Wednesday, 2021-11-10

*** melwitt is now known as Guest550810:12
*** dviroel|out is now known as dviroel11:16
*** senrique__ is now known as enriquetaso13:36
rosmaita#startmeeting cinder14:00
opendevmeetMeeting started Wed Nov 10 14:00:14 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
opendevmeetUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.14:00
opendevmeetThe meeting name has been set to 'cinder'14:00
rosmaita#topic roll call14:00
whoami-rajat__Hi14:00
fabiooliveirao/14:00
shoffmannhi14:00
*** whoami-rajat__ is now known as whoami-rajat14:00
enriquetasohi14:00
eharneyhi14:00
sfernandhi14:00
walshh_hi14:00
smorehi14:00
e0nehi14:01
angioluccio/14:01
rosmaitagood turnout!14:01
rosmaita#link https://etherpad.opendev.org/p/cinder-yoga-meetings14:02
rosmaitaok, let's get started14:02
LeoCampeloo/14:02
rosmaitathis one isn't on the agenda, but Yoga Milestone 1 is next week14:02
jungleboyjo/14:03
rosmaitaafter this meeting, there is another continuation of the secure RBAC goal for Yoga14:03
rosmaita#link http://lists.openstack.org/pipermail/openstack-discuss/2021-November/025666.html14:03
rosmaitaif you are interested ^^14:03
rosmaitawe are having an impromptu Festival of RBD driver reviews tomorrow14:04
rosmaita#link http://lists.openstack.org/pipermail/openstack-discuss/2021-November/025683.html14:04
rosmaitait will work like the other review festival, that is, we'll "meet" in video but work separately doing reviews14:04
rosmaitausing the video to ask questions or tell people something was just +2d and could use another look14:05
rosmaitai am trying out google hangouts, which i became aware yesterday may not work for everyone14:05
rosmaitaso, you can always join and just open the etherpad and ask questions that way14:05
rosmaita#link https://etherpad.opendev.org/p/cinder-festival-of-driver-reviews14:06
rosmaitaok, final thing i want to mention is that there's been a discussion on the ML about release cadence14:06
rosmaitathat is, should we continue to release openstack every 6 months, or try something else14:07
rosmaita#link http://lists.openstack.org/pipermail/openstack-discuss/2021-November/025684.html14:07
rosmaitathat's the start of the thread ^^14:07
rosmaitaso if you are interested in such stuff and/or have an opinion, you may want to get in on the action14:07
rosmaitaok, i mentioned earlier that M-1 is next week14:08
rosmaitaa key priority for early in the cycle is to get the sqlalchemy-migrate -> alembic work done14:09
rosmaita#link https://review.opendev.org/q/topic:%2522bp/remove-sqlalchemy-migrate%2522+(status:open+OR+status:merged)+project:openstack/cinder14:09
rosmaitaso that is what i am going to be mostly reviewing over the next few days14:09
rosmaitaof course, others need to review also14:09
enriquetasosure, i'll review as well 14:10
rosmaitabut my point is, asking for reviews of anything other than critical bugs is unlikely to result in action14:10
rosmaitathanks enriquetaso!14:10
rosmaitathat's it from me, is there anything else going on that the team should be aware of?14:10
rosmaitai know several people have had recent holidays in various locations, so welcome back and i hope you are refreshed!14:11
rosmaitaok, looks like everything has been announced ... moving on14:12
rosmaita#topic Reviews need for NVMe multipathing spec14:12
rosmaitaquick comment before you get started simondodsley14:13
rosmaitathis is for everyone putting up specs14:13
rosmaitawe write them in restructured text/sphinx, but end users will read them in HTML14:13
eharneyfg14:13
rosmaitaso it's important that they render in a way that makes sense14:13
eharney(wrong window, sorry)14:13
rosmaita:)14:13
rosmaitayou can check this yourself when putting together the spec by using your tox environment to build the docs and then open your spec in a browser14:14
whoami-rajatrunning tox -edocs and checking the rendered html file is a good practice before pushing patch ^^14:14
jungleboyj++14:15
rosmaitaor, if you are reviewing, the zuul job builds the html and you can look at it there14:15
rosmaitait's a little hard to find, so here are directions14:16
rosmaitatake this patch for example: https://review.opendev.org/c/openstack/governance/+/81515814:16
rosmaitago to the "Zuul Summary" tab14:16
rosmaitaclick on "openstack-tox-docs"14:16
rosmaitathere you will see an "Artifacts" tab ... click it14:16
rosmaitalast item in the list is "Docs preview site" ... click it14:17
enriquetasocool14:17
rosmaitathen you need to know how to find your doc ... this example is a community goal, so you will scroll to community goals, find accepted, find yoga14:17
rosmaitafor our specs, it's much easier because they're all basically listed on the first page that opens14:18
rosmaitaanyway, same process can be used when you are reviewing a docs change or a release note14:18
whoami-rajatI usually go to View log -> docs but good to know the alternative14:19
rosmaitaand if you are really stuck with some weird formatting that you can't fix, you can ask in the cinder channel, someone may be able to help14:19
enriquetaso#link https://review.opendev.org/c/openstack/cinder-specs/+/81561314:20
rosmaitamy other point is that reviewers tend to ignore patches that have -1 from zuul14:20
rosmaitaso it's kind of important to make sure your spec is getting a +1 if you want some review action14:21
rosmaitaok, that's it from me ... i think simondodsley's point is that that spec is available and can use some reviews14:21
rosmaita:)14:21
rosmaitaspec freeze is R-1514:21
rosmaita(this week is R-20)14:22
rosmaitaok, guess there's no more to say about that one14:24
rosmaita#topic Pure driver patch reviews needed14:24
rosmaita#link https://review.opendev.org/c/openstack/cinder/+/81708514:24
rosmaitalooks like a code simplification based on backend capabilities14:25
rosmaitaanyway, i left a question on the patch14:25
rosmaitaok, big topic for this week14:25
rosmaita#link ussuri final release14:26
rosmaitawhoami-rajat: you have the floor14:26
whoami-rajathi14:26
whoami-rajatso we've the final ussuri release targeted for 12th Nov14:26
whoami-rajatand an etherpad to keep track of patches that needs to get in it14:26
whoami-rajat#link https://etherpad.opendev.org/p/ussuri-final-release14:26
whoami-rajatbut looks like 2 of the os-brick patches haven't even merged in master yet14:27
whoami-rajatI wanted to know if we have an option to delay the release since they look important or just release with whatever we have merged?14:27
whoami-rajatThe ML has this line which i guess implies for whole openstack so not sure if we can change release date for cinder specifically "In a month Ussuri is planned to transition to Extended Maintenance phase14:28
whoami-rajat[1] (planned date: 2021-11-12)."14:28
rosmaitalet's discuss whether the 2 brick patches need to get into a release14:28
rosmaita1. NVMe multipath enabled kernels14:29
rosmaitawe haven't had that functionality, so i am not sure that it needs to be released in ussuri14:29
rosmaitait would be good to hear from NVMe users/deployers, however, if anyone is here14:29
rosmaitaanyone know who put it on the etherpad?  that color is 'unamed'14:31
whoami-rajatmaybe Zohar did since he's the author but not sure14:31
rosmaitagood reminder ... please "sign in" to an etherpad when you leave comments so we know it's you14:31
rosmaitacould be14:32
rosmaitai think the nvme connector in ussuri can remain as good as it has always been in ussuri, then14:32
jungleboyj++14:32
rosmaitaok, so that one is not needed14:32
rosmaitathe file locks, though, i think is kind of important14:32
simondodsleyHere now. Got delayed 14:33
eharneyyes, would be good to get that one in14:33
rosmaitadoesn't show up in our test environments, but looks like it cna be a problem for a not uncommon deployment configuration14:33
rosmaitaok, my -1 on that has to do with documenting the lock path in the release note, so should be pretty easy to address (or talk me out of)14:34
whoami-rajatpossible in a scenario where 2 volumes from 2 different cinder backend are going through attachment related operation14:34
rosmaitaok, so let's get "Use file locks in connectors" into the release14:35
rosmaitai think the code and tests are fine on that one, i need someone to look at my comment about the release note14:36
rosmaitagorka has been on away, i don't think he will mind someone revising the release note for him14:36
rosmaitaok, so that's brick14:37
rosmaitacinderclient patch has merged14:37
rosmaitacinderlib doesn't have anything open14:37
jungleboyjrosmaita: Do you want to update the release note then and we can help get it merged?14:37
rosmaitai can, but what i'm looking for is some assurance that my comment there is correct14:38
rosmaita:)14:38
rosmaitawhile people are looking, there are 2 items for cinder14:38
rosmaitafirst one is "Add libcgroup related packages in bindep.txt"14:39
rosmaita#link https://review.opendev.org/c/openstack/cinder/+/81588314:39
rosmaitanot sure why that is a WIP14:39
rosmaitait's kind of a nice-to-have, but people have gotten along without it, so i guess it's not essential14:40
eharneyqemu-img is in there?  need to refresh myself on this...14:40
eharneyoh. bad cherry-pick14:40
rosmaitathe other one fixes a regression in the rbd driver14:40
rosmaitaeharney: ok, that explains it14:41
enriquetasothe cherry picked everything.. even the qemu-img14:41
rosmaitais anyone motivated enough to correct that patch, or should we just let it go?14:42
whoami-rajatrosmaita, I'm not sure if we should pass the lock_path via os-brick (in code along with external=True) or use the configured one in cinder/nova/glance using os-brick14:42
enriquetasocan correct the patch and resubmit 14:44
rosmaitawhoami-rajat: i think we want to use the configured one for the service that is doing the attaching?14:45
rosmaitaok, so for cinder, enriquetaso is going to fix the c-p for the bindep.txt change14:47
rosmaitathe rbd one i think is uncontroversial, it's in the victoria gate now14:47
rosmaitaso this brings us to whoami-rajat's question about when we release14:47
rosmaitai think we can do it when we are ready ... i don't see any release patches proposed for ussuri-em yet14:48
whoami-rajatrosmaita, yeah, makes sense since os-brick isn't aware about the deployment tool/distro so won't be able to tell correct path for the locks14:48
rosmaitaok14:48
whoami-rajatrosmaita, should we ask release team regarding the same? just to be sure14:49
rosmaitayeah, that's a good idea14:49
rosmaitahttps://review.opendev.org/q/project:openstack/releases+status:open14:49
whoami-rajatack, i will ask Elod since he sent out the mail for ussuri EM14:49
rosmaitalooks like a few projects have proposed patches, but i don't see the generate ones yet14:50
rosmaitawhoami-rajat: thanks ... tell him that we have a specific list of what we need to get in there, so it's not like it will drag on forever14:50
rosmaitawe just need probably one more week14:50
whoami-rajatack, will do, thanks for the feedback14:51
rosmaitaok, looking good for ussuri final release (more or less)14:51
rosmaita#topic volume driver API14:51
rosmaitaneed a solidfire driver maintainer to vote onĀ https://review.opendev.org/c/openstack/cinder/+/81489814:52
rosmaitathat's because we are removing a pair of functions that no other drivers used14:52
rosmaitai'd like an ack on the patch since it will leave stale metadata in the backend (possibly)14:52
rosmaitathe other patch has to do with merging 2 functions into a single call14:53
rosmaitait touches a bunch of drivers14:53
rosmaita#link https://review.opendev.org/c/openstack/cinder/+/81565914:53
rosmaitai'm not going to hold the patch for acks from everyone, but i would like to at least have a few14:54
rosmaitaalso, gorka did some minor refactoring as he was fixing tests14:54
rosmaitasome good test patterns that you should be aware of14:54
rosmaitaso that's all for the volume driver API ... hopefully we can get the reviews completed and merged before next meeting, and then start on another function14:55
rosmaitaactually one other patch to look at:14:55
rosmaita#link https://review.opendev.org/c/openstack/cinder/+/815660/14:55
rosmaitathat changes the documentation for some functions14:55
rosmaitai have a formatting quibble 14:56
rosmaitabut it would be good to get feedback on that, because it will be the template on how we'll be documenting these from now on14:56
rosmaitaso act now if you want your opinion taken into account14:56
rosmaita#topic mypy patch of the week14:57
rosmaitai am happy to announce that last week's patch is merged14:57
rosmaitaalong with a bonus mypy patch14:57
jungleboyj\o/14:57
rosmaitaso we are moving right along!14:57
rosmaitathis week's patch:14:57
rosmaita#link https://review.opendev.org/c/openstack/cinder/+/73362214:57
eharneyyay14:58
rosmaitai need 2 core and on non-core volunteer to look this one over14:58
rosmaitai am willing to be 1 core, but i don't want to hog all the glory here14:58
whoami-rajati can take a look14:58
fabiooliveirai could take a look also14:59
rosmaitagreat! ty14:59
rosmaitaok, i guess i will be the other core while i am in a mypy state of mind15:00
rosmaitaall right, we are out of time15:00
rosmaitathanks everyone, and don't forget the RBD Festival of Reviews tomorrow15:00
rosmaita#endmeeting15:00
opendevmeetMeeting ended Wed Nov 10 15:00:57 2021 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)15:00
opendevmeetMinutes:        https://meetings.opendev.org/meetings/cinder/2021/cinder.2021-11-10-14.00.html15:00
opendevmeetMinutes (text): https://meetings.opendev.org/meetings/cinder/2021/cinder.2021-11-10-14.00.txt15:00
opendevmeetLog:            https://meetings.opendev.org/meetings/cinder/2021/cinder.2021-11-10-14.00.log.html15:00
*** dviroel is now known as dviroel|out21:15

Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!