20:00:08 <mmcgrath> #startmeeting
20:00:08 <zodbot> Meeting started Thu Jan 21 20:00:08 2010 UTC.  The chair is mmcgrath. Information about MeetBot at http://wiki.debian.org/MeetBot.
20:00:08 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
20:00:14 * ricky 
20:00:16 <mmcgrath> #topic Who's here?
20:00:18 * ricky 
20:00:21 * jds2001 
20:00:26 <jaxjax> Hi
20:00:26 * wzzrd 
20:00:49 * nirik is lingering in the back
20:00:50 <abadger1999> Here.
20:00:53 <nullsys> Here
20:00:57 * sijis is here
20:01:02 <jaxjax> I'm newby here
20:01:04 * a-k is
20:01:07 <mmcgrath> Lets see if we can get through the meeting without taking the wiki down or losing a chunk of one of our production switches.
20:01:39 <mmcgrath> OK, lets get started
20:01:43 <mmcgrath> #topic Infrastructure -- Tickets
20:01:47 * sijis is wondering how a meeting can take down the wiki
20:02:04 <mmcgrath> sijis: ehh, we lost the wiki twice last night for seemingly very strange reasons.
20:02:12 <mmcgrath> .tiny https://fedorahosted.org/fedora-infrastructure/query?status=new&status=assigned&status=reopened&group=milestone&keywords=~Meeting&order=priority
20:02:12 <zodbot> mmcgrath: http://tinyurl.com/47e37y
20:02:15 <mmcgrath> hey look, no tickets.
20:02:17 <mmcgrath> that makes this easy.
20:02:30 <mmcgrath> #topic boot.fedoraproject.org
20:02:32 <jaxjax> :)
20:02:44 <mmcgrath> So this is kind of an odd topic since we don't really deal directly with the OS all that often.
20:02:53 <mmcgrath> but - http://boot.fedoraproject.org/
20:02:59 <mmcgrath> is unofficially online.
20:03:06 <mmcgrath> I'd appreciate it if people with abilities can give it a try.
20:03:12 <mmcgrath> there's experimental stuff on there that flat out doesn't work
20:03:16 <mmcgrath> but install and rescue bits should
20:03:44 <mmcgrath> Has anyone had a chance to play with it yet?
20:03:44 <abadger1999> Anyway to generate a list of what's hosted semi-automatically?
20:04:10 <mmcgrath> abadger1999: on fedorahosted?
20:04:24 <abadger1999> On boot.fp.o
20:04:36 <abadger1999> The images we can boot.
20:04:55 <wzzrd> mmcgrath: boot is wicked; anything specific you want tested?
20:04:55 <mmcgrath> abadger1999: ah, I'll try to get one together, at the moment just install media.  I'm hoping to get live media going though
20:05:05 <mmcgrath> wzzrd: just it, over and over and over.
20:05:11 <wzzrd> doing it as we speak
20:05:18 <mmcgrath> if it's buggy I don't want to expose the users to it yet.
20:05:42 <mmcgrath> abadger1999: oh, and some other tools are on there.
20:05:46 <jds2001> is BKO officially launcehd yet?
20:05:51 <sijis> is this going to be the new way to install fedora?
20:05:59 <mmcgrath> the thing that's nice is we can update it without having to force people to re-burn the image.
20:06:13 <mmcgrath> jds2001: good question, I'm not sure if he's advertising it actively or not yet.
20:06:22 <mmcgrath> BFO (based on BKO) is not yet official but I have a feature request in
20:06:29 <mmcgrath> sijis: it's going to be "a new way" to install fedora
20:06:44 <wzzrd> won't this cost massive bandwidth? or doesn't that matter?
20:06:52 <sijis> gotcha, an alternative
20:07:05 <mmcgrath> wzzrd: actually, as it stands right now, it won't.
20:07:09 <mmcgrath> since it relies on the mirrors.
20:07:24 <mmcgrath> having said that... if we go into live media..  That gets much trickier.
20:07:31 <Oxf13> yeah
20:07:33 <mmcgrath> and involves some architectural decisions.
20:07:49 <mmcgrath> I may not post them at all, I may just point to what BKO already has setup.
20:07:53 <Oxf13> mmcgrath: is there anything about boot.fp.o that requires releng to put content somewhere different/else/triggersomething/whatever?
20:08:17 <mmcgrath> Oxf13: not really, the one thing I was hoping from you guys is to get some things into our default initrd's that aren't there now
20:08:20 <mmcgrath> like iscsi and nfs support
20:08:38 <mmcgrath> other then that I don't think there's any changes for anyone.
20:08:43 <Oxf13> I've been trying to get notting and warthog to talk about that for months :/
20:09:00 <mmcgrath> Oxf13: as I understand it there's an ordering problem with dracut.
20:09:07 <Oxf13> I really don't know what the deal is
20:09:26 <mmcgrath> Oxf13: perhaps after the meeting you can step me through exactly what command you run to build an ISO and I can do some tests and see if I can get it working.
20:09:32 <mmcgrath> I've gotten it working after the fact.
20:09:39 <mmcgrath> but in the actual build process I'm a little confused what is going on
20:09:39 <Oxf13> sure
20:10:03 <mmcgrath> Anyone else have any questions about that?
20:10:25 <mmcgrath> If not we'll move on.
20:10:28 <mmcgrath> k
20:10:32 <mmcgrath> #topic The outages last night
20:10:35 <mmcgrath> So last night was weird.
20:10:44 <mmcgrath> special thanks for ricky and nigel for taking care of things.
20:11:03 <mmcgrath> The wiki went out twice, then we lost a switch blade in one of our switches.
20:11:09 <mmcgrath> that's now replaced as of an hour or two ago.
20:11:19 <mmcgrath> We're not sure if the switch issues caused the wiki issues or not
20:11:33 <mmcgrath> really not much else to say about that.
20:11:36 <mmcgrath> #topic NTP
20:11:39 <mmcgrath> smooge: this all done?
20:11:56 <smooge> NTP is done
20:12:29 <ricky> I have some puppet errors about being unable to remove the ntp package
20:12:38 <ricky> Some machines need system-config-date for something, and thus need ntp
20:12:39 <mmcgrath> ricky: how many errors?
20:12:48 <ricky> So I'd juts disable the service and not remove the package in ntp::noclient
20:12:54 <smooge> all hardware systems should be running NTP
20:13:06 <smooge> oh ok
20:13:12 <smooge> I can change the logic
20:13:14 <ricky> compose machines, spin1, publictest2
20:13:15 <smooge> sorry didn't know that
20:13:21 <ricky> No worries, it didn't cause any harm
20:13:31 <pjones> mmcgrath: Oxf13: the ordering problem is that there's no way to tell rpm to order something earlier in the transaction /if and only if/ it's already there.
20:13:40 <mmcgrath> ricky: so non-RHEL hosts?
20:13:51 <smooge> well I was planning on installing it on all machines at first and just have it not running.. but I was worried if that was outside of standard scope
20:13:56 <ricky> Ah, looks like it, didn't notice that.
20:14:20 <mmcgrath> pjones: is there some command we can run in %post?
20:14:54 <pjones> not really; for mpath I'm probably going to change anaconda to set hostonly in the dracut config and rebuild the initrd at the end of install.  which sucks.
20:15:05 <pjones> (we could make other things do that as well)
20:15:30 <mmcgrath> pjones: mind if I ping you after the meeting for more?
20:15:45 <pjones> mmcgrath: no problem
20:15:51 <mmcgrath> smooge: ricky: we can take a look after the meeting
20:15:54 <mmcgrath> ricky: what is the error btw?
20:15:56 <smooge> np
20:16:11 <ricky> Thu Jan 21 17:10:55 +0000 2010 //ntp::noclient/Package[ntp]/ensure (err): change from 4.2.4p8-1.fc13 t
20:16:14 <ricky> o absent failed: Execution of '/bin/rpm -e ntp-4.2.4p8-1.fc13.x86_64' returned 1: error: Failed depend
20:16:18 <ricky> encies: ntp is needed by (installed) system-config-date-1.9.53-1.fc13.noarch
20:16:38 <mmcgrath> ricky: thanks.
20:16:59 <dgilmore> we should use yum to remove it :)
20:17:03 <mmcgrath> Ok, any questions on that?
20:17:11 <mmcgrath> dgilmore: I think puppet actually used to do that.
20:17:53 <mmcgrath> Ok, moving along
20:17:58 <mmcgrath> #topic Search Engine
20:18:01 <mmcgrath> a-k: around?
20:18:05 <a-k> yep
20:18:15 <a-k> One of our candidates, Xapian, is on publictest3
20:18:25 <a-k> I got it to crawl fedoraproject.org to a depth of 4 (after about 4 hours of running)
20:18:26 <mmcgrath> how'd the crawl go?
20:18:34 <a-k> I expected disk to be the issue, but it was memory, instead
20:18:45 <a-k> The indexer chewed up all the memory at a depth of 5 (after about 10 hours of running)
20:18:52 <a-k> So I backed off one level
20:19:04 <a-k> The search tool is available at:
20:19:11 <a-k> #link http://publictest3.fedoraproject.org/cgi-bin/omega
20:19:23 <a-k> Nutch is next to do
20:19:30 <a-k> I'll update the wiki with the new info on publictest
20:19:42 <a-k> huzaifas also says he'll be updating the wiki with the stuff he's evaluated in the past month or so
20:19:52 <a-k> That's about it for now
20:19:55 <mmcgrath> a-k: excellent.
20:19:58 <mmcgrath> seems pretty peppy
20:20:02 <mmcgrath> a-k: do you need more memory?
20:20:06 <mmcgrath> how much does pt3 have now?
20:20:24 <a-k> I don't know if the memory issue was a leak or if it really needs more
20:20:36 <a-k> We could always try more, if you're up for it
20:20:44 <a-k> But I don't know how much more more is
20:20:54 <jds2001> its got 2gb
20:21:03 <a-k> A 2g of swap
20:21:06 <mmcgrath> jds2001: it's x86_64 though :-/
20:21:07 <mmcgrath> hmm
20:21:11 <mmcgrath> I wonder if we should rebuild it i386
20:21:26 <jds2001> mmcgrath: it's at ibiblio, so we've got plenty of resources there
20:21:34 <a-k> Where will it run when deployed?
20:21:49 <a-k> ... if ever ...
20:21:59 <mmcgrath> yeah
20:22:02 <jds2001> on the app servers, i think....
20:22:05 <mmcgrath> a-k: probably across multiple hosts in PHX2
20:22:11 <mmcgrath> a-k: can I reboot publictest3 right now?
20:22:21 <a-k> reboot, sure
20:22:49 <a-k> But it will run on x86_64, right, so we should test there
20:22:54 <jds2001> mmcgrath: how are you doing this? :)
20:23:00 <mmcgrath> a-k: it may or may not.
20:23:12 <mmcgrath> a-k: we run lots of stuff on i386 and get a considerable memory savings on some.
20:23:32 <mmcgrath> a-k: ok, publictest3 is booting and now has 4G of RAM.
20:23:34 <mmcgrath> we'll see if that helps.
20:23:47 <mmcgrath> Anyone have any other questions about that?
20:24:32 <mmcgrath> jds2001: I'm doing it on ibiblio1 if that's what you're asking?
20:25:26 <mmcgrath> anywho, that's that
20:25:31 <mmcgrath> #topic Noobs?
20:25:38 <mmcgrath> Any new people around that want to say hey?
20:25:42 <nullsys> Im a noob
20:25:50 <_yawns_> hey
20:25:55 <mmcgrath> nullsys: yeah, you're working with the websites team right?
20:25:57 <ricky> Welcome
20:25:58 <mmcgrath> and helped with blogs?
20:26:01 <jaxjax> Hi
20:26:12 * wzzrd is a noob too :-)
20:26:25 * dgilmore is n00b
20:26:30 <nullsys> hiemanshu, been around  a week or two. Working on blogsadmin, helping out with website team on a few projects. Im an ambassador and in the freemedia group. Also working with MooDoo to head the fedora-uk stuff too.
20:26:40 <nullsys> sorry, "hi" + tab again lol.
20:26:52 <mmcgrath> nullsys: excellent, welcome
20:26:56 <nullsys> thanks :)
20:27:08 <mmcgrath> _yawns_: you're new?  what brings you to Infrastructure?
20:27:13 <hiemanshu> nullsys: fix it :)
20:27:25 <nullsys> Told you i'd bug you :P
20:27:39 <_yawns_> interested in helping how I can
20:27:57 <_yawns_> not really sure exactly in what fashion yet
20:28:33 <nullsys> Im sure plenty of projects require allsorts of happy. Have you had a look at the different ways you can "join"?
20:28:58 <_yawns_> I had a look around a bit last night and saw this meeting today
20:29:00 <mmcgrath> _yawns_: interested just in something in Fedora or in Infrastructure specifically?
20:29:35 <_yawns_> Infrastructure seemed to match my interest the most as well as skill set
20:29:55 <mmcgrath> _yawns_: well make sure you've read through - http://fedoraproject.org/wiki/Infrastructure/GettingStarted
20:30:08 <_yawns_> I hope you don't mind if I sit in and listen for a while
20:30:11 <_yawns_> will do
20:30:15 <mmcgrath> jaxjax: I know you're new here, tell us a bit about yourself and what you were interested in doing.
20:30:23 <mmcgrath> _yawns_: of course not.
20:30:24 <nullsys> If I could offer any advice here, it would be to contact any project leaders your interested in helping out with. Also a good setup for your wiki user page would help any questions asked. Good luck with it.
20:30:50 <_yawns_> Thanks :)
20:31:04 <jaxjax> Hi I'm interested about virtualization and stor
20:31:17 <jaxjax> Sorry storage
20:31:32 <mmcgrath> jaxjax: ah, have you read through the GettingStarted page?
20:31:38 <jaxjax> But will help with everything you think I can help with
20:31:50 <jaxjax> Yeah I did
20:32:09 <jaxjax> Sent my intro last days
20:32:27 <jaxjax> I think I can help a lot just let me how
20:32:43 <mmcgrath> jaxjax: excellent, lets meet up after the meeting
20:32:49 <jaxjax> Sure
20:33:30 <mmcgrath> Ok, anyone have anything else to say before I open the floor?
20:34:13 <mmcgrath> k
20:34:16 <mmcgrath> #topic Open Floor
20:34:23 <mmcgrath> Anyone have anything in particular they'd like to discuss?
20:34:49 <wzzrd> i saw gregdek has created a cloud sig
20:34:56 <mmcgrath> wzzrd: he did, I'm on it.
20:34:58 <Oxf13> "cloud"
20:35:11 <mmcgrath> :)
20:35:12 <wzzrd> is that something the Infra team will cooperate in, in any way?
20:35:32 <mmcgrath> wzzrd: well, that's a complex question.  Fedora was actually going to offer some free guests running in the cloud.
20:35:39 <mmcgrath> but the software to manage that cloud never really got put together.
20:35:53 <mmcgrath> so it's very much left in a state of limbo at the moment.
20:36:19 <wzzrd> i guess management software like that gets real complex real quick
20:36:38 <mmcgrath> wzzrd: espeically when the requirements are completely spelled out.
20:36:52 <skvidal> mmcgrath: or they are - but they're spelled out mostly in klingon :)
20:36:57 <mmcgrath> heh
20:37:02 <mmcgrath> Oh, one thing I wanted to mention
20:37:26 <mmcgrath> did everyone have a chance to look through - http://infrastructure.fedoraproject.org/csi/host-lifecycle-policy/
20:37:29 <mmcgrath> ?
20:37:38 <mmcgrath> Nothing too objectionable about it?
20:38:08 <wzzrd> pretty long read; not completely through it
20:38:13 <smooge> "cloud" poop
20:38:25 <mmcgrath> wzzrd: yeah, we also have a security policy that's worth a look.
20:38:36 <mmcgrath> oh, and the free software policy :)
20:38:37 <mmcgrath> http://infrastructure.fedoraproject.org/csi/free-software-policy/
20:38:43 <mmcgrath> http://infrastructure.fedoraproject.org/csi/security-policy/
20:38:55 <wzzrd> my eyes are almost square from reading fp.o stuff already :)
20:39:18 <skvidal> 'eyes  are almost square'
20:39:22 <skvidal> I'll have to remember that one
20:39:30 <mmcgrath> heh
20:39:37 <Oxf13> the cloud sig will first focus on getting our EC2 presence up to date
20:39:38 <wzzrd> is that the wrong way to say it?
20:39:42 <Oxf13> and moving forward keeping it up to date
20:39:52 <Oxf13> after that, not sure what the next goal will be
20:39:56 <skvidal> wzzrd: I'm sure it's not - I've just never heard it before
20:40:00 <mmcgrath> Oxf13: yeah, that's a worthy goal at least.
20:40:18 <wzzrd> skvidal: literally translated from Dutch
20:40:28 <mmcgrath> Ok, so anyone have anything else to say?  if not we'll end the meeting a little early today
20:40:30 <skvidal> wzzrd: it sounds great
20:41:29 <mmcgrath> ok then, done :)
20:41:31 <mmcgrath> #endmeeting