20:32:05 <prometheanfire> #startmeeting requirements
20:32:06 <openstack> Meeting started Wed Dec 12 20:32:05 2018 UTC and is due to finish in 60 minutes.  The chair is prometheanfire. Information about MeetBot at http://wiki.debian.org/MeetBot.
20:32:07 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
20:32:10 <openstack> The meeting name has been set to 'requirements'
20:32:15 <prometheanfire> #topic rollcall
20:32:36 <prometheanfire> tonyb, prometheanfire, number80, dirk, coolsvap, toabctl, smcginnis, dhellmann
20:32:40 <prometheanfire> ping
20:32:42 <prometheanfire> o/
20:36:58 <dhellmann> o/
20:37:08 <prometheanfire> net may be going down
20:37:45 <prometheanfire> #topic open issues in queue?
20:38:31 <prometheanfire> only thing is libvirt and they know what to do
20:38:44 <prometheanfire> there are some open things that could use votes though
20:39:15 <tonyb> \o
20:39:18 <tonyb> sorry I'm late
20:40:01 <tonyb> just libvirt as you say but I don't think that's urgent and tbarron and hodgepodge seem motivated to get it in
20:40:18 <prometheanfire> yep
20:40:41 <prometheanfire> #topic open floor
20:40:57 <dhellmann> is https://review.openstack.org/#/c/623988/1 still open just because it's a stable patch?
20:41:15 <dhellmann> I guess it's only 2 days old, but it's a constraint update
20:41:43 <dhellmann> I don't see any comments indicating a problem, so I was going to approve it if that's ok
20:41:56 <tonyb> dhellmann: Yup because it's stable, I just approved it
20:42:05 <dhellmann> ok
20:42:12 <dhellmann> does the failure on https://review.openstack.org/#/c/624759/ makes any sense to either of you?
20:42:22 <smcginnis> Do we need to get the libvirt stuff written down somewhere?
20:42:31 * smcginnis tries to catch up
20:42:31 <tonyb> dhellmann: Nope, that one is goign to require digging
20:42:45 <prometheanfire> smcginnis: I'd rather use gate jobs for the libvirt stuff
20:42:56 <smcginnis> Gate jobs as documentation?
20:43:06 <tonyb> smcginnis: I'm planning on turning the conversation into a doc once we've worked it through to conclusion
20:43:28 <smcginnis> It just seems there's recurring confusion around how to handle that and requires a discussion every time it happens.
20:43:28 <tonyb> I'm kinda expecting somethign new to pop up once we get to the next step
20:43:30 <prometheanfire> smcginnis: someone complains about failures, point them to update the gate job for their new broken OS
20:43:31 <smcginnis> tonyb: ++
20:43:37 <tonyb> smcginnis: Yup
20:43:44 <dhellmann> https://review.openstack.org/#/c/624048/ is a request to add a new thing to the stable/queens requirements, which seems controversial to me
20:43:59 <prometheanfire> dhellmann: ya, iirc new feature dev is done there?
20:44:00 <prometheanfire> tonyb: ?
20:44:01 <dhellmann> although maybe not highly so?
20:44:16 <prometheanfire> not highly so from my pov
20:44:17 <dhellmann> prometheanfire : yeah, we support feature work in stable versions of deployment tools
20:44:23 <tbarron> tonyb: it's not lack of motivation, it's being swamped by the other job(s)
20:44:30 <dhellmann> oh, although that's a neutron driver
20:44:41 <prometheanfire> hmm, OSA has been doing feature backports for a long time
20:44:52 <prometheanfire> tbarron: ack
20:44:53 <dhellmann> "networking-ansible ML2 driver invokes ansible..."
20:44:56 <tbarron> tonyb: sorry to respond out of context but my ircs are delayed due to thrashing and deadlines
20:44:57 <tonyb> tbarron: Yup, I fuigured as much.  I just meant it seems to have landed on you :(
20:45:34 <tonyb> dhellmann:, prometheanfire: I'll need to look at that ansible-runner review.  It is controversial but perhaps there are mitigating circumstances
20:45:53 <tonyb> tbarron: np
20:45:54 <tbarron> people want me to say what we did in this sprint and shit like that, and plan rfes for future 3 month releases, rather distracting
20:46:02 <dhellmann> tonyb : I've just asked for a ML thread
20:46:17 <prometheanfire> as far as capping grpcio I don't like it, I'd rather mask known bad versions until upstream is unbroken
20:47:06 <tbarron> tonyb: but I will go throug the livirt doc you pointed me to and try to figure from zuul what the various distros max is for each stable/* branch
20:47:06 <prometheanfire> or at least have a review in queue to uncap it
20:47:10 <tbarron> it's not in the doc
20:47:35 <tonyb> tbarron: Ok, it's possible that there isn't a max and I'm wrong on that point
20:47:41 <dhellmann> prometheanfire : I think before we give that feedback we probably want to fix the bug in the validate job, but I don't disagree
20:47:49 <tonyb> dhellmann: cool, comments passing in the ether ;P
20:47:58 <prometheanfire> dhellmann: I thought I was mistaken about the bug
20:48:06 <prometheanfire> unless you found something new
20:48:17 <tbarron> i need to learn in decentralized zuulv3 world how to find for each stable/* what release of each nodeset runs
20:48:24 <tbarron> unless there's a better way
20:48:25 * prometheanfire was going on little sleep for that email
20:48:28 <dhellmann> prometheanfire : the job failed reporting that a constraint that isn't present in the constraints list does not match the requirements settings
20:48:41 <dhellmann> unless there's already a constraint line there and bnemec added a new one instead of updating the old one
20:48:48 <prometheanfire> dhellmann: https://review.openstack.org/#/c/624759/ ?
20:48:55 <dhellmann> yes
20:49:02 <prometheanfire> ah ok, was thinking something else
20:49:19 <tonyb> tbarron: I'd probably just pick a job on the branch and look at the logs ?
20:49:28 <bnemec> I don't see an existing entry in g-r.
20:49:52 <dhellmann> bnemec : line 53 of upper-constraints.txt
20:49:57 <dhellmann> I left a comment
20:49:59 <tbarron> tonyb: i don't know e.g. where to find a job thar runs with gentoo on stable/queens
20:50:04 <bnemec> Oh, u-c, not g-r.
20:50:10 <bnemec> Fixing...
20:50:11 <prometheanfire> dhellmann: beat me to it :D
20:50:18 <tbarron> tonyb: for ubuntu and centos it's easy
20:50:24 <tonyb> it *probably* measn that something higher up already required grpcio and is grabbing the latest version
20:50:47 <prometheanfire> tonyb: it was defined earlier in UC
20:50:50 <dhellmann> tonyb : yeah, grpcio was already in the constraints list
20:51:20 <tonyb> tbarron: Ther wont be a job for gentoo, prometheanfire .... what OpenStack releases do you have in gentoo and which libvirt versions
20:51:40 <tonyb> Oh okay validate not check my bad
20:51:44 <prometheanfire> ya, rocky/queens (I'm removing pike soon)
20:51:45 <tbarron> tonyb: ah, you mentioned gentoo and SUSE
20:51:54 <prometheanfire> 4.5.0-r1(0/4.5.0){xpak} ~4.9.0(0/4.9.0)
20:51:57 <prometheanfire> for libvirt
20:52:03 <tbarron> tonyb: and if I look at nodesets in zuul there could be others
20:52:07 <prometheanfire> 4.9.0 is not marked stable yet
20:52:13 <tonyb> tbarron: I did we 'support' them but I don't think we have nodepool images yet
20:52:33 <prometheanfire> as a general note, I've never capped libvirt or constrained it in my gentoo packages for this reason
20:52:40 <prometheanfire> it's an exception
20:52:47 <prometheanfire> I do set a minimum though
20:52:57 <prometheanfire> haven't had problems (myself) or bug reports
20:52:59 <tonyb> prometheanfire: when would a user install 4.5.0-r1 vs 4.9.0 ? is it there choice?
20:53:13 <prometheanfire> default for amd64 is 4.5.0-r1
20:53:23 <tbarron> tonyb: kk, guess I'll ask prometheanfire for gentoo, zigo for debian, toabctl for suse, for each stable/*
20:53:41 <openstackgerrit> Ben Nemec proposed openstack/requirements master: Cap grpcio to below 1.16.0  https://review.openstack.org/624759
20:53:41 <prometheanfire> user can keyword 4.9.0 and install that if they want
20:53:41 <openstackgerrit> Ben Nemec proposed openstack/requirements master: Remove daiquiri from global-requirements  https://review.openstack.org/624760
20:53:48 <tonyb> Okay tbarron so gentoo is 4.5.0 which may be the new minimum version?
20:54:00 <prometheanfire> iirc cent is 4.5.0 right?
20:54:04 <tonyb> prometheanfire: Okay that's cool that answer the question at hand
20:54:13 <tonyb> tbarron: would know
20:54:26 <prometheanfire> tonyb: we constantly update, so don't use us for the min
20:54:28 <tbarron> tonyb: prometheanfire: thx, centos is currently also 4.5.0
20:54:44 <tonyb> cool
20:54:46 <prometheanfire> generally when a libvirt bug comes out we will stablize the new and remove the old versions as a matter of course
20:55:19 <prometheanfire> anything else (I think we've beaten this dead horse)
20:55:21 <prometheanfire> ?
20:55:21 <tonyb> bnemec: Thanks for going for the bonus points \o/ #awesome :)
20:56:10 * tonyb will find a suse person to identify the same datappint for suse
20:56:16 <tonyb> dirk: seems busy
20:56:20 * bnemec is the teacher's pet :-)
20:56:36 * prometheanfire pats bnemec on the head
20:56:41 <tbarron> tonyb: thx, toabctl last updated the libvit doc you pointe me to
20:57:00 <tbarron> libvirt
20:57:31 <tbarron> and he's back from paternity leave, nice one they get in DE
20:58:04 <tonyb> toabctl, dirk, jhesketh: Any chance one of you can confirm which OpenStack version ya'll support (in SLES and OpenSuse) and which version of libvirt ship with them and what, if any, impact us bumping the libvirt-python version will have on you
20:58:52 <prometheanfire> will wait another min before ending the meeting
21:01:21 <prometheanfire> #endmeeting