16:58:57 <knikolla> #startmeeting keystone
16:58:58 <openstack> Meeting started Tue Jan 19 16:58:57 2021 UTC and is due to finish in 60 minutes.  The chair is knikolla. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:58:59 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:59:01 <knikolla> o/
16:59:02 <openstack> The meeting name has been set to 'keystone'
17:00:46 <lbragstad> o/
17:04:24 <knikolla> cmurphy: gagehugo: around?
17:04:32 <gagehugo> o/
17:04:43 <cmurphy> o/
17:05:02 <knikolla> #topic Follow-up on gate instability
17:05:43 <knikolla> I read the discussion in last week's keystone meeting and saw that there were concerns about dropping lower constraint testing from the gates
17:06:22 <knikolla> In the last TC meting, we discussed the topic and reached out to packagers for RDO, ubuntu, and debian, and they didn't use lower-constraints in their packaging requirements, instead relying on upper-constraints
17:07:35 <knikolla> #link http://eavesdrop.openstack.org/meetings/tc/2021/tc.2021-01-14-15.00.log.html#l-105
17:08:40 <cmurphy> good to know
17:08:59 <knikolla> #link http://lists.openstack.org/pipermail/openstack-discuss/2021-January/019877.html
17:09:46 <knikolla> with regards to the timeouts, i have encountered those with the new pip dependency resolver when trying to fix l-c
17:09:59 <knikolla> the new dependency resolver will try all versions of a package to try and figure out which one works
17:10:18 <knikolla> and that can take a long time, between fetching, compiling and installing all known versions of a single package
17:10:37 <lbragstad> fun fact - i kicked off a job to do that last week and it took 8 hours
17:10:58 <lbragstad> before i stopped it :)
17:11:36 <knikolla> haha, yeah, i ran into the same issues
17:12:16 <knikolla> if you use a constraints file, you have to pin almost all the dependencies, otherwise it will try all the possible combinations of what you haven't pinned
17:12:42 <lbragstad> iiuc - that includes transitive dependencies
17:13:41 <cmurphy> i think including all transitive dependencies would be a good goal anyway? we've had problems where a dependency that wasn't in the constraints file broke the build and had to be added
17:14:22 <cmurphy> but i don't mind going with the flow if other projects are dropping that job
17:14:48 <knikolla> that would incur a maintenance burden to keep
17:15:22 <knikolla> ideally, i understand and would like such a thing
17:15:36 <lbragstad> i don't have an opinion - i just noticed it's going to take a while to figure out what sane list of dependencies we need (including transitive) to get things working again
17:15:37 <knikolla> but i feel that in a world where everything is containerized, that matters less nowadays
17:16:06 <lbragstad> i could see it making an different for deployments that need to verify versions of software for security reasons
17:16:09 <lbragstad> difference*
17:16:42 <lbragstad> (e.g., we can't deploy something until we comb through the source)
17:17:45 <knikolla> hmmm, i feel that in that scenario they would have the manpower to maintain their own pinned versions
17:17:57 <lbragstad> yeah - that's probably true
17:19:13 <knikolla> for now, we can disable the l-c job in stable branches that are broken, and see how the openstack-wide discussion progresses
17:22:07 <knikolla> #topic oslo.db 8.5 update breaks keystone
17:22:28 <knikolla> prometheanfire reported that in the -keystone channel
17:22:28 <bnemec> I'm in another meeting, but it looks like it's failing on the deprecation warning?
17:23:17 <knikolla> bnemec: do you have a link to zuul logs?
17:25:36 <knikolla> i have a spare hour after the meeting to do some investigating
17:28:15 <knikolla> we can circle back to the topic during open discussion
17:28:54 <knikolla> #topic PTL in Xena
17:28:58 <bnemec> Sorry, I naturally got called on in my other meeting. :-)
17:29:10 <bnemec> The logs are here: https://zuul.opendev.org/t/openstack/build/93390590a90f4ff59b3dc4d008371c41
17:29:24 * bnemec runs away from the PTL topic
17:29:45 <knikolla> Lately, I have been unable to devote the required amount of time and attention to keystone
17:30:04 <knikolla> I am not sure if I will (or should) be continuing in the role during the Xena cycle
17:31:37 <lbragstad> thank you for putting in the time you have, knikolla
17:31:55 <cmurphy> ++
17:32:04 <gagehugo> ^
17:32:12 <gagehugo> ++ thanks knikolla!
17:32:28 <knikolla> Thank you all. I appreciate that.
17:33:06 <knikolla> I wish I could have been doing a much better job as I owe a lot to keystone and OpenStack.
17:33:36 <knikolla> If anyone is interested in the role, talk to me. Otherwise, we can experiment with distributed leadership, or figure something else out.
17:36:33 <bnemec> If you have questions about DPL, I know a guy. ;-)
17:38:12 <knikolla> the tc resolution is here https://governance.openstack.org/tc/resolutions/20200803-distributed-project-leadership.html :)
17:39:47 <knikolla> bnemec: oslo is using distributed leadership right?
17:40:01 <bnemec> Yeah, we started this cycle.
17:41:17 <knikolla> cool
17:41:32 <knikolla> how is that working?
17:42:58 <bnemec> Pretty well so far.
17:43:08 <bnemec> Herve has taken over a lot of the PTL in duties in practice.
17:43:27 <bnemec> The rest of the liaisons mostly say "nothing to report" every week in the meeting. :-)
17:43:36 <knikolla> i see
17:43:53 <bnemec> Err, that should have been "PTL duties in practice" s/in//
17:45:24 <knikolla> alright, circling back to the oslo.db 8.5 topic
17:45:37 <knikolla> #topic oslo.db 8.5 update breaks keystone (part 2)
17:45:44 <knikolla> #link https://zuul.opendev.org/t/openstack/build/93390590a90f4ff59b3dc4d008371c41
17:46:03 <knikolla> DeprecationWarning: Using function/method 'db_version()' is deprecated in version '8.3.0': sqlalchemy-migrate support in oslo_db is deprecated; consider migrating to alembic
17:46:42 <knikolla> it does seem to be related to the deprecation of sqlalchemy-migrate support
17:47:02 <bnemec> It's the only thing I see in the logs that looks at all wrong.
17:47:15 <knikolla> yeah
17:47:42 <knikolla> i am unfamiliar with the amount of work required to migrate to alembic from sqlalchemy-migrate
17:47:58 <knikolla> i know we had a person working on it a while back, vishakha
17:48:35 <knikolla> we have a backlogged spec at https://specs.openstack.org/openstack/keystone-specs/specs/keystone/backlog/alembic.html
17:50:13 <knikolla> can we non-error on deprecation warnings for now while we figure this out?
17:52:28 <bnemec> That seems like the immediate path forward. I'm not immediately aware of what makes the deprecation warning fatal though.
17:54:19 <knikolla> i am not aware either
17:56:38 * lbragstad has to drop early
17:57:14 <knikolla> guess i'll debug while i cook lunch, thanks all! thanks bnemec for your time.
17:57:27 <bnemec> np
17:57:59 <knikolla> #endmeeting