20:31:25 <nirik> #startmeeting EPEL (2010-12-13)
20:31:25 <zodbot> Meeting started Mon Dec 13 20:31:25 2010 UTC.  The chair is nirik. Information about MeetBot at http://wiki.debian.org/MeetBot.
20:31:25 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
20:31:25 <nirik> #meetingname epel
20:31:25 <zodbot> The meeting name has been set to 'epel'
20:31:25 <nirik> #topic init process/agenda
20:31:31 <nirik> anyone around for an epel meeting?
20:32:45 <stahnma> ohhai
20:32:56 <nirik> morning.
20:33:09 * nirik doesn't see tremble
20:33:19 <nirik> who else do we often ping? ;)
20:34:34 <stahnma> that's about it from what I remember
20:34:36 <nirik> stahnma: you have any topics today?
20:34:44 <stahnma> every now and then a peanut gallary person chimes in
20:34:55 <stahnma> well, I have lots of packages with broken deps in el6
20:34:58 <nirik> well, sometimes abadger1999 or dgilmore or rsc or the like chime in, yeah.
20:35:05 <nirik> yes indeed.
20:35:11 <stahnma> but I don't own most of the packages I depend upon
20:35:19 * abadger1999 on vacation -- will disappear shortly.
20:35:22 <rsc> evening.
20:35:34 <nirik> abadger1999: no worries.
20:35:43 <stahnma> greetings
20:35:46 <nirik> so, epel6 status would be a topic.
20:36:59 <nirik> any others?
20:37:38 <nirik> ok, lets go ahead with that then. ;)
20:37:40 <stahnma> Should we have some sort of drive for new contributors in epel?  I still feel like we are vastly undermanned
20:37:53 <nirik> we could, but not sure how to do that. ;(
20:38:10 <nirik> #topic EPEL6
20:38:19 <nirik> #info broken deps reports are going out now.
20:38:32 <nirik> I've been mailing people who have not yet built things I need...
20:38:33 <stahnma> that's been nice
20:38:37 <nirik> so far with limited luck
20:38:38 <stahnma> other than my email filling up :)
20:39:54 <dgilmore> nirik:
20:40:07 <rsc> ?
20:40:51 <nirik> I posted a thing asking for us to clarify which things we consider in the base os and which not, but no one replied to me. ;)
20:41:04 <nirik> dgilmore: was just seeing if you wanted to be around for the epel meeting. ;)
20:41:14 <dgilmore> nirik: sure
20:41:46 <nirik> so, how about this (and will wait for screaming):
20:41:58 <rsc> I also need a package which is obviously not in RHEL.
20:42:08 <dgilmore> nirik: i missed the email
20:43:05 <nirik> Subject: EPEL6 release date + channels
20:43:14 <nirik> we have several options:
20:43:47 <nirik> a) The base os is everything that ships a src.rpm to mirrors in server/server-optional/workstation/workstation optional. We won't override or replace those.
20:44:12 <nirik> or b) the above + we will ship things for arches rhel doesn't, but will try and keep them in sync with rhel at all times.
20:44:46 <nirik> or c) the above + other channels (HighAvailability/etc). Unforntunately we don't have any way to know whats in those for sure.
20:44:53 <stahnma> b will make more people happy, but is also more difficult.
20:45:13 <rsc> c will make even more people happy, but even more difficult, right?
20:45:26 <stahnma> seems likely
20:45:26 <dgilmore> right
20:45:42 <dgilmore> and could make things uninstallable
20:45:47 <nirik> they are not shipping src.rpm's for those other channels like they did in rhel5. ;(
20:45:53 <rsc> nirik: was it you mentioning, that RHEL lacks x86_32 packages in parts?
20:46:05 <nirik> yes. the virt stack is x86_64 only.
20:46:18 <nirik> so, all those and their deps are only x86_64.
20:46:20 <nirik> in rhel
20:46:28 <nirik> but some epel packages need the perl deps, etc.
20:47:05 <nirik> also, none of the above covers cases like pacemaker. ;(
20:48:08 <nirik> anyone have brilliant ideas?
20:48:30 <stahnma> we should try to handle the most common cases first
20:48:44 <stahnma> I am guessing the majority of users of epel are on x86_64
20:48:57 <nirik> probibly.
20:49:03 <stahnma> perhaps we should concentrate on making that as usable as possible before we really pull in the edge cases
20:49:23 <stahnma> I mean, getting x86_64 correct will pull in/fix many problems in x86 and possibly ppc
20:49:30 <stahnma> at least with deps
20:49:34 <stahnma> in some cases
20:49:39 <nirik> well, not really.
20:49:50 <stahnma> hmm
20:49:52 <nirik> there will still be the excluded stuff in i386...
20:49:53 <stahnma> let me rephrase
20:50:04 <nirik> and ppc64 has no workstation / misses a lot of things like java.
20:50:21 <stahnma> right now even x86_64 isn't usable, IMHO.  Let's start with trying to get that usable before worrying about stuff that's only in one arch, etc
20:51:07 <nirik> well sure... we can work on fixing things in x86_64...
20:51:15 <nirik> but someday we will have to decide things. ;)
20:52:21 <stahnma> true
20:52:43 <stahnma> In my opinion, we should work to make the most software available to the most people
20:52:45 <nirik> I'm fine with people thinking about it more and replying on list.
20:52:55 <nirik> I am leaning forward b...
20:53:20 <stahnma> because the more standard packages we have, the less oppurtunity for each site to do it the wrong way we have :)
20:53:23 <nirik> srpms on mirrors with the exception that we can ship something in epel if it's missing an arch we care about
20:53:44 <nirik> yeah.
20:53:44 <stahnma> but that still excludes things like clustering?
20:54:18 <nirik> well, sorta. thats a nice edge case.
20:54:22 <stahnma> ah
20:54:23 <nirik> consider pacemaker.
20:54:24 <stahnma> ok
20:54:30 <nirik> the src.rpm is in server/optional.
20:54:44 <nirik> but the binary package is not in server or server optional.
20:54:54 <nirik> but the pacemaker-docs package is in server/optional
20:55:00 <stahnma> that seems like a bug
20:55:20 * nirik has no idea the rationale there.
20:55:54 <nirik> so, I would say if we went with process b, we could ship it in epel, as it's missing in arches we care about (all of them), but would have to keep it in sync with the src.rpm.
20:56:16 <stahnma> agreed
20:56:54 <nirik> but then we are shipping some functionality that rhel people normally need another channel for.
20:57:53 <nirik> so, I guess lets work on deps and consider more on list for a week?
20:58:07 <stahnma> that's probably ok.  We have much work to do in that space
20:58:42 <nirik> #agreed will ponder more and work on it on list and over the next week.
20:58:51 <nirik> any other epel6 business?
20:59:05 <nirik> we talked last time about 2011-01-11 for going out of beta.
20:59:31 <stahnma> I can tell you that until I get a centos6 box, I am going to have some packages I hold back, because I want to test them thoroughly
20:59:49 <nirik> yeah.
21:00:09 <nirik> #topic Meetings
21:00:25 <nirik> do we want to meet the 20th/27th?
21:00:33 <nirik> or will likely everyone be gone on holidays?
21:01:07 * nirik will be around, can try and meet and see who shows up.
21:01:22 * stahnma may be gone
21:01:24 <stahnma> not sure yet
21:01:29 <stahnma> 20th I should be around
21:02:25 <nirik> ok
21:02:43 <nirik> I guess we will see if people are around or not and meet or not. ;)
21:02:48 <nirik> #topic Open Floor
21:02:51 <nirik> anything for open floor?
21:05:32 <Oxf13> I got nothing :)
21:05:40 <Oxf13> actually i have something
21:05:45 <nirik> ok, lets continue on list / next week.
21:05:47 <nirik> oh, ok
21:05:55 <Oxf13> what resources exist for maintainers to test their EL6 stuff?
21:06:03 <Oxf13> (since centos6 isn't available yet)
21:06:04 <nirik> not much.
21:06:15 <nirik> I have a rhel6b2 test instance.
21:06:27 <nirik> but there's differences between it and final I am pretty sure.
21:06:28 <rsc> Oxf13: no free RHEL6 for EPEL maintainers? :)
21:06:40 <Oxf13> rsc: yes, but not an enjoyable way
21:06:48 <nirik> rsc: it keeps being mentioned, but never happens. ;)
21:06:51 <Oxf13> you can register for an evaluation copy, and submit lots of info to RHT
21:07:26 <rsc> Oxf13: is it still one month evaluation period? Only 12 trials a year to have a free RHEL? ;)
21:07:31 * nirik guesses he could do that for his test instance.
21:07:42 <Oxf13> I do not know what the details are
21:07:55 <Oxf13> I have pushed internally for free RHEL entitlements for EPEL maintainers
21:08:07 <nirik> Oxf13: cool.
21:08:20 <Oxf13> it seemed to be received OK, what's missing is the process to go from "Yes I am an epel maintainer" to "Here is my RHEL entitlement"
21:08:31 <nirik> yeah.
21:08:56 <nirik> we can look up maintainers easily enough, but not sure how 'active' we would want them to be.
21:08:56 <stahnma> Oxf13: if we could close that loop, it would help greatly
21:09:30 <Oxf13> there are people working on it, but I think, as usual, it's part of a larger project, and opaque to those outside of the project :(
21:10:10 <Oxf13> however
21:10:31 <Oxf13> The person involved with the project said that in the mean time, epel maintainers can contact him directly to get RHEL entitlements
21:10:41 <Oxf13> I don't want to share his email address on a public meeting space though
21:10:46 <nirik> cool.
21:10:50 <stahnma> feel free to PM me
21:10:52 <Oxf13> so...  I guess the better way is to contact me
21:10:54 <nirik> yeah, so how do we get that info out?
21:10:57 <Oxf13> and I'll pass your requests along
21:11:03 <nirik> ok, email good? or ?
21:11:03 <stahnma> ok
21:11:08 <Oxf13> email is fine
21:11:17 <Oxf13> my address is public enough, doing this shouldn't hurt much
21:11:18 * nirik notes there are 430 primary mainainers of EPEL packages currently in pkgdb
21:11:25 <Oxf13> haha...
21:11:38 <nirik> how many of those are active is unknown. ;)
21:11:58 <rsc> nirik: just count how many ever fired a build in koji for .el*?
21:12:02 <stahnma> yeah, if we had some type of enrollment process, I would bet that less than half woudl do  it
21:12:21 <stahnma> and if they are RH employees, they probably already have access to RHEL
21:12:33 <Oxf13> so put it in the meeting notes, and on a wiki page somewhere, and lets see what kind of turn out we have
21:12:44 <Oxf13> I'll be checking FAS names to look for EPEL activity
21:12:47 <nirik> #info contact Oxf13 about RHEL entitlement.
21:13:20 <nirik> it might work such that we could have a vetting process, then add people to a fas group... but perhaps thats too much overhead.
21:13:39 <Oxf13> yeah, lets go lo-fi for now
21:13:52 <Oxf13> once the RHT backend is ready for doing some sort of automation we can look at that on ourside
21:13:55 <nirik> cool. Thanks Oxf13
21:14:00 <halfline> well rhn has an organizational system already i think
21:14:27 <halfline> so presuambly you'd have one person deemed the org admin and he would dole out entitlements to the appropriate people
21:15:50 <nirik> whatever works. ;)
21:16:25 <nirik> ok, anything further? Or shall we close out the meeting?
21:18:02 * nirik will close out in a minute.
21:19:12 <nirik> thanks everyone.
21:19:15 <nirik> #endmeeting