14:00:16 <rosmaita> #startmeeting cinder
14:00:17 <openstack> Meeting started Wed Apr 14 14:00:16 2021 UTC and is due to finish in 60 minutes.  The chair is rosmaita. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:00:19 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:00:21 <openstack> The meeting name has been set to 'cinder'
14:00:22 <sfernand> hi
14:00:24 <jungleboyj> \o/
14:00:26 <enriquetaso> hi
14:00:28 <walshh_> hi
14:01:02 <e0ne> hi
14:01:14 <eharney> hi
14:01:49 <hemna> hi
14:02:02 <whoami-rajat> Hi
14:02:36 <rosmaita> hemna: did you want to go over the backup change you're working on at the PTG?
14:02:43 <hemna> sure
14:02:54 <rosmaita> we have some time on friday, i'll add it
14:03:01 <hemna> ok sounds good
14:03:06 <rosmaita> hello everyone
14:03:08 <rosmaita> let's get started
14:03:15 <rosmaita> #topic announcements
14:03:27 <rosmaita> #link https://etherpad.opendev.org/p/cinder-xena-meetings
14:03:36 <rosmaita> that's the new agenda etherpad ^^
14:04:11 <rosmaita> if you like getting the courtesy reminder in the cinder channel before each meeting, make sure your nick is listed at lines 30-32
14:04:27 <rosmaita> and if you don't like it, make sure your name is *not* listed
14:04:43 <rosmaita> next item: friday is the 3rd friday of April
14:04:52 <rosmaita> so time for another festival of XS reviews
14:05:05 <rosmaita> it will be 1400-1600 UTC
14:05:15 <rosmaita> and we are holding it in meetpad
14:05:28 <rosmaita> #link https://meetpad.opendev.org/cinder-festival-of-reviews
14:06:08 <jungleboyj> Cool.
14:06:12 <rosmaita> not sure if you are aware, but there is a Special Technical Committee Election happening
14:06:22 <rosmaita> it closes tomorrow at 23:45 UTC
14:06:33 <hemna> special ?
14:06:38 <jungleboyj> ++ Please vote!
14:06:44 <rosmaita> someone resigned or something
14:06:49 <rosmaita> jungleboyj: do you know?
14:06:50 <e0ne> rosmaita: what does 'special' mean?
14:06:51 <jungleboyj> We were short one member in the election.
14:07:09 <rosmaita> anyway, i heard about it and realized that i never got a ballot
14:07:11 <jungleboyj> Then had a candidate that realized they missed the deadline.
14:07:16 <rosmaita> it was stuck in my spam filter
14:07:27 * hemna looks i spam folder
14:07:32 <rosmaita> so, look for email with subject "Poll: April 2021 Special Technical Committee Election"
14:07:35 <jungleboyj> So, we approved a special election.
14:07:43 <jungleboyj> rosmaita: Thank you.  I need to go find that.
14:07:52 <rosmaita> it may have been flagged because different sent from and reply-to in the email
14:08:12 <rosmaita> it's the usual election deal, you can't share the link in the email
14:08:20 <rosmaita> otherwise i would just post it
14:08:58 <rosmaita> if you can't find it, put a shout in #openstack-tc and maybe someone there will help you
14:09:22 <rosmaita> ok, last announcement
14:09:30 <rosmaita> PTG next week!
14:09:37 <rosmaita> make sure you register
14:09:48 <rosmaita> #link https://april2021-ptg.eventbrite.com/
14:10:11 <rosmaita> and i went through the planning etherpad and came up with a schedule
14:10:31 <rosmaita> it's a rough schedule, we'll do the usual start at the beginning and give each thing as much time as it needs
14:10:43 <rosmaita> #link https://etherpad.opendev.org/p/apr2021-ptg-cinder
14:11:02 <rosmaita> if we run out of time (or have too much), i will move my topics around
14:11:15 <rosmaita> so if you have a topic, you at least know what day it will be discussed
14:12:06 <hemna> is the link for cinder-festival-of-reviews working for others?   it's not working for me (DNS ?)
14:12:21 <rosmaita> e0ne: don't forget that you were going to find a rally test example we can use for quotas testing
14:12:44 <rosmaita> we will talk aobut that on wednesday ... is that ok for you?
14:12:53 <e0ne> rosmaita: yes, thanks
14:13:08 <rosmaita> hemna: it's working for me
14:13:09 <e0ne> hemna: works for me
14:13:21 <hemna> ok thanks, must be my local network
14:13:46 <rosmaita> hemna: worst case, we will also be in #openstack-cinder during the festival, so can communicate that way
14:14:22 <enriquetaso> ++
14:14:36 <rosmaita> last thing about the PTG ... look at the schedule and let me know as soon as possible if there's a conflict for your topic
14:14:52 <rosmaita> otherwise, feel free to enhance the etherpad with links to what you will be discussing, etc
14:15:10 <rosmaita> and, as usual, we will be taking notes in that etherpad as the discussions happen
14:16:09 <rosmaita> last thing: no cinder meeting next week because we'll all be at the ptg
14:16:09 <whoami-rajat> rosmaita: can we still propose topics for PTG?
14:16:45 <rosmaita> whoami-rajat: yes, i think we have some room for Friday
14:17:10 <rosmaita> and i am going to talk about my requirements topic today, since the meeting agenda was so light
14:17:43 <whoami-rajat> ack thanks
14:17:55 <enriquetaso> no cinder bug meeting neither  :P
14:18:03 <rosmaita> good point!
14:18:36 <rosmaita> we'll be using bluejeans again for the PTG
14:18:50 <rosmaita> it seems to work well for everyone except when it crashes my laptop
14:19:05 <rosmaita> but i have a new set of fans, and that seems to have helped
14:19:26 <rosmaita> that's all the announcements i have
14:19:29 <rosmaita> anyone else?
14:20:21 <rosmaita> ok, moving along then
14:20:28 <rosmaita> #topic backportable?
14:20:39 <rosmaita> #link https://review.opendev.org/c/openstack/cinder/+/782917
14:21:16 <rosmaita> this is adding fast-diff support to the ceph backup driver
14:21:26 <rosmaita> the objection has come up that it's kind of feature-y
14:21:54 <rosmaita> on the other hand, fast-diff has been around for a long time on the ceph side, so it's kind of a bug that we don't support it when it's available
14:22:08 <rosmaita> and, it's a backup driver, not a volume driver
14:22:34 <rosmaita> so, i can see this going either way, and want to get a sense of the team whether this is an appropriate backport or not
14:22:40 <e0ne> I'm OK to backport it. it looks like a good improvement rather than new feature
14:23:31 <jungleboyj> I was worried about it being featury.
14:23:35 <eharney> seems low risk and useful to me, and isn't a user-visible feature, so i'd be ok with it
14:23:40 <jungleboyj> But am open to others input.
14:25:37 <rosmaita> i'm on the pro-backport side, it's not configured, it just checks whether the backend supports it and does the right thing
14:25:58 <jungleboyj> Ok.  Then I won't push back.  :-)
14:26:45 <rosmaita> sounds good ... thanks for bringing it up, it's good to keep the discussion out in the open
14:27:19 <rosmaita> i guess the other point is, please review that patch!
14:27:38 <rosmaita> #topic mTLS support in cinderclient
14:27:58 <rosmaita> this one missed the wallaby python-cinderclient release
14:28:13 <rosmaita> and it was the subject of a ML item yesterday
14:28:21 <rosmaita> #link http://lists.openstack.org/pipermail/openstack-discuss/2021-April/021822.html
14:28:27 <rosmaita> and i put a -1 on it this morning
14:28:44 <rosmaita> so it is getting some visibility
14:28:57 <rosmaita> i think that's all i have to say about that
14:29:12 <rosmaita> though it did prompt the next item
14:29:25 <rosmaita> #topic reviewing standards, practices, and strategy
14:29:46 <rosmaita> ok, so if you look at that patch, it has a bunch of +1s with no comments
14:29:54 <rosmaita> here is what i still think about that
14:30:03 <rosmaita> #link http://lists.openstack.org/pipermail/openstack-discuss/2019-May/006445.html
14:30:26 <rosmaita> that is my personal opinion ^^ and if you read the thread, not everyone agrees
14:30:34 <rosmaita> (though all the right-thinking people do)
14:30:59 <jungleboyj> :-)
14:30:59 <rosmaita> but here is why i am bringing this up
14:31:11 <rosmaita> if you want to speed up reviews on your patch, don't pile on +1s with no comment
14:31:21 <rosmaita> because i have been seeing a lot of that lately
14:31:30 <rosmaita> +1s with helpful comments are ... helpful
14:31:41 <rosmaita> you can speed up reviews by providing info that is helpful to reviewers
14:31:45 <rosmaita> here are some examples
14:31:54 <rosmaita> if this change is tested in CI, say where
14:32:02 <rosmaita> by "where", i mean link to the log
14:32:13 <rosmaita> make sure the bug description is clear ... especially for driver bugs, don't expect that everyone knows details about your backend
14:32:21 <enriquetaso> rosmaita++
14:32:31 <rosmaita> especially make sure you are clear about backend operations vs. cinder-side operations
14:32:52 <rosmaita> some of them have the same names, and it can be very confusing figuring out what the heck you are talking about
14:33:04 <rosmaita> make sure the commit message is clear about the fix
14:33:12 <rosmaita> #link https://wiki.openstack.org/wiki/GitCommitMessages#Information_in_commit_messages
14:33:33 <rosmaita> it's worth going back and looking at ^^ occasionally, some good advice in there
14:33:45 <rosmaita> make sure the release note is clear and correctly formatted
14:33:54 <jungleboyj> ++
14:33:56 <rosmaita> #link https://docs.openstack.org/cinder/latest/contributor/releasenotes.html
14:34:34 <rosmaita> and like for the mTLS review, it would be helpful to say +1 this is just like the fix in python-barbicanclient
14:34:53 <rosmaita> (except that in this case it isn't, which is why i have a -1 with a question on that patch)
14:35:46 <rosmaita> ok, so not to over do this
14:36:32 <rosmaita> please provide info when you +1 about why you are good with the fix
14:36:47 <rosmaita> any questions?
14:36:54 <rosmaita> or points other people want to make?
14:38:06 <jungleboyj> This shouldn't be surprising to people.  +1/-1 with no comments has always been frowned upon.  :-)
14:38:48 <rosmaita> :(
14:39:09 <jungleboyj> Why does that make rosmaita sad?
14:39:16 <rosmaita> i was frowning
14:39:20 <rosmaita> :)
14:39:34 * jungleboyj yells "Get off my lawn!"
14:39:46 <rosmaita> #topic requirements and lower-constraints
14:39:54 <rosmaita> #link http://lists.openstack.org/pipermail/openstack-discuss/2021-March/021204.html
14:40:03 <rosmaita> i have patches up removing the lower-constraints job from all stable branches back to rocky (where the l-c job was introduced)
14:40:12 <rosmaita> #link https://review.opendev.org/q/topic:%22cinder-drop-lc%22+(status:open%20OR%20status:merged)
14:40:22 <rosmaita> the good news is that our gates are mostly functional on the stable branches back to rocky
14:40:30 <rosmaita> there are three exceptions
14:40:33 <jungleboyj> ++
14:40:38 <rosmaita> #link https://review.opendev.org/q/topic:%2522cinder-drop-lc%2522+status:open+label:Verified%253C%253D-1
14:40:47 <rosmaita> for brick-cinderclient-ext, both stein and rocky, brick-cinderclient-dsvm-functional has TIMED_OUT
14:40:56 <rosmaita> i'm not sure what the problem is, and since that's not a high-traffic repo, and those are not releaseable branches, i'm not interested in following up
14:41:04 <rosmaita> (but that doesn't mean that no one should ... by all means, take a look if you are interested)
14:41:16 <rosmaita> for stable/rocky cinder, cinder-tempest-dsvm-lvm-lio-barbican is failing 8 network scenario tests
14:41:25 <rosmaita> i suspect a configuration problem, because those tests aren't even run in tempest-full or neutron-grenade
14:41:28 <tosky> oh, that requires a backport
14:41:33 <tosky> of a patch I send for stein
14:41:50 <rosmaita> ok, cool then
14:42:01 <tosky> https://review.opendev.org/c/openstack/cinder/+/778476
14:42:34 <rosmaita> got aheead and propose it for rocky
14:42:47 <rosmaita> it will be nice to have that gate working
14:43:20 <rosmaita> ok, next topic
14:43:35 <rosmaita> actually, a side topic
14:44:11 <rosmaita> those patches remove the l-c job, but not all the l-c stuff (like the tox testenv or the lower-constraints.txt file)
14:44:28 <rosmaita> i figured leave them there in case someone finds them useful?
14:44:56 <eharney> seems reasonable
14:45:17 <rosmaita> except for rocky, where it was such a PITA trying to figure out what was breaking that i ripped it all out on my patch
14:45:21 <rosmaita> :D
14:45:43 <rosmaita> ok, so the next topic is what to do in master
14:46:09 <rosmaita> but first, i should say something about the whole idea of lower-constraints so we can come to a reasonable decision
14:46:21 <rosmaita> The original lower-constraints.txt was generated from ... actually i'm not sure
14:46:30 <rosmaita> it was roughly pip-installing the minimum versions of stuff listed in {requirements.txt + test-requirements.txt + setup.cfg} and doing a pip freeze on the result
14:46:46 <rosmaita> but, it was not auto-maintained ... the only enforcement mechanism is that the job would fail if there was anything in {requirements.txt + test-requirements.txt + setup.cfg} that wasn't listed in lower-constraints.txt
14:47:09 <rosmaita> so, as a result, indirect dependencies of new additions to our requirements didn't make it into l-c unless someone was super-careful
14:47:37 <rosmaita> ok, so what is the point of the lower-constraints?
14:47:44 <rosmaita> supposed to give you a check on what minimum versions of dependencies can be used
14:47:54 <rosmaita> the l-c tox testenv does a pip install -c lower-constraints.txt
14:48:02 <rosmaita> (that's usually where the upper-constraints file is specified)
14:48:11 <rosmaita> and then runs the unit tests in that environment
14:48:19 <rosmaita> so it will fail if the current requirements aren't co-installable with the current lower-constraints
14:48:31 <rosmaita> or if our unit tests use some feature of a dependency that isn't available in the version listed in lower-constraints.txt
14:48:56 <rosmaita> but it doesn't run functional tests (i guess we could make it do that, though)
14:49:16 <tosky> just a note: the patch will be more complicated, as rocky uses the legacy version of that job
14:49:29 <rosmaita> tosky: barf
14:49:50 <rosmaita> maybe i will go back to my original plan of doing nothing for stable/rocky
14:50:09 <tosky> you may want to temporarily make the lvm-lio job non-voting
14:50:25 <rosmaita> i guess i would say, it's not a project priority to get stable/rocky working
14:50:26 <eharney> what's up with the lvm-lio job?
14:50:34 <tosky> on rocky, it's a legacy job
14:50:46 <rosmaita> unless someone on the team is supporting stable/rocky and has an interest in getting the gates working
14:50:48 <tosky> it's a _bit_ more complicated to fix it
14:51:11 <tosky> so... when are we going to schedule a discussion about abandoning a few other old branches?
14:51:17 <tosky> after xena releases?
14:51:21 <eharney> is it broken because of requirements issues?  maybe i missed something
14:51:34 <rosmaita> eharney: i think just general bit rot
14:51:54 <tosky> eharney: it is broken because some neutron tests need to be skipped, see https://review.opendev.org/c/openstack/cinder/+/778476
14:52:29 <rosmaita> to tell you the truth, i think  a higher priority is to fix cinderlib train gate
14:52:43 <rosmaita> because that is still releasable
14:52:58 <rosmaita> ok, we have got off track here
14:53:23 <tosky> isn't the next point "open discussion" anyway?
14:53:27 <rosmaita> we need to decide what to do about lower-constraints in master, because having a l-c job is optional
14:53:37 <rosmaita> option (1): drop completely
14:53:44 <rosmaita> pro: don't need to worry about it!
14:53:53 <rosmaita> con: lose the sanity check on the minima listed in our requirements
14:54:02 <rosmaita> option (2): reduce l-c.txt to direct dependencies only (i.e., those in {requirements.txt + test-requirements.txt + setup.cfg}
14:54:10 <rosmaita> pro: easier to maintain (smaller set of dependencies)
14:54:17 <rosmaita> pro: gives us a sanity check on the minima listed in the requirements files - we know our unit tests run with those versions
14:54:26 <rosmaita> con: still has to be maintained
14:54:34 <rosmaita> con: don't need the sanity check if we aggressively update the minima requirements each cycle
14:54:40 <rosmaita> (which we have been doing)
14:54:50 <rosmaita> option (3): keep full l-c.txt (basically, everything in pip list of one of the unit test tox envs)
14:54:59 <rosmaita> pro: even better sanity check than (2)
14:55:07 <rosmaita> con: harder to maintain than (2)
14:55:51 <rosmaita> are people ready to make a decision now, or should we revisit this next week?
14:56:24 <tosky> we don't always update the minimal requirements (for good reasons), so I believe 2 would still provide some value
14:57:55 <rosmaita> tosky: thanks ... let's revisit next week, because we also need a raise-the-min-requirements policy discussion
14:58:02 <rosmaita> #topic open discussion
14:58:07 <rosmaita> for 2 minutes
14:58:54 <rosmaita> i will change the requirements!!! on the PTG etherpad to requirements + lower-constraints + stable branch retirements
14:59:12 <jungleboyj> ++
15:00:14 <rosmaita> ok, don't forget about the festival of reviews on Friday, 1400-1600 UTC
15:00:20 <rosmaita> see you in meetpad!
15:00:24 <jungleboyj> Thanks!
15:00:26 <rosmaita> (except maybe hemna)
15:00:31 <rosmaita> #endmeeting