15:38:31 <dgilmore> #startmeeting RELENG (2014-07-07)
15:38:31 <zodbot> Meeting started Mon Jul  7 15:38:31 2014 UTC.  The chair is dgilmore. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:38:31 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:38:35 <dgilmore> #meetingname releng
15:38:35 <zodbot> The meeting name has been set to 'releng'
15:38:35 <dgilmore> #chair dgilmore nirik tyll sharkcz bochecha masta
15:38:35 <zodbot> Current chairs: bochecha dgilmore masta nirik sharkcz tyll
15:38:36 <dgilmore> #topic init process
15:38:42 * nirik waves. morning
15:39:00 <tyll> Hi
15:39:14 * sharkcz is here
15:40:49 * nirik goes to grab some more coffee, back in a min
15:40:55 <masta> hi
15:44:25 <dgilmore> hi
15:44:37 <dgilmore> sorry in a bunch of conversations right now
15:45:09 <dgilmore> #topic #5931 [Proposal] Move new branch and unretire requests to pkgdb2
15:45:13 <dgilmore> https://fedorahosted.org/rel-eng/ticket/5931
15:47:08 * pingou has nothing new on this for this week
15:47:11 <dgilmore> okay
15:47:19 * pingou has planned to send an email to the list but, still haven't :)
15:47:37 <dgilmore> #topic #5870 rawhide signing
15:47:46 <dgilmore> tyll: how is this going?
15:48:03 <tyll> nothing new here as well, I still cannot reach the secondary signers from autosign01
15:48:24 <tyll> nirik: did RHIT do something?
15:48:43 <nirik> I noted that the change wasn't working, but I haven't seen any action taken yet from that.
15:48:59 <nirik> I'll try and bug folks in the next few days ago
15:49:00 <nirik> again
15:49:30 <dgilmore> nirik: thanks
15:49:40 <dgilmore> lets move on
15:50:03 <dgilmore> #topic #5914 Move fedmsg based blocking service to Fedora Infrastructure
15:50:08 <dgilmore> https://fedorahosted.org/rel-eng/ticket/5914
15:50:19 <tyll> nothing new here from me as well :-(
15:50:33 <dgilmore> :( okay, whats needed?
15:51:36 <tyll> currently it was mostly lack of time, but the next step would be to decide which koji key to use or to generate one, then the script could at least run as it is in Fedora Infrastructure
15:52:57 <dgilmore> tyll: okay, we could either reuse one that exists like masher, or bodhi. or make a new one
15:53:09 <dgilmore> pkg-blocker or something like it
15:54:22 <dgilmore> okay, lets move in
15:54:23 <dgilmore> on
15:54:30 <dgilmore> #topic Secondary Architectures updates
15:54:31 <dgilmore> #topic Secondary Architectures update - ppc
15:54:37 <dgilmore> masta: you're up
15:54:41 <masta> ok
15:55:03 <masta> nothing to report really, I'm not aware of any new developments.
15:55:19 <dgilmore> hows the mass rebuild going?
15:55:31 <dgilmore> ppc64le working well?
15:55:44 <masta> uh, I'm not sure... have not checked since I got back in this morning.
15:56:03 <masta> as far as I know the combined hub is working fine
15:56:08 <dgilmore> okay
15:56:25 <masta> I suspect the rebuild is done, but have no tlooked at the failures yet
15:56:28 <dgilmore> any big blockers?
15:57:09 <sharkcz> not big, but there are some, failing test for gnulib which is part of coreutils/findutils/...
15:57:16 <sharkcz> https://fedoraproject.org/wiki/Architectures/PowerPC/WorkQueue
15:57:36 <dgilmore> sharkcz: okay cheers
15:57:49 <dgilmore> sharkcz: how did ppc64le support get dropped?
15:59:06 <sharkcz> sorry, don't understand, but the workaround was to use a special builder, on one type the hang doesn't occur
15:59:38 <dgilmore> sharkcz: okay. there is some other pretty big issues with systemd-215
15:59:53 <dgilmore> right now composing is broken
16:00:08 <dgilmore> https://bugzilla.redhat.com/show_bug.cgi?id=1116651
16:00:40 <sharkcz> oh yes, systemd, they added some functionality, but didn't for ppc64le
16:01:27 <sharkcz> it's hard to react with an official solution in few hours
16:01:27 <dgilmore> lovely
16:01:53 <sharkcz> it's about debian multiarch triplets ...
16:02:09 <sharkcz> which is not defined for ppc64le yet ...
16:03:57 <masta> oh great
16:04:11 * masta shivers at the mention of triplets
16:04:58 <dgilmore> okay, lets move on
16:05:27 <dgilmore> #topic Secondary Architectures update - s390
16:05:40 <dgilmore> sharkcz: hows things?
16:06:14 <sharkcz> one gcc issue has a fix proposed, second one is still being worked on, so I'll wait for fixed gcc before continuing
16:06:25 <dgilmore> okay cool thanks
16:06:34 <dgilmore> #topic Secondary Architectures update - arm
16:07:04 <dgilmore> so aarch64 is just a little behind build wise
16:07:40 <dgilmore> the current kernels should be working for physical hardware
16:07:48 <dgilmore> but they seem to only be working for some
16:08:13 <dgilmore> but overall aarch64 should be able to follow the branched processes pretty well
16:08:21 <dgilmore> #topic Open Floor
16:08:28 <dgilmore> anyone have anything to talk about?
16:08:31 <nirik> so are we ready for branching?
16:08:37 <dgilmore> nirik: not really
16:08:39 <tyll> I retired a lot of packages
16:08:45 <sharkcz> yep - https://fedorahosted.org/rel-eng/ticket/5933 and https://fedorahosted.org/rel-eng/ticket/5934
16:09:08 <nirik> pingou: pkgdb2 has a mass branching thing right?
16:09:32 <pingou> nirik: it has a script yes
16:09:42 <dgilmore> pingou: is the process documented?
16:10:13 <dgilmore> pingou: http://fedoraproject.org/wiki/Mass_Branching_SOP needs updated
16:10:31 <pingou> https://github.com/fedora-infra/pkgdb2/blob/master/utility/pkgdb2_branch.py this is the script
16:10:41 <pingou> dgilmore: indeed, I'll do that
16:11:01 <masta> sharkcz: so we might have to manually untag it looks like, but not sure the root cause.
16:11:16 <dgilmore> masta: thats not an option
16:11:33 <dgilmore> we will need to work out why the syncing of blocked packages is doing something wrong
16:11:52 <sharkcz> for the f20-updates-testing thing? I guess yes
16:11:58 <pingou> abadger1999: any idea if we looked at pkgdb2branch.py
16:12:01 <pingou> ?
16:12:32 <masta> sharkcz: yes, sry the first ticket with that big list of builds in both tags
16:12:35 <abadger1999> pingou: pkgdb2branch is used for day-to-day new packages so it must work to some extent.
16:12:54 <pingou> abadger1999: cool, thanks :)
16:12:59 <abadger1999> pingou: but it does have a few CLI switches that optimize the mass branching process so those could be broken and we didn't test them.
16:13:09 <sharkcz> masta: ok, there is probably a flow in bodhi worflow in the "right before GA" phase
16:13:27 <dgilmore> sharkcz: bodhi probably needs fixed
16:13:45 <sharkcz> dgilmore: for the sync - you have the logs somewhere, don't you?
16:14:16 <dgilmore> sharkcz: local email.
16:14:41 <dgilmore> sharkcz: likely will need to run manully and add some debugging
16:14:57 <abadger1999> pingou: The --branch-for and -c options of pkgdb2branch are the pathway for mass branch
16:16:01 <sharkcz> dgilmore: I guess the workaround would be to add GConf-dbus in f19 tag in primary and block it then, but knowing the root should be preferred
16:16:34 <dgilmore> sharkcz: yeah, id rather work oyt what exactly is happening
16:16:43 <nirik> oh, one more thing I had... not sure it's really for here, but will mention it: centos7 is almost out, so we might want to look at moving epel7 from beta sometime soon (time permitting).
16:16:54 <sharkcz> dgilmore: maybe it has something with the f19 branch only on secondaries
16:17:44 <sharkcz> like it thinks, f18 blocked, so it should be unblocked in f19 because it exists
16:18:00 <sharkcz> but you will see
16:18:08 <dgilmore> sharkcz: will work it out
16:18:18 <sharkcz> thanks
16:20:16 <dgilmore> i will have to work with pingou on branching
16:20:36 <dgilmore> the branching side of things shouldn't be too hard
16:20:45 <dgilmore> but TC1 is another issue
16:20:47 <pingou> dgilmore: I'll update the doc in the morning and make sure to be free in the afternoon
16:20:57 <dgilmore> pingou: cheers
16:23:48 * dgilmore would like to ask everyone to keep me on task in updating the compose documentation as we do them
16:24:10 <dgilmore> I would like to make sure that at least nirik, and masta can do primary composes
16:24:23 <nirik> sounds good.
16:24:39 <dgilmore> reminds me
16:25:05 <dgilmore> Red Hat hired Bochecha as a releng person, part of his tasks is to help with fedora
16:25:19 * dgilmore is hoping he can work on some of the coding things we need done
16:25:42 <pingou> (he left irc 5 min ago :()
16:26:01 <dgilmore> we have hired another person to backfill ppc, but thats still be finalised as to the details and when he starts. so i wont name him yet
16:26:05 <nirik> let me just say: hurray!
16:26:15 <dgilmore> backfil dwa
16:26:19 <dgilmore> backfill
16:26:36 <dgilmore> so there should be some more active people to help with things
16:26:48 <masta> dgilmore: that sounds great, I'd like to try a compose
16:26:58 <dgilmore> we need to sort out composing for secondary arches in the new world
16:27:20 <dgilmore> ideally I would like everyone active in releng to be able to do composes
16:29:13 <masta> that sounds good.
16:29:36 <masta> we should strive to fully cross train
16:31:30 <dgilmore> indeed
16:32:07 <dgilmore> okay, anyone have anything else they want to cover today?
16:33:51 <dgilmore> #endmeeting