18:00:14 <smooge> #startmeeting EPEL
18:00:14 <zodbot> Meeting started Wed May 25 18:00:14 2016 UTC.  The chair is smooge. Information about MeetBot at http://wiki.debian.org/MeetBot.
18:00:14 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
18:00:14 <zodbot> The meeting name has been set to 'epel'
18:00:14 <smooge> #meetingname EPEL
18:00:14 <zodbot> The meeting name has been set to 'epel'
18:00:14 <smooge> #chair smooge nirik Evolution bstinson avij
18:00:14 <smooge> #topic aloha
18:00:14 <zodbot> Current chairs: Evolution avij bstinson nirik smooge
18:00:14 <smooge> #info avij is away on travel
18:00:28 <nirik> morning
18:00:32 <smooge> Hello fellow humans and robots
18:01:01 <bstinson> happy wednesday
18:01:10 <smooge> and we have quorum
18:01:12 <smooge> hi bstinson
18:01:50 <smooge> avij is out for the week.
18:02:06 <smooge> Evolution I guess is recovering his liver from OScon?
18:02:21 <smooge> #topic Outstanding issues
18:02:21 <smooge> #info Orphans and removals
18:02:21 <smooge> #info Recharter EPEL
18:02:21 <smooge> #info Removing EL-5 in 7 months.
18:02:22 <smooge> #info alternative arch issues for CentOS
18:02:46 <smooge> I saw tyll had orphaned a bunch of packages last week
18:02:53 <nirik> yep
18:03:11 <smooge> so I think that issue is 'closed' until next month :)
18:03:44 <smooge> I got stuck on 2 other issues so my time to deal with EPEL items was small to none.
18:04:05 <nirik> it's been a crazy week
18:04:09 <Evolution> amen.
18:04:16 <smooge> I would like to work on setting up a pagure and moving the EPEL items to that from the trac
18:05:06 <nirik> I'd suggest waiting a bit until the importer tool is fully ready...
18:05:10 <smooge> I will be putting the current EPEL pages in it as rst documents that can be 'forked' and updated by others so if others do have time to work on EPEL charter and other things they can fork and make pull requests
18:05:19 <nirik> unless you want to salt the old earth and make a new start
18:05:27 <smooge> well we don't have many current issues in the system
18:05:39 <smooge> I am in the salt the earth mood today
18:05:50 <nirik> ok.
18:06:06 <nirik> if you are willing to do the work, you can decide how to do it... fine with me. ;)
18:06:11 <smooge> salt the earth and spray the napalm :)
18:06:13 <bstinson> at least having the charter and other docs in tehre would be quite nice
18:06:28 <nirik> we can setup a epel group with access/commits
18:06:54 <smooge> yeah I started this in the trac side but it never got very far because pagure looked like a better place :)
18:07:09 <smooge> [cough] revisionist history [cough]
18:07:53 <smooge> I have a 3 day weekend and only need to put up curtains so I am sure I can work on it this weekend [cough]
18:08:16 <smooge> OK epel 5 is going to be EOL in 7 months.
18:08:21 <Evolution> see, by saying that, you're jinxing yourself.
18:08:46 <smooge> OK epel 5 is going to be Johnny Hughes problem in 7 months.
18:08:50 <smooge> there better?
18:09:02 <Evolution> no, I meant about "only" needing to hang curtains.
18:09:15 <smooge> oh yeah..
18:09:16 <Evolution> we'll all cheer the death of el5 and epel5 together
18:09:56 <smooge> oh I expect it will be like the XP support in Cygwin
18:10:05 <smooge> we will never be able to get rid of it..
18:10:33 <smooge> anyway. last up is alternative arch items for CentOS.
18:11:17 <smooge> Evolution, you posted a couple of things on that this week. That was on aarch64 I think?
18:11:39 <Evolution> yes. we're cycling the builds. it's available now as 'unsigned', use at own risk.
18:11:50 <Evolution> we're not calling it epel, just saying that it's built from epel srpms
18:12:18 <Evolution> once EPEL is available for aarch64 (according to pbrobinson around late august), we'll kill it
18:12:37 * nirik nods
18:12:43 <Evolution> the other builds for armv7, and i686 still need to be worked out.
18:12:57 <yselkowitz> with a few minor exceptions, my changes should hit epel7 stable next week
18:13:32 <yselkowitz> and ghc will take a little while to coordinate
18:13:36 <Evolution> from the initial pass, we had 1500 that failed to build.
18:14:02 <Evolution> I suspect most of those will be resolved just cycling the build again for dep issues.
18:14:06 <yselkowitz> yep
18:14:11 <nirik> yeah.
18:14:15 <yselkowitz> you've got some bootstrapping to do too
18:14:19 <Evolution> yup.
18:14:45 <Evolution> the one thing to be aware of for now that I'll try to resolve for 7.3 is that we don't include ada for gcc
18:14:50 <smooge> do you think that the way you did the builds for aarch64 could work for armv7 and i686?
18:14:51 <Evolution> so that might impact the ghc builds.
18:15:21 <smooge> oooh so that was the ada conversation I was reading about earlier
18:15:40 <Evolution> bootstrapping ada is... ugly
18:16:20 <Evolution> and I chose not to do that for 7.1 or 7.2 for sanity/time reasons.
18:16:28 <Evolution> I'm hoping I'll have it resolved for 7.3
18:16:39 <yselkowitz> ghc != ada
18:17:05 <Evolution> yselkowitz: right. but I thought ada was a dep for some of the ghc bits.
18:17:21 <Evolution> I seem to recall there were 3-4 packages in epel that needed ada.
18:17:49 <Evolution> beyond that we've patched and rev'd the redhat-rpm-config package with yselkowitz's patch
18:17:56 <smooge> cool
18:18:18 <Evolution> so most other things should just work (apart from the occasional bootstrapping)
18:19:41 <smooge> Evolution, I remember for one of the rebuild loops it took about 3 rebuilds and then a pull in of some temp fedora packages to get ppc done.. but I am only dimmly remmebering other people talking
18:19:53 <Evolution> smooge: have you had a chance to propose your point-release update yet? tying in rebases to the version bumps?
18:20:32 <Evolution> smooge: we're doing passes until nothing succeeds. then the bootstrap/debug starts.
18:20:46 <Evolution> easier to debug 150 packages than 1500.
18:21:19 <yselkowitz> lucky for you we did most of the work already :-P
18:21:28 <smooge> The point-release needs to be finished. I have been tying it in with the recharter
18:22:19 <Evolution> yselkowitz: I know, it's nice. thanks for making us look good!
18:22:20 <Evolution> :-P
18:22:20 <smooge> so let us see where it is on Tuesday next week
18:23:02 <smooge> anything else on the alternative arches?
18:23:42 <Evolution> for now, no.
18:24:18 <smooge> okie dokie.
18:24:30 <smooge> #topic Current opportunities
18:24:30 <smooge> #info CentOS-6.8 released. Rebuilds of packages and upgrades
18:24:30 <smooge> #info dnf needs in EL-7
18:24:30 <smooge> #info mirrormanager problems
18:24:30 <smooge> #info ???
18:24:46 <smooge> Congrats to the CentOS team on getting 6.8 out the door
18:24:54 <smooge> if I read the emails today correctly
18:25:02 <Evolution> you did.
18:25:41 <smooge> This is the last major change release in the 6 series. So no more rebases expected for .9 [oooh like it iddnt happen in 5.9..]
18:26:01 <Evolution> "in theory"
18:26:19 <nirik> In theory, theory and practice are the same. In practice, they aren't.
18:26:24 <smooge> The dnf item actually is erroneous. I think nirik and dgilmore dealt with it completely last week
18:26:37 <nirik> they withdrew the updates and blocked the packages.
18:26:48 <nirik> I don't know what their plan is moving forward tho, no one talked to me at least
18:27:06 <smooge> I think their plan is to see if they can get it in CentOS extras or something
18:27:06 <Evolution> so, I have some insight on this one but the short answer is "it's complicated"
18:27:18 <smooge> ok go ahead
18:28:19 <Evolution> they want to bump the libsolve bits, and get dnf and a yum compat package into an optional repository, giving folks the option to test.
18:28:40 <Evolution> it can't go into extras directly, because extras is enabled by default, and that breaks compat
18:28:53 <smooge> ah ok makes sense
18:29:09 <Evolution> however, there's a possible impact with mock for folks who build across fedora, centos, and epel.
18:29:18 <Evolution> er, s/epel/rhel/
18:29:34 <Evolution> but only if they're using centos or rhel as the dev platform.
18:29:58 <Evolution> that userbase is *probably* rather small, but still important.
18:30:40 <Evolution> but since epel is built against rhel, it's an exercise we'd need to walk through to be sure we don't cause breakage.
18:31:41 <smooge> got it
18:31:57 <Evolution> first step is figuring out what gets dropped into the addon repo, and how things are named for compat with existing users.
18:32:24 <Evolution> I need to get with langdon and blc to start working on this one a bit.
18:33:22 <Evolution> end-of-fairytale
18:34:44 <smooge> okie dokie
18:35:46 <smooge> Alright on the mirrormanager issues. We had some earlier this week because another program ate all the database connections
18:36:43 <nirik> that should have been a pretty short window...
18:36:44 <smooge> However, I was able to 'replicate' the problem of timeouts outside of that on Monday where i was getting timeouts but our nagios didn't trigger. However I couldn't figure out why it was happening
18:37:10 <smooge> by the time I got to debugging, the problem went away on my end
18:37:42 <smooge> then later it came back but again.. by the time I could get to debugging. poof fixed
18:38:39 <nirik> odd. note that db connections shouldn't affect mirrorlists directly
18:38:52 <nirik> mirrorlists pull data from a local static pkl file.
18:39:45 <smooge> correct. something is going on somewhere and I am not sure what yet. I will work with the CentOS people who are seeing it more often to try and pin it down.
18:41:35 <smooge> anyway.. it is something I will try to figure out soon.
18:41:39 <smooge> ok
18:41:43 <smooge> #topic Open Floor
18:41:50 <nirik> sure, would be good to make sure we fix anything like that
18:42:46 <smooge> i am mostly grumpy on it that I can see it but can't see it at the same time :)
18:43:41 <smooge> ok anyone have any open floor issues? if not I will close out and go work on stuff for my progress report
18:44:00 * nirik has nothing for open floor
18:45:06 <smooge> #endmeeting