Meeting Log - 2008-10-16

Ricky Zhou ricky at fedoraproject.org
Thu Oct 16 20:38:38 UTC 2008


19:59 -!- mmcgrath changed the topic of #fedora-meeting to: Infrastructure -- Who's here?
19:59 -!- giallu [n=giallu at fedora/giallu] has joined #fedora-meeting
20:00 < G> moo!
20:00 < ivazquez> Pong.
20:00  * skvidal is
20:01  * SmootherFrOgZ is
20:01 < kcadieux> hello
20:01 < mmcgrath> Ok, lets get started
20:01 -!- mmcgrath changed the topic of #fedora-meeting to: Infrastructure -- Outstanding tickets
20:01 < mmcgrath> .tiny https://fedorahosted.org/fedora-infrastructure/query?status=new&status=assigned&status=reopened&group=milestone&keywords=%7EMeeting&order=priority
20:01 < zodbot> mmcgrath: http://tinyurl.com/2hyyz6
20:01 -!- lfoppiano_ [n=lfoppian at host5-167-dynamic.8-87-r.retail.telecomitalia.it] has quit "Ex-Chat"
20:01 -!- Sonar_Guy [n=Who at fedora/sonarguy] has quit "IRC is just a real time Flame Fest"
20:01 < mmcgrath> .ticket 395
20:01 < zodbot> mmcgrath: #395 (Audio Streaming of Fedora Board Conference Calls) - Fedora Infrastructure - Trac - https://fedorahosted.org/fedora-infrastructure/ticket/395
20:01 < mmcgrath> jcollie: anything new?
20:02  * ricky 
20:02 < jcollie> um nope
20:02 < mmcgrath> skvidal: just so I know, how high of a priority is this from the boards point of view?
20:02 < skvidal> I wouldn't stress it
20:02  * dgilmore is here
20:02 < G> jcollie: nope's get one zap in the electric chair you know :)
20:02 < mmcgrath> k
20:02 < mmcgrath> .ticket 446
20:03 < zodbot> mmcgrath: #446 (Possibility to add external links on spins page) - Fedora Infrastructure - Trac - https://fedorahosted.org/fedora-infrastructure/ticket/446
20:03 < mmcgrath> dgilmore: ^^
20:03 -!- jmbuser [n=jmbuser at 94.98.47.119] has quit "Leaving"
20:04 < mmcgrath> dgilmore: anything new there?  Can that ticket be closed?
20:05 < mmcgrath> ok, we'll skip that
20:05 < mmcgrath> and 740 for now
20:05 -!- mmcgrath changed the topic of #fedora-meeting to: Infrastructure -- Change freeze
20:06 < mmcgrath> The pre-release change freeze is going to happen again for the preview release.  The release is on the 4th so the freeze starts on the 21st.  Thats next Tuesday.
20:06 < mmcgrath> Those of you having any major changes to do, do it before that.
20:06 < mmcgrath> any questions?
20:06 < G> mmcgrath: err...
20:06 < G> release is on the 4th?
20:06 < G> oh... preview release
20:07 < mmcgrath> yeah, not the final, this is still just a pre-release freeze.
20:07 < abadger1999> When's the final freeze start?
20:07  * abadger1999 determining the window.
20:07 < mmcgrath> the final freeze starts on the 11th
20:07 < mmcgrath> assuming no slip.  so the window between pre-freeze and the final freeze is large.
20:08 -!- balor [n=balor at gimili.plus.com] has quit Remote closed the connection
20:08 < mmcgrath> Also, I've heard there's a new switch going in at some point in time soon.  Initial estimates put it in the third week in November.  I said no to that.
20:08 < mmcgrath> which means it might get done sooner, might get done after.
20:08 < G> mmcgrath: you said no to finally getting a new switch? :P
20:09 < jcollie> G: he said no to putting it in right before the release
20:09 < mmcgrath> Even if its during the pre-release freeze I'd like to get that switch in.  I generally consider it lower risk but it will involve downtime.  There's no room in that rack so we have to power one down, remove it, put the new one in, hook it up.
20:09 < G> yeah yeah, I know, third week is a pain for release :)
20:09 < mmcgrath> hehehe
20:09 < mmcgrath> yeah.
20:09 < mmcgrath> Yeah.
20:09 < G> jcollie: I was trying to be sarcastic
20:09 < mmcgrath> Ok, so thats a heads up.
20:10 -!- mmcgrath changed the topic of #fedora-meeting to: Infrastructure -- bapp
20:10 < mmcgrath> err sorry, next topic :)
20:10 < mmcgrath> so I've moved a couple more things to the bapp server like the translations cron job.
20:10 < mmcgrath> Hopefully before the freeze I'll have the docs site over there as well
20:11 < mmcgrath> anyone have any comments on our new server setup?  So far its worked well.
20:11 < abadger1999> What's the best puppet file to look at as an example?
20:11 < G> Yep, I have a comment... It's great now we have it finally working :)
20:12 < mmcgrath> abadger1999: appRhel.pp now has a section in it for bapp servers and non bapp servers.
20:12 < abadger1999> Cool.  I'll see about moving some other crob jobs over
20:12 < mmcgrath> so, with the translations app for example, in the bapp section I have a "update translations" include.  In the else section I have the "rsync db" include
20:13 < abadger1999> Are we always going to have a single bapp server?
20:13 < mmcgrath> always one active.
20:13 < abadger1999> Cool.
20:13 < mmcgrath> until we find its being overwhelmed by cron jobs :)
20:13 < mmcgrath> its pretty easy to set another one up if it crashes, and that should be transparent to the users which is good.
20:14 < mmcgrath> Anyone have anything else on that?
20:14 < mmcgrath> Ok
20:14 -!- mmcgrath changed the topic of #fedora-meeting to: Infrastructure -- db2 upgrade
20:14 -!- Valholla_work [i=0c29727e at gateway/web/ajax/mibbit.com/x-a8a2fdbfcfe68d57] has joined #fedora-meeting
20:14 -!- Valholla_work is now known as work_valholla
20:15 -!- JSchmitt [n=s4504kr at fedora/JSchmitt] has quit "Konversation terminated!"
20:15 < mmcgrath> so we did a db2 upgrade yesterday.  It went well.  the db was down only for a few minutes and AFAIK has been working great.
20:15 < mmcgrath> this was the last upgrade to postgres8.3 that we've done.
20:15 < ricky> Woo, so no more vacuuming scripts, etc. then?
20:15 < mmcgrath> nope, all gone
20:15 < G> mmcgrath: so I take it we are assuming RHEL6 will have postgres >= 8.3?
20:15 < abadger1999> Right :-)
20:15  * work_valholla grabs a chair in back,  because he is running late
20:16 < mmcgrath> G: not sure actually.  I guess its up to them.  But even if it doesn't, we have our infrastructure repo.
20:16 < G> mmcgrath: yeah I know :)
20:16 < mmcgrath> I'm a little surprised they haven't shipped 8.3 in some other channel or something after all the problems we had with the stock version
20:16 -!- cassmodiah [n=cass at fedora/cassmodiah] has quit Read error: 104 (Connection reset by peer)
20:16 -!- AndreasR [n=zeus at 80-218-143-60.dclient.hispeed.ch] has joined #fedora-meeting
20:16 < mmcgrath> actually I guess I never checked, we might ship an 8.3 :)
20:17 < G> mmcgrath: I don't think RH does
20:17 < G> mmcgrath: RH Application Stack only ships 8.2
20:17 < mmcgrath> :: sigh ::
20:17 < mmcgrath> ahwell.
20:17 < mmcgrath> yeah, I'm just running under the hope that RHEL6 will have it.
20:17 < mmcgrath> though by then it might be 8.4 :)
20:17 < G> or the RHEL6 App Stack
20:17 < G> mmcgrath: true
20:17 < mmcgrath> Ok, so anyone have any questions on that?
20:18 -!- mmcgrath changed the topic of #fedora-meeting to: Infrastructure -- Performance issues
20:18 < mmcgrath> So I've been investigating some of the more odd performance issues we have to try to get a better handle on how our applications interact.
20:19 < mmcgrath> this includes simple things like too many httpd processes running on a host, to having unusually slow connections from app server to db over the vpn.
20:19 < mmcgrath> I think there are some efficiences to be gained there.
20:19 < mmcgrath> Does anyone have anything they'd like to comment on there?
20:20 < mmcgrath> ok, next item
20:20 -!- mmcgrath changed the topic of #fedora-meeting to: Infrastructure -- fedoraproject.org/maps
20:20 < dgilmore> is there one or two small thinsg that would have a big impact?
20:20 < abadger1999> Keep it up :-)
20:20 < abadger1999> If we need code changes I can help.
20:20 < mmcgrath> dgilmore: I'm not sure yet, we changed the database server in the middle of some of my tests so that kind of threw things off :)
20:21 < mmcgrath> really we're in decent shape, but sometimes things pop up like, why does haproxy report pkgdb as dead on app4 so often?
20:21 < mmcgrath> especially since all the app servers should be the same now.
20:21 < mmcgrath> so I don't think any major changes, just lots of little ones.
20:21 < mmcgrath> Anywho, huzaifas has volunteered to get our maps stuff back in order on log1.
20:22 < mmcgrath> huzaifas's been doing a great job recently, and he's on the other side of the planet like G so he's a great addition to our team.
20:22 < mmcgrath> I think the maps script is all worked out and fixed, but we're missing a perl module.  The owner has been contacted and we're going to try to get an EPEl branch for it
20:22 < mmcgrath> Thats really all there is on that.
20:22 < ricky> Coool
20:23 < mmcgrath> So
20:23 -!- mmcgrath changed the topic of #fedora-meeting to: Infrastructure -- Open Floor
20:23  * ricky wonders if there are any python plans for version 2 :-)
20:23 -!- ke4qqq [n=ke4qqq at fedora/ke4qqq] has quit Remote closed the connection
20:23 -!- ke4qqq [n=ke4qqq at fedora/ke4qqq] has joined #fedora-meeting
20:24 < mmcgrath> anyone have anything they'd like to discuss?
20:24 < G> mmcgrath: two items maybe... 1) Ticket handling on trac
20:24 < mmcgrath> ricky: oh who knows with these things ;)
20:24 < G> and I just forgot 2) dammit
20:24 < mmcgrath> heh
20:24 < mmcgrath> well lets hit 1.  what about ticket handling?
20:25 < G> mmcgrath: really, 'it'd be nice' to start using statuses of some sort
20:25 < mmcgrath> statuses?
20:25 < G> mmcgrath: like 'instaging' etc
20:25 < G> trac doesn't really support statuses so I guess we'd have to use the keywords component
20:26 < mmcgrath> what problem are we trying to solve?
20:26 -!- No5251 [n=No5251 at p50990d02.dip0.t-ipconnect.de] has quit Read error: 113 (No route to host)
20:26 < G> mmcgrath: make it easier when triaging etc
20:26  * mmcgrath often doesn't know if people are actually working on something when its assigned to them.
20:26 -!- AndreasR [n=zeus at 80-218-143-60.dclient.hispeed.ch] has quit "Verlassend"
20:26 < G> (or looking for good candidates to close etc)
20:26 < G> mmcgrath: I agree
20:27 < mmcgrath> we can try it.  Put together a workflow and send it to the list.
20:27 < ricky> How much burden does it add to reporters/handlers, and what do we get in exchange?
20:27 < mmcgrath> If people follow it great, if they don't we're no worse off then we are now.
20:27 < G> mmcgrath: I've been guilty of it too, I've assigned a ticket to myself and 'forgotten about it'
20:27 < mmcgrath> ricky: The big one I run into is people assigned a task that aren't working on it.
20:27 < ricky> https://fedorahosted.org/fedora-infrastructure/report/8 :-)
20:27 < mmcgrath> I'd also like to be more liberal about people taking stalled tickets.
20:27 < G> ricky: well, look at the BZ workflow @redhat'ers have, imo it's great because it's somewhat obvious what stage a ticket is in
20:28 < mmcgrath> like "if someone takes your ticket and you haven't been working on it for X days, don't get a sour face about it"
20:28 < G> mmcgrath: like I did with the Fedora Hosted post-update ticket? :P
20:28 < ricky> Could we just not get sour faces period?  :-)
20:29 < mmcgrath> ricky: yeah but some people still get that way.
20:29 < mmcgrath> G: well, if you're interested propose a workflow and send it to the list so we can comment and adopt it.
20:29 < G> mmcgrath: right
20:29 -!- LetoTo [n=paul at bofh.xelerance.com] has quit No route to host
20:30 < mmcgrath> Anyone have anything else to discuss?
20:30 < G> oh yeah...
20:30 < G> 2) RH App Stack
20:30 < G> It has PHP 5.2 iirc which could be nice for when we are looking at CMS'
20:31 -!- LetoTo [n=paul at bofh.xelerance.com] has joined #fedora-meeting
20:32 -!- zcat [n=zcat at pdpc/supporter/active/zcat] has quit Read error: 110 (Connection timed out)
20:33 < mmcgrath> I'm fine with moving to that if the CMS supports it.
20:33 < mmcgrath> but lets not move to it if the CMS doesn't gain from it.
20:33 -!- J5 [n=quintice at c-76-24-17-105.hsd1.ma.comcast.net] has joined #fedora-meeting
20:33 < G> mmcgrath: +1
20:33 < mmcgrath> K, anyone have anything else to discuss?  If not we can close the meeting in 30
20:33 < ricky> Hmm,  newer versions of RHEL packages?  I wonder how tough that it to manage (and if it's in the scope of EPEL stuff)
20:34 -!- zcat [n=zcat at pdpc/supporter/active/zcat] has joined #fedora-meeting
20:34 < G> ricky: well it's a seperate channel, which is how RH like to do stuff
20:34 < mmcgrath> ricky: the RH App Stack is a channel in RHN and would replace RHEL packages.  But EPEL can't touch it because RHN does technically ship it.
20:34 < skvidal> ricky: newer ver of things in rhel?
20:34 < ricky> As in, is that something that EPEL can do as well?  I was just curious
20:35 < G> ricky: we'd kinda have to do a "EPEL App Stack" repo in a way
20:35 < G> that could build against "RHEL+EPEL+RH App Stack"
20:35 < skvidal> ricky: I think that would be 'frowned on'
20:35 < G> skvidal: +1 :)
20:35 < ricky> Heh, OK
20:35 -!- tibbs [n=tibbs at fedora/tibbs] has quit "Konversation terminated!"
20:36 < mmcgrath> Ok, anyone have anything else to discuss?
20:36 < mmcgrath> if not we'll close the meeting in 30
20:36 < skvidal> not to mention the dep resolution would make me want to kill people
20:36 < skvidal> 'oh look, we have a newer glibc... doom'
20:36 < G> skvidal: :P
20:36 < mmcgrath> 15
20:36 < mmcgrath> 5
20:37 < G> Have a good day all!
20:37 -!- mmcgrath changed the topic of #fedora-meeting to: Infrastructure -- Meeting End
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 197 bytes
Desc: not available
URL: <http://listman.redhat.com/archives/fedora-infrastructure-list/attachments/20081016/3c42fd99/attachment.sig>


More information about the Fedora-infrastructure-list mailing list