16:00:51 <jungleboyj> #startmeeting Cinder
16:00:52 <openstack> Meeting started Wed May 15 16:00:51 2019 UTC and is due to finish in 60 minutes.  The chair is jungleboyj. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:53 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:00:55 <openstack> The meeting name has been set to 'cinder'
16:00:58 <whoami-rajat> Hi
16:01:10 <davidsha> o/
16:01:11 <jungleboyj> courtesy ping:  jungleboyj diablo_rojo, diablo_rojo_phon, rajinir tbarron xyang xyang1 e0ne gouthamr erlon tpsilva ganso patrickeast tommylikehu eharney geguileo smcginnis lhx_ lhx__ aspiers jgriffith moshele hwalsh felipemonteiro lpetrut lseki _alastor_ whoami-rajat yikun rosmaita enriquetaso hemna _hemna
16:01:15 <smcginnis> o/
16:01:16 <_alastor_> \o
16:01:17 <xyang> hi
16:01:21 <rajinir> o/
16:01:22 <jungleboyj> @!
16:01:22 <_pewp_> jungleboyj ( ´ ▽ ` )ﾉ
16:01:27 <gouthamr> o/
16:01:30 <rosmaita> o/  -- i like the courtesy ping
16:01:40 <jungleboyj> rosmaita:  So do I.  :-)
16:01:51 <jungleboyj> rosmaita:  I miss meetings that don't have it.
16:01:58 <rosmaita> same here
16:01:59 <woojay> hi
16:02:11 <jungleboyj> Look all these wonderful people showing up.
16:02:29 <_erlon_> hey
16:02:50 <jungleboyj> Give people one more minute here.
16:02:59 <jungleboyj> Take it everyone made it home safely?
16:03:19 <smcginnis> I made it home dangerously.
16:03:28 <jungleboyj> smcginnis:  Nice!
16:03:31 <smcginnis> :P
16:03:31 <rosmaita> i made it home eventually
16:03:43 <jungleboyj> Get the pilot to do some barrel rolls?
16:03:48 <e0ne> hi
16:03:58 <jungleboyj> rosmaita:  :-(
16:04:07 <jungleboyj> Well, eventually is better than not at all.
16:04:12 <rosmaita> yep
16:04:16 <jungleboyj> My 6 am flight was brutal.
16:04:30 <jungleboyj> Anyway, we have a lot to cover, so lets get started.
16:04:36 <jungleboyj> #topic announcements
16:04:53 <jungleboyj> First, we have defined the Week for the Train Mid-Cycle!
16:05:04 <jungleboyj> #link https://etherpad.openstack.org/p/cinder-train-mid-cycle-planning
16:05:20 <jungleboyj> If you are interested in attending in person or remotely please add your name to the etherpad.
16:06:05 <jungleboyj> I let Lenovo know we are planning to be there again and haven't heard a no.  I am pretty sure we will be ok as I had already budgeted for it.
16:06:27 <jungleboyj> Hope we can get a few more people to attend this time as the last one was really productive.
16:07:12 <jungleboyj> Also, there is the requirement for drivers to move to Python 3 for their 3rd Party CI.
16:07:24 <jungleboyj> #link http://lists.openstack.org/pipermail/openstack-discuss/2019-May/006001.html
16:07:41 <smcginnis> We said by milestone-2, right?
16:07:41 <jungleboyj> If you own a driver and aren't already doing this/thinking about it ... Please start.
16:07:45 <jungleboyj> Correct.
16:07:54 <jungleboyj> I just realized I should add that to the schedule I just proposed.
16:08:01 <smcginnis> Oh, good idea.
16:08:24 <whoami-rajat> VNX driver specifically ? ^
16:08:36 <smcginnis> If they find issues with running under py3 - milestone-3 to get that resolved?
16:08:39 <jungleboyj> whoami-rajat:  Is that the one that is currently failing?
16:08:46 <smcginnis> Or running successfully by milestone-2?
16:08:53 <jungleboyj> smcginnis:  ++  I think that is fair.
16:09:03 <whoami-rajat> jungleboyj:  yes, IIRC
16:09:06 <jungleboyj> I am ok with giving them a bit to work out the failures.
16:09:31 <smcginnis> We will need to disable the VNX unit tests if that is not resolved soon since py3.7 is one of the official runtimes for train.
16:09:41 <jungleboyj> whoami-rajat:  Ok.  I still have a todo to ping Helen about that.  She isn't in the meeting so I will e-mail her.
16:09:52 <mszwed_> Hi :-)
16:10:07 <whoami-rajat> jungleboyj: sounds good. thanks.
16:10:35 <enriquetaso> o/
16:10:36 <jungleboyj> Is everyone in agreement to the Python3 dates above?
16:10:43 * jungleboyj marks enriquetaso tardy
16:10:53 <enriquetaso> u.u
16:11:04 <smcginnis> Should we say py3.7, or is py3.6 enough?
16:11:25 * smcginnis thinks py3.7 really would be the interesting one
16:11:57 <jungleboyj> smcginnis: I don't know enough to have a strong opinion.
16:12:10 <jungleboyj> I defer to your expertise there.
16:12:20 <e0ne> I prefere 3.7 from developer's perspective
16:12:26 <whoami-rajat> I think py37 would be better
16:12:33 <e0ne> but it would be good to check with disctros
16:12:40 <e0ne> #link https://governance.openstack.org/tc/reference/runtimes/train.html
16:12:42 <smcginnis> 3.7 is where they really broke some compatibility, so if we are saying 3.7 is an official runtime for train, then I think the CIs probably should use that version to make sure there are no surprises for end users.
16:12:57 <jungleboyj> smcginnis:  ++
16:13:14 <rosmaita> smcginnis: ++
16:13:15 <jungleboyj> That makes sense with me and then we should be confident that whatever distros are using should be good.
16:14:12 <jungleboyj> So, lets call it Py37 then.
16:14:29 <smcginnis> ++
16:14:36 <jungleboyj> The people actually impacted are being too quiet.
16:15:02 <jungleboyj> But, I can move on.
16:15:16 <jungleboyj> I have posted the summary of the PTG and Forum:
16:15:28 <jungleboyj> #link https://wiki.openstack.org/wiki/CinderTrainSummitandPTGSummary#Train_PTG_Summary
16:15:34 <mszwed_> +
16:15:39 <jungleboyj> Thank you to everyone who came and helped make our time successful.
16:15:50 <jungleboyj> I have outlined who has action items in there.
16:16:08 <jungleboyj> Please take a look and plan to take your action items.  :-)
16:16:36 <jungleboyj> If you see any inconsistencies or mis-interpretations please let me know.
16:17:00 <jungleboyj> Finally, I have proposed the Train release schedule:
16:17:08 <rosmaita> is it ok if we just edit the wiki?
16:17:14 <jungleboyj> #link https://review.opendev.org/659330
16:17:26 <jungleboyj> rosmaita:  Sure, that would be fine.  Just let me know what I did wrong if you can.
16:17:29 <jungleboyj> :-)
16:17:31 <rosmaita> ok
16:17:43 <_erlon_> jungleboyj: nice summary Jay. Thanks a lot!
16:17:55 <jungleboyj> Have two things to update after the meeting, it sounds like, so I will push up another patch.
16:17:56 <_alastor_> I notice the 3rd party CIs are supposed to be running the Cinder tempest plugin, but all-plugin has been deprecated in tempest
16:18:15 <jungleboyj> _erlon_:  Thanks.  Glad to do it.
16:18:45 <jungleboyj> _alastor_:  Yeah, I saw that.  Can we chat about that later in the meeting?
16:19:01 <_alastor_> yeah np
16:19:06 <jungleboyj> _alastor_:  Thank you sir.
16:19:21 <jungleboyj> So, take a look at the release schedule and please add review comments if you have concerns.
16:19:48 <jungleboyj> Ok.  That was what I had for announcements.  That was quite a bit.
16:20:07 <jungleboyj> #topic the unwritten rules about same-company approvals
16:20:19 <jungleboyj> #link http://lists.openstack.org/pipermail/openstack-discuss/2019-May/005954.html
16:20:31 <jungleboyj> So, this is a discussion that has been happening on the ML.
16:20:47 <jungleboyj> I haven't followed the later updates, honestly, because it started getting a little political and silly.
16:20:58 <jungleboyj> rosmaita:  Wanted to know our expectations.
16:21:11 <rosmaita> yeah, it seemed like "don'
16:21:25 <rosmaita> t have all approvals be from a single company" could be written down
16:21:30 <rosmaita> so i was wondering if there was more to it
16:21:32 <jungleboyj> Long story short, this used to be a bigger issue when we had multiple cores from the same company trying to push features in.
16:21:52 <jungleboyj> Now, given our current landscape it is less of an issue.
16:21:59 <smcginnis> I think like how we said trivial things like small doc changes can be single approved, if there is no potential concerns it's OK to approve with two from the same company. Just discourage it in normal practice.
16:22:30 <smcginnis> I'd only be really concerned if the proposer and both cores were all from the same company.
16:22:31 <jungleboyj> But yes, the short answer is, be a good citizen and don't have all the reviews on a patch be from one company.  Especially if it is a feature or something you are pushing.
16:22:32 <rosmaita> ok, so for bugfixes, not a big deal ... features, think twice
16:22:47 <smcginnis> rosmaita: I think that's a great way to frame it.
16:22:53 <rosmaita> ok, cool, thanks
16:22:55 <jungleboyj> smcginnis: rosmaita ++
16:22:59 <e0ne> smcginnis: +1
16:23:13 <jungleboyj> I would rather have you ping me on IRC and push the patch in my face.
16:23:27 <rosmaita> ok, i will keep that in mind!
16:23:32 <jungleboyj> It just keeps us looking more credible.
16:24:12 <jungleboyj> Any other questions around that?
16:24:21 <rosmaita> not from me
16:25:02 <jungleboyj> Ok.
16:25:16 <jungleboyj> #topic dell_emc.vnx.test_adapter.TestCommonAdapter failing py37 tests
16:25:18 <jungleboyj> rosmaita:
16:25:32 <rosmaita> yeah, i just wanted to bring the bug to people's attention
16:25:49 <rosmaita> i approved rajat's patch adding the py37 job as non-voting
16:25:50 <jungleboyj> #link https://bugs.launchpad.net/cinder/+bug/1795957
16:25:51 <openstack> Launchpad bug 1795957 in Cinder "VNX unit tests are not Python 3.7 compatible" [Undecided,Confirmed]
16:26:06 <rosmaita> and walter had mentioned that we should make sure some attention is paid to the failing tests
16:26:13 <jungleboyj> rosmaita:  Right.
16:26:20 <rosmaita> so i figured this way we all know about it
16:26:25 <rosmaita> that's all
16:26:27 <jungleboyj> Need to ping Helen Walsh about that.
16:26:33 <jungleboyj> I will do that after the meeting.
16:26:42 <rajinir> rosmaita: the vnx driver team is looking into the unit failures. Helen is vmax
16:26:54 <jungleboyj> rajinir:  Oh.  Gotcha.
16:26:57 <rosmaita> rajinir: thanks
16:27:01 <jungleboyj> rajinir:  Who is VNX?
16:27:33 <rajinir> jungleboyj: larger team in china, sam and sun hao
16:28:04 <jungleboyj> rajinir:  Ok.  Thanks.  So it is being worked?
16:28:06 <rajinir> jungleboyj: I will update them
16:28:19 <rajinir> jungleboyj: yes,
16:28:25 <jungleboyj> rajinir:  Great.  Thank you!
16:28:30 <whoami-rajat> i saw ryan liang pushing a lot of patches for VNX recently, not sure if he's on IRC.
16:28:58 <whoami-rajat> rajinir: thanks for the update.
16:29:02 <rajinir> ryan liang is also part of that hao's team. Time zone issues
16:29:20 <jungleboyj> Yay timezones.
16:29:26 <whoami-rajat> rajinir: oh ok
16:29:31 <jungleboyj> Ok, so if that is being worked we can move forward.
16:29:47 <jungleboyj> #topic Move away from courtesy pings?
16:29:49 <rajinir> can we assign this bug to ryan https://bugs.launchpad.net/cinder/+bug/1795957
16:29:50 <openstack> Launchpad bug 1795957 in Cinder "VNX unit tests are not Python 3.7 compatible" [Undecided,Confirmed]
16:29:59 <jungleboyj> rajinir:  ++
16:30:26 <whoami-rajat> rajinir: ++
16:30:29 <rosmaita> rajinir: is ryan at emc.com?
16:30:49 <rajinir> rosmaita: dell emc is one company now, yes
16:31:04 <rosmaita> ok, just wanted to make sure i had the correct person
16:31:09 <jungleboyj> :-)
16:31:16 <rajinir> rosmaita: ++
16:31:32 <jungleboyj> So, there has been discussion with some groups that Courtesy Pings for meetings are bad form.
16:31:42 <jungleboyj> Though, it doesn't seem that everyone is in agreement.
16:32:15 <jungleboyj> For those of us that use IRC Cloud I have not yet figured out a way to highlight on 'startmeeting <cinder|oslo>' etc.
16:32:26 <jungleboyj> Highlighting on just Cinder gets me spammed all the time.
16:32:55 <jungleboyj> So, I personally think that the pings work nicely for those of us that see meeting reminders, get busy and forget.
16:33:05 <rosmaita> jungleboyj: ++
16:33:19 <rajinir> jungleboyj: ++
16:33:25 <_erlon_> it works for me
16:33:31 <bnemec> Ick. I'm finding that the docs for Quassel's highlight configuration are somewhat lacking too.
16:33:35 <jungleboyj> I was going to propose that I remove the current ping list as I know there are a number of people on there that don't attend anymore.
16:33:56 <jungleboyj> If you want to be pinged, please add your name back in.
16:33:59 <e0ne> jungleboyj: it works better than calendar reminder for me:)
16:34:00 <rosmaita> we can call it a "discourteous ping", that would still work for me
16:34:01 <smcginnis> I agree with Morgan's post that as long as the list is self-managed, there's no reason not to use it.
16:34:09 <smcginnis> rosmaita: ++ :)
16:34:25 <bnemec> It might be worth requiring everyone to re-subscribe each cycle?
16:34:31 <jungleboyj> I can also send an e-mail to the mailing list so people know.
16:34:35 <enriquetaso> rosmaita++
16:34:37 <bnemec> I know in Oslo we almost certainly have some stale entries in our ping list.
16:34:37 <jungleboyj> bnemec:  ++
16:34:53 <jungleboyj> bnemec:  Yeah, if you lose the pings, I am probably going to miss your meeting more often than not.
16:35:06 <mszwed_> jungleboyj ++
16:35:07 * bnemec needs to do more than skim the responses to that thread
16:35:09 <jungleboyj> Especially on a Monday morning.  :-)
16:35:18 <bnemec> Yeah, no kidding.
16:35:30 <bnemec> The PTL should propose a reschedule to a less terrible time. ;-)
16:35:57 <jungleboyj> Ok, so are we in agreement we keep the pings and I am just going to remove the current list in the Train etherpad so that people can add themselves back in if wanted?
16:36:07 <rosmaita> sounds good
16:36:12 <mszwed_> You can't satisfy every time zone ;-)
16:36:21 <jungleboyj> mszwed_:  Sad but true.
16:36:31 <bnemec> Yeah, I'm pretty sure that's why it is when it is.
16:36:46 <bnemec> The previous PTL is in China so it was the only time that sort of worked.
16:36:48 <mszwed_> jungleboyj but send mail with that info
16:37:05 <rosmaita> we once tried having different meeting times every other week in glance, but that didn't work, everyone always had the time wrong
16:37:11 <jungleboyj> mszwed_:  Of course.
16:37:46 <jungleboyj> rosmaita: ++
16:38:06 <enriquetaso> I agree with the pings
16:38:07 <jungleboyj> This time doesn't work for China unfortunately, but at least we are getting most people.
16:38:25 <jungleboyj> Ok, seems like we more or less have agreement on the pings.
16:38:30 <mszwed_> In SPDK we are using two time slots: US/China and US/Europe friendly, but it's not the best
16:39:26 <jungleboyj> I have been with groups like that and it seemed to just result in two less useful meetings with a lot of repetition.
16:39:36 <jungleboyj> Anyway, barring disagreement, lets move on.
16:39:44 <mszwed_> Exactly
16:40:00 <jungleboyj> #topic proposal to change stable/pike,queens u-c to address a CVE
16:40:11 * jungleboyj nominates rosmaita as shadow PTL
16:40:20 <rosmaita> i just can't keep my mouth shut
16:40:31 <rosmaita> anyway, this is looking like it may be a non-issue
16:40:31 <jungleboyj> #link http://lists.openstack.org/pipermail/openstack-discuss/2019-May/005956.html
16:40:46 <smcginnis> The requests CVE is something that the requirements team will need to control, but we can provide input on that thread.
16:40:51 <jungleboyj> #link https://review.opendev.org/#/q/topic:cve-2018-18074+status:open
16:41:17 <rosmaita> short version, there was a proposal to change hte upper-constraints for requests to address a CVE
16:41:23 <rosmaita> on stable/pike and stable/queens
16:41:32 <rosmaita> which is a pretty unusual thing to do
16:41:51 <rosmaita> anyway, i put up some patches to see what would happen to cinder, os-brick, and the cinderclient
16:41:53 <jungleboyj> Thanks for catching that.
16:42:01 <jungleboyj> rosmaita:  Awesome!
16:42:06 <rosmaita> they look ok, except for all the failures on non-voting jobs
16:42:35 <rosmaita> so it might be a good idea for people to take a quick look if they have any interest in some of the non-voting jobs
16:42:50 <rosmaita> but like i said, it's starting to look like the change won't happen
16:43:01 <rosmaita> but this is openstack, so you never know
16:43:08 <rosmaita> that's all
16:43:43 <_erlon_> rosmaita: what is the point for not having the change?
16:43:53 <jungleboyj> Cool.  Thank you for bringing that up.
16:44:14 <rosmaita> that it destablilizes the branches, and it's up to each consumer to make the update if they want ot
16:44:20 <rosmaita> _erlon_: ^^
16:44:52 <_erlon_> rosmaita: hmm, got it
16:45:14 <rosmaita> _erlon_: the justification is a bit more nuanced, but that's basically it
16:45:48 <jungleboyj> rosmaita:  Will you keep an eye at as to whether that looks like it is going to go forward?
16:45:56 <jungleboyj> Let us know if we need to look closer.
16:46:01 <rosmaita> jungleboyj: yes, i will follow the action on that
16:46:09 <jungleboyj> rosmaita:  Cool.  Thanks.
16:46:44 <jungleboyj> So, next topic.
16:47:02 <jungleboyj> #topic Unified limits implementation
16:47:06 <jungleboyj> whoami-rajat:
16:47:18 <whoami-rajat> I should have probably brought that up at the PTG but anyways
16:47:32 <whoami-rajat> i've encountered bugs in past related to quotas, most of which will (i think) be solved after unified limits implementation.
16:47:42 <whoami-rajat> nova has already started working on the same,
16:48:30 <whoami-rajat> in a brief: all quota management will be handled at keystone database and limits will be enforced by the oslo.limits lib.
16:48:37 <jungleboyj> whoami-rajat:  I think we touched on this at the PTG and that we did want to make the move.
16:48:55 <whoami-rajat> I'm willing to work on it if everyone has an agreement to start with the implementation in cinder.
16:49:00 <jungleboyj> It was secondary to a discussion that smcginnis started.
16:49:29 <jungleboyj> whoami-rajat:  I have no objections
16:49:41 <whoami-rajat> oh, i was unaware that work has already started on cinder side jungleboyj
16:49:49 <smcginnis> We may want to wait until U so we can also remove that nested quota driver, but if it can be done now and worked into the current implementation, that's great.
16:50:17 <smcginnis> The work itself hasn't been started, but that was deprecated to try to clean up some of the code before we make any major changes.
16:50:21 <jungleboyj> whoami-rajat:  Work hadn't started.  It was just related to smcginnis wanting to remove the nested quota driver.
16:50:28 <jungleboyj> smcginnis: ++
16:50:46 <bnemec> Note that oslo.limit isn't implemented yet. It's just a bare repo.
16:50:56 <bnemec> We'd love to have more input on the initial implementation though.
16:51:09 <smcginnis> Though really, not sure if we need to deprecate that like we do with backend drivers. There will probably be major changes anyway, so since we do a major version bump each cycle, we can make backwards incompatible changes.
16:51:37 <smcginnis> bnemec: Good point. I didn't think we were quite ready for everyone to jump on unified limits quite yet.
16:52:47 <whoami-rajat> #link https://review.opendev.org/#/c/602201/8
16:52:47 <whoami-rajat> spec from nova side
16:53:25 <smcginnis> Would definitely be good to keep an eye on that to make sure things don't end up too nova-centric and have issues for others to adapt to.
16:53:36 <jungleboyj> smcginnis:  ++
16:54:08 <jungleboyj> So, at this point whoami-rajat  it sounds like your work would to be to keep an eye on how the development of UL is going and make sure it matches with what we will need.
16:54:29 <jungleboyj> We do want to move to it and have it fix our quota issues, so it is time well spent.
16:54:43 <rosmaita> here's sean's patch deprecating the nested quota driver: https://review.opendev.org/#/c/657511/
16:54:53 <whoami-rajat> jungleboyj:  ack
16:55:12 <smcginnis> I think Cinder is the only project (that I know of) that ever tried implementing hierarchical quotas, so Cinder probably would be a better place to start. So I guess it just comes down to someone having the time to push it.
16:55:30 <jungleboyj> smcginnis:  ++
16:55:38 <jungleboyj> Sounds like we have a volunteer?
16:56:25 <whoami-rajat> jungleboyj: can do
16:56:34 <jungleboyj> whoami-rajat:  Thank you!
16:56:57 <whoami-rajat> jungleboyj: mp :)
16:56:58 <jungleboyj> Ok, I had wanted to cover priorities for the release but we are short on time.  I can work on getting things organized for next week and cover it then.
16:57:07 <jungleboyj> _alastor_:  Tempest.
16:57:14 <_alastor_> yeah
16:57:23 <_alastor_> what gives?
16:57:36 <jungleboyj> So, is it just our documentation that is out of date?
16:57:55 <jungleboyj> Is there a different process that needs to be followed?
16:58:20 <_alastor_> I have no idea.  What I know is when I run tempest with 'all-plugin' in won't run
16:58:20 <jungleboyj> enriquetaso:  You and Alan have been looking at Tempest some.  RIght?
16:58:25 <smcginnis> Someone should ask on the ML or in #openstack-qa what the current correct way is to run the tempest plugins.
16:58:36 <jungleboyj> smcginnis: ++
16:58:42 <smcginnis> I think it might just pick them up and run them as long as the repos are added to the list that gets installed.
16:59:20 <jungleboyj> _alastor_:  Would you be willing to start that discussion on the ML?
16:59:38 <_alastor_> sure
16:59:45 <jungleboyj> Awesome.  Thank you.
17:00:09 <enriquetaso> jungleboyj, yup I've been working on cinder tempest
17:00:30 <jungleboyj> enriquetaso:  Ok, if you have any insights for _alastor_  Can you let him know?
17:00:44 <jungleboyj> And we are at time so lets continue that in the Cinder channel.
17:00:53 <enriquetaso> 'all-plugin' is not working for me neither
17:00:53 <jungleboyj> Thank you to everyone for joining us!
17:00:58 <enriquetaso> jungleboyj, _alastor_ yes, sure
17:01:03 <jungleboyj> Cool.
17:01:08 <jungleboyj> Talk to you all next week.
17:01:13 <jungleboyj> #endmeeting