14:00:16 <rosmaita> #startmeeting cinder
14:00:17 <openstack> Meeting started Wed Nov 11 14:00:16 2020 UTC and is due to finish in 60 minutes.  The chair is rosmaita. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:00:18 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:00:21 <openstack> The meeting name has been set to 'cinder'
14:00:29 <rosmaita> #topic roll call
14:00:50 <whoami-rajat__> Hi
14:00:58 <michael-mcaleer> Hi everyone
14:01:07 <TusharTgite> hi
14:01:09 <tosky> hi
14:01:11 <lseki> o/
14:01:21 <e0ne> hi
14:01:28 <geguileo> hi! o/
14:01:45 <rosmaita> looks like good turnout!
14:01:55 <eharney> hi
14:02:02 <jungleboyj> o/
14:02:17 <zoharm> hi!
14:02:22 <rosmaita> #link https://etherpad.openstack.org/p/cinder-wallaby-meetings
14:02:31 <rosmaita> ok, let's get started
14:02:36 <rosmaita> #topic announcements
14:02:49 <rosmaita> first thing: whoami-rajat has volunteered to be the release manager
14:03:00 <rosmaita> we will hear from him a bit later
14:03:32 <rosmaita> there are still roles available for people who want to get more involved in cinder
14:03:45 <rosmaita> see the Wallaby PTG summary on the wiki for details
14:05:36 <rosmaita> second item:
14:05:46 <rosmaita> "consistent and secure" policy testing strategy still under discussion
14:05:56 <rosmaita> #link http://lists.openstack.org/pipermail/openstack-discuss/2020-November/018636.html
14:06:11 <rosmaita> there's a link to the etherpad from yesterday's discussion in that email
14:06:17 <rosmaita> whoami-rajat__ and i both attended
14:06:29 <whoami-rajat__> also tosky  attended
14:06:42 <rosmaita> sorry ... how can i forget tosky ?
14:06:49 <rosmaita> tosky was there too
14:06:51 <tosky> it's totally possible
14:07:10 <rosmaita> tosky: you are ubiquitous
14:07:38 <rosmaita> the issue has to do whether full-on tempest api tests using the roles & scopes for the nine personas
14:07:48 <rosmaita> or some kind of unit or functional testing
14:08:04 <rosmaita> the problem is that we want the responses from keystone to be realistic
14:08:23 <rosmaita> on the other hand, we don't want to do a bunch of rechecks for extraneous failures
14:08:42 <rosmaita> so if you have suggestions or thoughts, please put them on the etherpad or attend the next meeting on monday
14:08:52 <rosmaita> (info in the email linked earlier)
14:09:26 <rosmaita> final announcement is that as jungleboyj reported at the PTG, the TC is taking a "stance" on the openstackclient
14:09:36 <rosmaita> #link https://review.opendev.org/#/c/759904/
14:09:48 <jungleboyj> :-)
14:09:56 <rosmaita> has not yet been approved, so leave comments if you have opinions
14:10:11 <rosmaita> there's no timeline on the proposal, it's just aspirational at this point
14:10:35 <rosmaita> i guess one more thing
14:10:58 <jungleboyj> Right.  Just a note on that.
14:11:09 <jungleboyj> We know that it is going to be a heavy lift.
14:11:32 <jungleboyj> We aren't mandating things at the moment but we are making note that we need to start heading in that direction.
14:11:41 <jungleboyj> Everyone has been warned.  ;-)
14:13:04 <rosmaita> ok, i still think they need to fix the 'openstack help' feature to make it usable, but that's just me
14:13:09 <eharney> how's the microversion stuff for osc going?
14:13:16 <rosmaita> one last point about the community goals for wallaby: the support-yaml-only-for-policy-files proposal is still under vote, so no action on that goal yet
14:13:31 <rosmaita> eharney: "it works for nova"
14:13:37 <eharney> ok, same as a few years ago then
14:13:43 <jungleboyj> eharney:  Well, they say it is working.  If it doesn't have what we need, then we need to let them know.
14:13:53 <jungleboyj> eharney:  Yes.  Same old same old.
14:14:47 <rosmaita> i have on my to-do list to  check the mv support and then talk with someone on the osc team, and also tell them about python-brick-cinderclient-ext
14:14:56 <rosmaita> namely, that it needs to be supported
14:15:12 <rosmaita> that's all from me -- anyone have anything to announce?
14:15:18 <jungleboyj> rosmaita:  Ok.  So that is a good starting point.
14:17:04 <rosmaita> #topic Bug Review
14:17:11 <michael-mcaleer> thanks rosmaita
14:17:15 <rosmaita> #link https://etherpad.opendev.org/p/cinder-wallaby-r23-bug-overview
14:17:21 <michael-mcaleer> Wallaby R-23 bug review #link https://etherpad.opendev.org/p/cinder-wallaby-r23-bug-overview
14:17:40 <michael-mcaleer> 4 bugs opened this week, 3 in cinder and 1 in os-brick
14:17:46 <michael-mcaleer> cinder first
14:17:53 <michael-mcaleer> cinder-manage traceback on missing action https://bugs.launchpad.net/cinder/+bug/1902852
14:17:54 <openstack> Launchpad bug 1902852 in Cinder "cinder-manage traceback on missing action" [Undecided,In progress] - Assigned to Gorka Eguileor (gorka)
14:18:01 <michael-mcaleer> currently in progress
14:18:08 <michael-mcaleer> Running the cinder-manage command with a category but without an action will throw a Traceback.
14:18:13 <michael-mcaleer> $ cinder-manage backup
14:18:21 <michael-mcaleer> 2020-11-04 10:44:49.541 9465 ERROR cinder oslo_config.cfg.NoSuchOptError: no such option action_fn in group [DEFAULT]
14:18:42 <rosmaita> that's pretty ugly
14:19:06 <michael-mcaleer> any comments or happy to move onto second bug?
14:19:25 <whoami-rajat__> there's already a fix upstream so looks good
14:19:45 <michael-mcaleer> agree, I had a look at the fix on Monday
14:20:02 <michael-mcaleer> second bug, NFS: Failed to extend volume after blockRebase
14:20:02 <michael-mcaleer> #link https://bugs.launchpad.net/cinder/+bug/1903319
14:20:04 <openstack> Launchpad bug 1903319 in Cinder "nfs: failed to extend volume after blockRebase" [Undecided,Triaged] - Assigned to Rajat Dhasmana (whoami-rajat)
14:20:20 <michael-mcaleer> It fails with the following error in cinder-volume:
14:20:25 <michael-mcaleer> ERROR cinder.volume.manager [req-5f939fc7-f4a5-4f6a-9b70-6dcb52ee7f2e req-a2702c72-fc40-41df-a600-903a5293b898 admin None] Extend volume failed.: oslo_concurrency.processutils.ProcessExecutionError: Unexpected error while running command.Command: sudo cinder-rootwrap /etc/cinder/rootwrap.conf qemu-img resize /opt/stack/data/cinder/mnt/896fb15da6036b68a917322e72ebfe57/volume-2a351c27-abb0-4965-a16e-5378bd2c2afe 2G
14:20:26 <michael-mcaleer> Exit code: 1
14:20:26 <michael-mcaleer> Stdout: ''
14:20:26 <michael-mcaleer> Stderr: "qemu-img: Could not open '/opt/stack/data/cinder/mnt/896fb15da6036b68a917322e72ebfe57/volume-2a351c27-abb0-4965-a16e-5378bd2c2afe': Could not open '/opt/stack/data/cinder/mnt/896fb15da6036b68a917322e72ebfe57/volume-2a351c27-abb0-4965-a16e-5378bd2c2afe': No such file or directory\n"
14:20:44 <michael-mcaleer> Currently only triaged
14:20:55 <whoami-rajat__> I'm currently working on a fix for this, but this might be affecting more than just the extend operation
14:21:16 <whoami-rajat__> lseki: also this might be affecting netapp nfs driver extend operation as well
14:21:31 <lseki> I'll try to reproduce it
14:22:14 <michael-mcaleer> any further comments or happy to move onto third and last cinder bug?
14:22:31 <rosmaita> keep on moving!
14:22:41 <michael-mcaleer> VNX: LUN allocation owner issue #link https://bugs.launchpad.net/cinder/+bug/1903500
14:22:42 <openstack> Launchpad bug 1903500 in Cinder "VNX: LUN allocation owner issue" [Undecided,In progress] - Assigned to Yong Huang (yonghuang)
14:22:52 <michael-mcaleer> driver specific bug here for DellEMC VNX
14:22:59 <michael-mcaleer> currently in progress
14:23:14 <michael-mcaleer> ill move onto os-brick now as thats well covered by driver owner
14:23:27 <michael-mcaleer> NVMe not flushing on disconnect #link https://bugs.launchpad.net/os-brick/+bug/1903032
14:23:27 <openstack> Launchpad bug 1903032 in os-brick "NVMe not flushing on disconnect" [Undecided,New]
14:23:28 <rosmaita> very nice that these all have people working on them already
14:23:45 <rosmaita> that seems like it should be high priority
14:23:59 <michael-mcaleer> yeah I was going to discuss this at the end about priority/importance
14:24:12 <michael-mcaleer> good example here
14:24:13 <rosmaita> sorry, did not mean to get ahead of you
14:24:16 <jungleboyj> I think that bug came out of discussion at the PTG.
14:24:22 <michael-mcaleer> it did yeah
14:24:41 <jungleboyj> Wasn't someone from the new Vendor that joined looking at that?
14:24:42 <geguileo> yeah, I found that one while preparing for the PTG
14:24:46 <michael-mcaleer> I also cannot set the state to traiged
14:25:06 <michael-mcaleer> jungleboyj: yes he was discussing this
14:25:13 <rosmaita> oh, maybe your cinder permissions don't apply to os-brick space in launchpad
14:25:14 <michael-mcaleer> we were watching it in the powermax team also
14:25:29 <rosmaita> i will look into that after the meeting
14:25:31 <michael-mcaleer> rosmaita: I can change it to anything apart from triaged or wont fix oddly
14:25:42 <whoami-rajat__> I'm also not able to set it to triaged
14:26:28 <michael-mcaleer> we can sort that out later, might impact a number of people
14:26:52 <michael-mcaleer> thats all the bugs that came in this week, housekeeping only has one point and that was on bug importance/priority
14:27:07 <rosmaita> #action rosmaita look into bug permissions for cinder-related projects in launchpad
14:27:10 <michael-mcaleer> should bug supervisor set this or should we reach consensus each week on the new bugs and their importance at these meetings
14:27:33 <rosmaita> i think set the ones you'
14:27:43 <rosmaita> re sure about and we can discuss ones you are not sure about
14:27:49 <michael-mcaleer> sounds good to me
14:27:55 <rosmaita> thank you!
14:27:56 <jungleboyj> ++
14:27:57 <michael-mcaleer> thats all I have rosmaita
14:28:08 <rosmaita> ok, next up is releases
14:28:13 <rosmaita> #topic stable releases
14:28:19 <whoami-rajat__> Hi
14:28:50 <whoami-rajat__> as you can see we've planned a release on Nov 15 which doesn't seem possible as there are a lot of unmerged patches in all branches
14:28:53 <whoami-rajat__> #link https://launchpad.net/cinder/+series
14:29:14 <whoami-rajat__> the stable releases are for victoria, ussuri and train
14:29:53 <whoami-rajat__> I've mentioned the links to branches where we have open patches, other branches doesn't have release specific patches (like one in cinderlib train related to pep8 gate fix)
14:30:08 <whoami-rajat__> also I've prepared a summary for all open patches
14:30:11 <whoami-rajat__> #link https://etherpad.opendev.org/p/stable-releases-review-tracker
14:30:25 <whoami-rajat__> It's still not complete for train as I haven't looked at all patches there
14:31:08 <whoami-rajat__> My current query is, when can we plan a stable release and then merge patches accordingly
14:31:39 <whoami-rajat__> (Also there are 2 feature backports in ussuri that we can abandon)
14:31:42 <rosmaita> let's aim for 26 November
14:32:07 <whoami-rajat__> ok, works for me
14:32:59 <rosmaita> ok, so stable-maint-cores, please look at rajat's etherpad at your convenience
14:33:11 <jungleboyj> Can do.
14:33:29 <rosmaita> and if you see any bad backports, please note them on the etherpad
14:33:36 <whoami-rajat__> I will complete the train patches list today EOD
14:33:38 <rosmaita> "bad" == not appropriate to backport
14:33:53 <rosmaita> whoami-rajat__: thank you for your work on this
14:34:01 <whoami-rajat__> sure, np
14:34:17 <rosmaita> ok, on to the main part of our show
14:34:30 <rosmaita> #topic removing Block Storage API v2 support from python-cinderclient
14:34:40 <rosmaita> #link https://etherpad.opendev.org/p/python-cinderclient-v2-support-removal
14:34:47 <rosmaita> we had an initial discussion last week
14:35:00 <rosmaita> decided to put an announcement to the ML for comments
14:35:11 <rosmaita> #link http://lists.openstack.org/pipermail/openstack-discuss/2020-November/018531.html
14:35:22 <rosmaita> and there were no comments ... so looks like it's up to us
14:35:56 <rosmaita> the contenders are the "quick" and "one more" strategies
14:36:22 <rosmaita> any discussion, or have people thought about this and just want to vote?
14:37:19 <rosmaita> (i will wait a few minutes while people look at the options again in the etherpad)
14:40:03 <rosmaita> ok, i cannot keep my mouth shut ... my thought is that ordinarily i would not be in favor of "quick", but since you can pin your v3 requests to use mv 3.0, you get back the v2 responses and won't have to change scripts immediately
14:40:16 <rosmaita> so i am leaning toward "quick"
14:41:13 <rosmaita> either i am disconnected or nobody cares about this issue
14:41:46 <jungleboyj> I don't have a strong feeling.  Quick is fine if we haven't gotten any response.
14:42:00 <rosmaita> jungleboyj: ty!
14:42:07 <rosmaita> ok, let's vote
14:42:47 <rosmaita> #startvote python-cinderclient v2 support removal strategy? "quick", "one more", "don't care"
14:42:48 <openstack> Begin voting on: python-cinderclient v2 support removal strategy? Valid vote options are , quick, one, more, don, t, care, .
14:42:49 <openstack> Vote using '#vote OPTION'. Only your last vote counts.
14:43:06 <rosmaita> apparently the quotation marks don't work
14:43:17 <jungleboyj> #vote quick
14:43:23 <rosmaita> #vote quick
14:43:26 <whoami-rajat__> #vote quick
14:43:41 <eharney> #vote quick
14:43:43 <lseki> #vote quick
14:44:38 <rosmaita> geguileo: e0ne: hemna: tosky: ^^ any opinions?
14:44:56 <geguileo> #vote quick
14:44:58 <tosky> no strong opinion (not sure whether to choose "don", "t" or "care"
14:45:10 <rosmaita> lots of options there!
14:45:23 <e0ne> #vote one more
14:45:24 <openstack> e0ne: one more is not a valid option. Valid options are , quick, one, more, don, t, care, .
14:45:30 <e0ne> #vote "one more"
14:45:31 <openstack> e0ne: "one more" is not a valid option. Valid options are , quick, one, more, don, t, care, .
14:45:37 <e0ne> hm...
14:45:38 <rosmaita> e0ne: i think just "one"
14:45:45 <rosmaita> we will know what you mean
14:45:52 <e0ne> :)
14:45:55 <e0ne> #vote one
14:45:59 <tosky> e0ne: so, do you have a specific reason for that?
14:46:40 <e0ne> just to give operators more time in case if they are uses the same client for clouds with different versions
14:47:01 <rosmaita> it's a fair point
14:47:18 <rosmaita> i think they will be using the victoria client for a while, which has v2 support
14:47:19 <e0ne> I know, v3 is pretty old, but I know that  there are people who uses v2 only :(
14:47:30 <e0ne> rosmaita: +1
14:47:40 <e0ne> my vote won't win for sure
14:48:08 <rosmaita> e0ne: about the "just use mv 3.0" argument ... how plausible is that?
14:48:13 <e0ne> I'm just trying to be fair
14:48:32 <rosmaita> what i mean is, it's easy to say, in practice will it be easy?
14:49:17 <e0ne> rosmaita: it should be
14:49:33 <rosmaita> ok
14:49:38 <rosmaita> any last votes?
14:50:00 <rosmaita> #endvote
14:50:01 <openstack> Voted on "python-cinderclient v2 support removal strategy?" Results are
14:50:02 <openstack> quick (6): eharney, jungleboyj, lseki, whoami-rajat__, geguileo, rosmaita
14:50:03 <openstack> one (1): e0ne
14:50:29 <rosmaita> ok, i will send an announcement to the ML and see if there are any last minute complaints
14:50:40 <rosmaita> otherwise, looks like "quick" is the way to go
14:50:56 <rosmaita> #topic Restricting the versions of Ceph that Cinder supports
14:50:58 <jungleboyj> Sounds good.
14:51:07 <e0ne> at least, it will allow us to maintain less code :)
14:51:16 <rosmaita> we discussed this at the PTG, but in writing it up, i found a problem
14:51:26 <rosmaita> #link https://etherpad.opendev.org/p/cinder-restrict-ceph-version
14:51:41 <rosmaita> it has to do with the way the ceph and openstack release cycles are out of alignment
14:52:17 <rosmaita> at the PTG we said we would support Ceph project releases Plus one
14:52:34 <rosmaita> ceph will release in March, just before us
14:52:42 <rosmaita> which will put Mimic out of support
14:52:55 <rosmaita> from our survey, some people are just now upgrading to Mimic
14:53:17 <rosmaita> so my question is, should we say: we support Ceph project releases Plus two ?
14:53:21 <eharney> are we counting from when cinder releases (the end of our cycle)?
14:53:40 <rosmaita> yes
14:54:03 <rosmaita> so when we release W, ceph will only be supporting O and P
14:54:41 <eharney> then +2 sounds pretty ok to me, i think i was envisioning the +1 model counting from the beginning of our cycle before
14:54:50 <rosmaita> me too
14:55:04 <rosmaita> but i think we need to count from release time
14:55:55 <eharney> makes sense
14:56:16 <rosmaita> i have an action item from the PTG to write this up for the docs, i will go ahead and go with Ceph+2 and we can discuss further on the patch
14:56:23 <rosmaita> thanks!
14:56:26 <jungleboyj> ++
14:56:39 <rosmaita> ok, i will move the next topic to next week
14:56:44 <rosmaita> #topic Open Discussion
14:57:20 <rosmaita> anyone?
14:58:08 <rosmaita> btw, the Victoria Community Meeting is tomorrow
14:58:41 <rosmaita> 16:00 UTC
14:58:54 <rosmaita> though i realize that i have no connection information
14:59:11 <jungleboyj> Ugh.  How many meetings can I have at one time?  :-)
14:59:51 <rosmaita> well, that's all folks! have a good and productive rest of the week
15:00:01 <rosmaita> #endmeeting