15:01:00 <apevec> #startmeeting RDO packaging meeting (2015-06-17)
15:01:00 <zodbot> Meeting started Wed Jun 17 15:01:00 2015 UTC.  The chair is apevec. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:01:00 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:01:07 <apevec> #topic rollcall
15:01:09 <v1k0d3n> it doesn't seem like the signing key is available?
15:01:11 <apevec> o/
15:01:27 <number80> apevec: if we host other meetings +1
15:01:41 <number80> if not, i suggest reusing either fedora or centos channels
15:01:47 <blinky_ghost> slagle: and eventually the stack times out
15:01:54 <jpena> o/
15:01:55 <rbowen> apevec: Consensus so far has been that putting it off on a different channel would make it less visible, and thus decrease community participation. However, I don't feel strongly either way
15:01:57 <number80> o/
15:02:09 <number80> for one meeting, it's much overhead
15:02:14 <apevec> #info agenda at https://etherpad.openstack.org/p/RDO-Packaging
15:02:41 <mburned> 0/
15:02:54 <apevec> number80, yeah, I'm just thinking that blocking channel for 1h is not so nice  :) but yeah visibility trumps it
15:02:56 <mburned> err... o/
15:03:06 <social> v1k0d3n: fixed in newer opm
15:03:07 <apevec> mburned, headache? :)
15:03:17 <apevec> #chari number80 mburned jpena rbowen
15:03:26 <mburned> apevec: something like that...
15:03:32 <v1k0d3n> social: sorry, newer opm?
15:03:41 <apevec> #chair number80 mburned jpena rbowen
15:03:41 <zodbot> Current chairs: apevec jpena mburned number80 rbowen
15:04:04 <apevec> let's keep meeting short today, only two topics on agenda!
15:04:11 <social> v1k0d3n: rabbitmq error is from puppet-rabbitmq which is probably delivered to you by openstack-puppet-modules package
15:04:13 <apevec> #topic Migrate delorean to F22
15:04:17 <v1k0d3n> i just ran through install (RDO: Kilo) fresh on install just a few minutes ago.
15:04:25 <apevec> jpena, derekh ^ who proposed that? Any blockers?
15:04:35 <number80> apevec: me from a discussion with mrunge
15:04:46 <social> v1k0d3n: this didn't hit repos yet http://koji.fedoraproject.org/koji/buildinfo?buildID=651289
15:04:51 <apevec> I did notice some issues with fedora-review on F22, not sure what it was, but worked fine on F21 ...
15:04:57 <derekh> apevec: not me
15:04:58 <apevec> so F22 might have stability isssues
15:05:01 <number80> F21 will be EOLed when Liberty will be released
15:05:14 <apevec> number80, ok, so that's not urgent
15:05:28 <v1k0d3n> social: ah, crap. when? is there a bette way to obtain for the workaround? i went to the workarounds page on RDO and didn't see this error.
15:05:28 <number80> *nods*
15:05:38 <apevec> I vote we keep it on F21 for now, but start RDO Liberty testing repo early
15:05:40 <apevec> for new deps
15:05:41 <v1k0d3n> social: sorry...meaning when will it hit repos?
15:06:02 <social> v1k0d3n: that's more of a question for number80 and apevec, sorry
15:06:04 <number80> +1 for delaying it
15:06:10 <jpena> +1 for delaying
15:06:41 <apevec> social, OT but have you pushed Rawhide build?
15:06:58 <apevec> I can push rdo update after the meeting
15:07:15 <apevec> #info delay Delorean to F22 migration
15:07:34 <apevec> #action apevec switch Delorean Trunk to early RDO Liberty testing repo asap
15:07:47 <apevec> that's worth separate topic actually
15:08:03 <apevec> #topic Liberty-1 prep
15:08:19 * apevec looks up schedule
15:08:35 * eggmaster waves
15:08:37 <apevec> liberty-1 (Jun 23-25)
15:08:43 <v1k0d3n> social: thanks for the heads up. so should i override the repo opm in favor of this updated rpm?
15:08:45 <number80> next week
15:08:59 <apevec> and Branch Fedora 23 is 2015-07-14
15:09:09 <apevec> so we need to fake distgit somewhere
15:09:12 <v1k0d3n> social: thanks for the heads up btw...really appreciate that you guys are all over this already.
15:09:14 <social> apevec: apevec https://review.gerrithub.io/#/c/236794/ ?
15:09:40 <apevec> so what about creating liberty-1 branch in github/openstack-packages as a preparation?
15:09:52 <apevec> and when f23 branch is ready, we import it
15:10:19 <apevec> social, OT but ACK, I'll add CBS builds and push it
15:10:56 <apevec> number80, jpena - any other suggestions?
15:11:23 <number80> apevec: why not rpm-liberty?
15:11:56 <number80> btw, alphacc is in PTO, his backup for CBS is bstinson
15:11:56 <apevec> hm, maybe - I wanted to make it clear it's unofficial
15:12:26 <number80> rpm-liberty-1 N
15:12:29 <apevec> rpm-liberty would be normally created after liberty GA
15:12:34 <apevec> ok
15:12:42 <apevec> rpm-liberty-1 sold
15:12:46 <number80> i prefer keeping branch naming consistent :)
15:12:51 <apevec> yeah
15:13:03 <apevec> thanks for alphacc backup info
15:13:18 <apevec> that was next question, I need to get liberty tags created in CBS
15:13:40 <apevec> number80, so I just open centos bug and ping bstinson ?
15:14:08 <number80> yup
15:14:11 <apevec> ok
15:14:48 <apevec> #action apevec to  open centos bug for liberty tags in CBS and ping bstinson
15:15:14 <apevec> #action apevec to mass-create rpm-liberty-1 branches to test liberty-1 distgit merge
15:15:36 <apevec> ^ that's what we have as "synced once per cycle from delorean" in https://jruzicka.fedorapeople.org/pkgs/rdo-workflow.jpg
15:15:56 <apevec> but I want to test/automate it with liberty-1 now
15:16:08 <apevec> so we don't end  up with backlog at the end of cycle
15:17:13 <apevec> ok, we'll check status on this next week
15:17:43 <apevec> #topic summary from https://etherpad.openstack.org/p/openstack-rpm-packaging meeting
15:18:08 <apevec> on that meeting it was me number80 and derekh from RDO and two SUSE guys
15:18:17 <apevec> my take aways:
15:18:35 <apevec> - rdo should proceed with upstream import but into rdo-* not rpm-* namespaces
15:19:22 <apevec> - we'll start greenfield specs repo (one for all?) using templates (see Anvil examples?) starting with oslo libraries
15:19:34 <apevec> number80, derekh ^ anything more to add from your side?
15:19:51 * rbowen tries to catch up ...
15:20:11 <jruzicka> +1
15:20:21 <jruzicka> oh, sry
15:20:29 <apevec> rbowen, nobody asked about middleman migration yet :)
15:20:47 <derekh> Yup, thats about it, I'll pick up again on the rdo upstream import to get the ball rolling
15:20:53 <number80> apevec: not really, except that we far more advanced than other groups to move upstream :)
15:20:56 <rbowen> Alas, it's delayed yet again, due to the folks in front of us in the queue having troubles.
15:21:24 <apevec> derekh, do we actually need to import to openstack/ or stackforge/ would be enough?
15:21:44 <apevec> derekh, given that Ubuntu is pulling off, I don't think there'll be TC support?
15:21:57 <derekh> apevec: as far as I can tell fungi wants it in  openstack/
15:22:08 <pabelanger> yes
15:22:17 <pabelanger> -infra does want openstack/
15:22:18 <apevec> rbowen, just keep them in front of us, they're our testing bunnies :)
15:22:23 <derekh> apevec: we can aim for that and see what the TC says
15:22:24 <rbowen> +1
15:22:41 <apevec> derekh, pabelanger, ok, let's how it goes
15:22:47 <rbowen> Unfortunately, it means that various website things are on hold because I don't want to do them in the wiki and then have to do them again.
15:23:29 <apevec> #action derekh to pick up again on the rdo upstream import
15:23:36 <derekh> apevec: its the rdo packaging team so I don't see the absence of other distros doing something similar effecting it much
15:24:12 <apevec> derekh, ack
15:24:19 <pabelanger> apevec, derekh actually, if you didn't see both Ubuntu / Debian are pulling pack from the upstream packaging effort
15:24:24 <pabelanger> things fell a part for them
15:25:04 <jruzicka> official reasons being?
15:25:07 <apevec> #topic open floor
15:25:19 <derekh> pabelanger: thanks must catch back up on the thread but I think we can try to plough on
15:25:38 <pabelanger> derekh, Right. I think this is a good chance for RPMs to shine :)
15:25:48 <derekh> pabelanger: yup
15:25:49 <mburned> number80: did that common repo refresh happen?
15:25:54 <number80> +1
15:26:01 <apevec> mburned, which refresh?
15:26:13 <mburned> apevec: we were missing a couple packages a couple weeks ago
15:26:30 <number80> mburned: yes, but fio and sysbench were untagged since :/
15:26:34 <mburned> packages built and tagged properly, but was awaiting signing and update
15:26:39 <pabelanger> jruzicka, http://lists.openstack.org/pipermail/openstack-dev/2015-June/066942.html
15:26:39 <number80> heat-cfntools were pushed
15:26:40 <mburned> number80: why?
15:26:42 <pabelanger> explains the reasoning
15:26:46 <number80> mburned: no idea :(
15:26:50 <apevec> we don't have separate common repo, it's just subfolder in kilo repo
15:27:04 <v1k0d3n> social: thanks for the tip...after updating the opm rpm, the installation went perfectly!
15:27:06 <number80> mburned: I'm planning to ask a new refresh since I fixed a CVE in cinder
15:27:10 <apevec> number80, which NVRs ?
15:27:28 <apevec> look at tag history
15:27:37 <number80> apevec: fio-2.2.8-1.el7
15:28:23 <v1k0d3n> apevec: when will the kilo repo be updated with opm-2015.1.6-2.fc23? i can tell you the installation went perfectly for centos7 once i updated that rpm in yum.
15:28:43 <apevec> v1k0d3n, as I said, after the meeting
15:29:00 <v1k0d3n> oh, i'm so sorry! i missed it...was working on other things too.
15:29:04 <v1k0d3n> sorry for the repeat
15:30:05 <apevec> number80, mburned - http://paste.openstack.org/show/298653/
15:30:11 <apevec> it was never tagged -testing ?
15:30:55 <number80> apevec: I remember it was
15:31:09 <apevec> Koji never lies :)
15:31:21 <number80> :/
15:31:40 <apevec> number80, unless it was restored from backup??
15:32:04 <number80> maybe, i'll ping bstinson
15:32:05 <mburned> apevec: perhaps not
15:32:22 * mburned wasn't sure of the process
15:33:04 <apevec> for now, cbs tag-build cloud7-openstack-{kilo|common}-testing $NVR
15:33:09 <number80> anyway, re-tagging them and asking for kilo update
15:33:14 <number80> *repo
15:33:21 <apevec> I need to figure out how to integrate this with rdopkg update
15:34:27 <apevec> ok, any other topics?
15:34:53 <number80> atm, no
15:35:13 <number80> but integrating rdopkg update w/ CBS is top priority for liberty
15:35:22 <apevec> yep
15:35:42 <number80> do we have a box for cron tasks ?
15:35:54 <number80> I plan reusing debtor for RDO
15:36:10 <number80> so that we can get regular reports about downstream patches
15:36:13 <apevec> number80, try instance on OS1
15:36:16 <number80> ack
15:36:26 <apevec> dogfooding :)
15:36:59 <apevec> number80, but we should not have many in rdo, only horizon has more b/c RDO has non-upstream theme
15:37:09 <apevec> but that's mrunge's choice
15:37:24 <apevec> he doesn't other upstream to test it
15:37:30 <number80> we still have few pbr/oslosphinx patches in kilo ...
15:37:46 <number80> planning to kill them all during liberty full review
15:37:49 <apevec> really?? Drop them asap!
15:37:53 <apevec> ack
15:38:12 <number80> yes, debtor was the one who caugh these :)
15:38:16 <apevec> number80, you'll propose those cleanup in openstack-packages rpm-master ?
15:38:22 <number80> yes
15:38:39 <number80> liberty should have none
15:38:43 <apevec> cool, we could invite debtor author to the next meeting to introduce his nice tool
15:38:53 <number80> flepied^
15:39:10 <number80> I think we can wrap up :)
15:39:11 <apevec> number80, bribe^Winvite him please
15:39:17 <apevec> yep
15:39:19 <apevec> #endmeeting