15:01:38 <enriquetaso> #startmeeting cinder_bs
15:01:38 <opendevmeet> Meeting started Wed Aug 18 15:01:38 2021 UTC and is due to finish in 60 minutes.  The chair is enriquetaso. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:01:38 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:01:38 <opendevmeet> The meeting name has been set to 'cinder_bs'
15:01:46 <enriquetaso> #link https://etherpad.opendev.org/p/cinder-bug-squad-meeting
15:01:55 <enriquetaso> Hello, welcome to the bug meeting, we have a bunch of bugs for this week. Full list of bugs:
15:01:55 <enriquetaso> #link http://lists.openstack.org/pipermail/openstack-discuss/2021-August/024220.html
15:02:07 <enriquetaso> The way I used to request the new bugs to Launchpad is failing or something so I lost track of most of the new reported bugs. For this week's report I manually checked the launchpad. If anyone knows how to filter dates on Launchpad that actually works please let me know.
15:02:32 <enriquetaso> OK.. the first two bugs are High importance bugs and that's why I've mentioned them here but both have patches proposed to master.
15:03:02 <enriquetaso> #topic bug_1 "Original volume size is lost when uploading a volume to a qcow2 image"
15:03:07 <enriquetaso> #link https://bugs.launchpad.net/cinder/+bug/1939972
15:03:13 <enriquetaso> Summary: The virtual size of the image(2GB) doesn't fit in a volume size(1GB), but this error is not exposed to users. This makes it difficult to let users know the actual cause of the error.
15:03:13 <whoami-rajat> enriquetaso, i think you can Order by: Date last updated
15:03:51 <enriquetaso> whoami-rajat, yep, i need to double check again I guess, because it wasnt working for me today
15:04:02 <whoami-rajat> ok
15:04:16 <enriquetaso> Also, regarding the bug_1 Rajat proposed a new user message that will notify users regarding any async failure with the given operation. In addition, the reporter wrote that cinder already has a mechanism to validate a given volume size based on the min_disk property of a source image.
15:05:02 <whoami-rajat> yeah so i will be working on the user message part for this bug (as I've a series of patches for user-messages and can add this in one of them)
15:05:31 <enriquetaso> whoami-rajat++
15:05:32 <enriquetaso> thanks
15:05:36 <enriquetaso> moving on..
15:05:39 <rosmaita> i guess cinder could set virtual_size for glance, and/or min_disk when we upload
15:06:12 <enriquetaso> The patch proposed upstream uses min_disk https://review.opendev.org/c/openstack/cinder/+/804584
15:06:16 <enriquetaso> #link https://review.opendev.org/c/openstack/cinder/+/804584
15:06:21 <rosmaita> min_disk would be better, we check that
15:06:28 <rosmaita> thanks for the link
15:07:15 <rosmaita> that sounds like a good solution
15:07:17 <whoami-rajat> glance has code inplace to set virtual_size after creating an image, maybe they can add that to this flow as well
15:08:47 <enriquetaso> do you mind doing that whoami-rajat ? I'm not sure what should glance do
15:09:25 <enriquetaso> OK, i can check that later, moving on..
15:09:31 <enriquetaso> #topic bug_2 " Nova thinks deleted volume is still attached"
15:09:37 <enriquetaso> #link https://bugs.launchpad.net/cinder/+bug/1937084
15:09:43 <enriquetaso> Summary: There are cases where a cinder volume no longer exists yet nova still thinks it is attached to an instance and we cannot detach it anymore. This has been observed when running cinder-csi, where it makes a volume delete request as soon as the volume status says it's available.
15:10:07 <enriquetaso> Fix proposed to branch: master
15:10:08 <enriquetaso> #link https://review.opendev.org/c/openstack/cinder/+/801913
15:10:46 <enriquetaso> Please review when you have time, nothing else to add there from my side and I think gorka is on holiday
15:11:43 <rosmaita> and it's #4 in a series of patches!
15:11:45 <eharney> the whole series of 4 patches is stuck on a merge conflict on the first patch that needs resolution
15:12:41 <opendevreview> Eric Harney proposed openstack/cinder master: Fix detach notification  https://review.opendev.org/c/openstack/cinder/+/778534
15:12:43 <enriquetaso> oh true, sorry. forgot to mention that
15:12:59 <rosmaita> i think eric just resolved the conflict
15:13:01 <eharney> rebased ^
15:13:10 <enriquetaso> eharney++
15:13:20 <opendevreview> Eric Harney proposed openstack/cinder master: Delete attachment on remove_export failure  https://review.opendev.org/c/openstack/cinder/+/801912
15:13:38 <opendevreview> Eric Harney proposed openstack/cinder master: Expose volume_attachments in Volume OVO  https://review.opendev.org/c/openstack/cinder/+/779576
15:13:50 <enriquetaso> #topic bug_3 " Create instance from volume fails on reschedule "
15:13:56 <enriquetaso> #link https://bugs.launchpad.net/cinder/+bug/1936474
15:14:01 <enriquetaso> whoami-rajat sorry for bothering you here but I think it's easier, are you facing this problem with rbd for both glance and cinder or it's something different?
15:14:36 <whoami-rajat> enriquetaso, it's backend independent, I've also replied on the bug
15:14:49 <enriquetaso> thanks!
15:14:58 <enriquetaso> Last 2 bugs:
15:15:07 <eharney> should be moved back to "in progress" from "incomplete"
15:15:30 <enriquetaso> done
15:15:39 <enriquetaso> #topic bug_4 "Removing a volume in wrong state doesn't report its state"
15:15:47 <enriquetaso> #link https://bugs.launchpad.net/cinder/+bug/1940437
15:15:52 <enriquetaso> Summary: When trying to remove a volume in a wrong state, the error message says about what's the states should be for the volume to be removable; however we don't report its current state. It would be useful for operators to know about it.
15:16:00 <enriquetaso> I thought that we already covered this somewhere, but it looks like not. Feel free to reproduce this and review Emilien's patch.
15:16:18 <eharney> i think the reason we don't do this is because the logged status could be wrong due to race conditions
15:16:22 <eharney> i remember looking at it a while back
15:17:00 <eharney> to do it completely accurately you need conditional_update() to report the relevant values back, not query them again later
15:18:26 <enriquetaso> yes.. sounds a bit tricky
15:18:56 <enriquetaso> Fix proposed on master
15:18:59 <enriquetaso> #link https://review.opendev.org/c/openstack/cinder/+/805027
15:19:42 <enriquetaso> guess this fix isn't completely accurate then
15:19:59 <eharney> i'm not sure, need to take a closer look again
15:20:18 <enriquetaso> sure
15:20:24 <enriquetaso> Last one, we've already discussed this in the main meeting but a little reminder.
15:20:30 <enriquetaso> #topic bug_5 "[LVM] lvextend command crashes with code 139"
15:20:35 <enriquetaso> #link https://bugs.launchpad.net/cinder/+bug/1940436
15:20:42 <enriquetaso> #action(enriquetaso) take a look to os-brick
15:20:49 <enriquetaso> The rest of the bugs are driver related sort of bugs and already have patches proposed upstream so feel free to check the list and review them :)
15:20:55 <enriquetaso> #topic open discussion
15:20:59 <enriquetaso> Do we have any bugs for open discussion?
15:24:01 <rosmaita> nothing from me ... thanks for organizing this, sofia
15:24:07 <enriquetaso> I guess we don't have any bugs for the open discussion then :P
15:24:11 <enriquetaso> thank you brian
15:24:23 <enriquetaso> #endmeeting