder-specs master: Add zed directory for specs
rosmaitahemna: you've done a lot of work with backups, can you take a look at ?15:12
hemnaok checking15:12
rosmaitait's only 1 line, but i wonder whether it could cause a regression15:12
rosmaita(but not for any good reason, i am just a worrier)15:13
opendevreviewIago Filipe proposed openstack/cinder master: Improving delete volumes exception log
hemnarosmaita, I'm wondering how the AZ could be different in this case. the original backup could be on 1 AZ and a new volume being created from that backup is in a different AZ?15:19
hemnahow could that be?15:19
opendevreviewMerged openstack/cinder master: PowerFlex documentation contains invalid paths
rosmaitahemna: Environment:15:20
rosmaita2 azs, each with a c-vol and c-bak15:20
rosmaitacinder servers at each az have their own ceph and the ceph is only accessible within the AZ15:20
rosmaitacinder backup uses swift driver and swift is globally accessible15:20
hemnaI guess I wasn't sure they'd be able to see the backup in the list being in 2 different AZ's for the project15:21
hemnacinder backup-list15:21
rosmaitahemna: yeah, take a look at the description in the bug:
hemnamaybe I'm thinking of our deployment too much.  we don't have cinder managing multiple AZs.  they are totally separate deployments15:22
rosmaitai'm slightly worried that this change may work great in his deployment, but may cause a problem elsewhere15:22
rosmaitathough the change does seem to make sense15:23
hemnaso, I guess the question is, if there are multiple AZ's in a cinder deployment, will that logic always be the case, where a backup backend always has access to both AZs?15:23
hemnaI'm not sure that's a guarantee15:25
hemnamaybe it makes sense to have a config option to allow cross AZ restores/creates from backups.  the admin would know if that same backend was available in both AZs15:26
hemnaok added a -1 on it and those comments15:28
opendevreviewOpenStack Release Bot proposed openstack/os-brick stable/yoga: Update .gitreview for stable/yoga
opendevreviewOpenStack Release Bot proposed openstack/os-brick stable/yoga: Update TOX_CONSTRAINTS_FILE for stable/yoga
opendevreviewOpenStack Release Bot proposed openstack/os-brick master: Update master for stable/yoga
rosmaitahemna: thank you15:33
rosmaitalooks like yoga os-brick has been released!15:33
jungleboyjNeed to jump to another meeting.  Good to see you guys.15:34
rosmaitajungleboyj: thanks for attending, have a good weekend15:35
jungleboyjThanks.  You too!15:36
whoami-rajatfabiooliveira, is the bug report for a local run or a gate job?
hemnaso I'm running into another backup issue.   A volume exists on a pool that has limited space leftover and a user wants to do a backup.  The volume is attached, so that requires a snapshot of the volume being made and on our backend that results in a clone effectively.   Since there isn't enough space on that pool to hold the clone, the backup fails15:41
hemna100% of the time.15:41
hemnathat's not very 'cloudy'15:41
fabiooliveirawhoami-rajat: from a local run15:41
hemnaif I migrate the volume, it could end up on another pool that has the same problem.15:42
hemnaI have a similar problem with extend volume too.15:42
whoami-rajatfabiooliveira, can you share the final part of the stack trace (omitted in the bug report) if handy15:42
hemnacustomers just want the actions to work.  they don't care where the volumes live.15:42
hemnakinda need an option for some of these actions that auto migrate to pools that have space for the actions to work.15:45
hemnaand migrations can take a while, so it needs to be async.15:45
hemnaI also think cinder needs transaction tracking for actions on volumes as a paper trail.15:46
hemnawe have so many inconsistent volumes from attaches, migrations it's overwhelming at this point.15:46
rosmaitawell, if we're going to start auto migrating stuff, we definitely need some kind of transaction trail15:47
rosmaitabut if an operator runs pools real close to capacity, they have to expect there will be some problems15:48
hemnafor sure.   if there isn't space, there isn't space15:49
hemnabut the scheduler would know that.15:49
hemnain our case we are always on the edge, but when we reach low capacity, we add more storage15:49
whoami-rajatI will be leaving now, happy weekend everyone!15:49
hemnaif the auto migrate was available, then it would just keep things working and re-balancing basically15:49
rosmaitabye Rajat!15:50
fabiooliveirawhoami-rajat: don't have it, right now, but I can provide it soon15:50
fabiooliveirawhoami-rajat: bye, have a nice weekend15:50
whoami-rajatfabiooliveira, ack, i will remove my -1 as i see the shuffling between the base and Iscsi driver __getattr__ methods but just want to confirm it15:51
opendevreviewTushar Trambak Gite proposed openstack/cinder master: Deleting a volume in 'downloading' state
hemnaextend volume already goes through the scheduler15:59
*** dviroel is now known as dviroel|lunch16:29
zigoWhat new feature of PBR is os-brick using?16:32
zigoI mean, it now declares pbr >= 5.8.0, I'd like to know why.16:32
zigoSame for all the 12 other updated depends...16:33
rosmaitazigo: doesn't matter, we've been testing with the new versions only for the number of months specified in the commit message16:33
rosmaitai mean, we've only been testing with the new versions for the number of months specified16:34
zigorosmaita: It does mater, requiring a version which is not already in Xena is very bad practice and is very annoying for in production deployments upgrades.16:35
zigoIf that's what's going to happen, then we MUST re-introduce lower-constraints testing quickly.16:35
zigoI'm opening a list thread right away...16:37
rosmaitaok, that is a good idea16:37
opendevreviewMerged openstack/cinder master: zuul: configure irrelevant-files for the rally job
opendevreviewMerged openstack/cinder master: Remove unused variable
opendevreviewMerged openstack/cinder master: PowerMax docs -  Include a note in cinder backup section
sreerammounikaHi reviewers, Requesting you to review the commit
opendevreviewMerged openstack/cinderlib master: setup.cfg: Replace dashes by underscores
opendevreviewMerged openstack/os-brick master: Update master for stable/yoga
opendevreviewMerged openstack/os-brick stable/yoga: Update .gitreview for stable/yoga
opendevreviewMerged openstack/os-brick stable/yoga: Update TOX_CONSTRAINTS_FILE for stable/yoga
*** dviroel|lunch is now known as dviroel17:32
opendevreviewLucas Morais de oliveira proposed openstack/cinder master: WIP api-ref: Add docs for clusters
*** dviroel is now known as dviroel|afk20:09

Generated by 2.17.3 by Marius Gedminas - find it at!