From Fedora Project Wiki

Meeting of 2006-11-16

*** Time shown in EST

15:00 < dgilmore> Everyone here and ready for the meeting?
15:01 < warren> meh
15:01 < warren> I'm still exhausted from yesterday
15:01 < dgilmore> warren: you can sleep latter
15:01  * mmcgrath is here
15:01 < warren> yeah
15:02 < mmcgrath> but also in another meeting so won't be responding extremely quickly
15:02 < warren> I just can't think of topics and lead discussion in my current mental state.
15:02 < dgilmore> packagedb abadger1999, c4chris anything on this?
15:02 < dgilmore> warren: sit back relace and chime in when you need to
15:03 < abadger1999> heh -- I'm writing an email to the list write now.
15:03 < dgilmore> i know i need to spend some more time on tha package database
15:03 < dgilmore> warren: relax even
15:03 < abadger1999> Basicaly, the summit had a lot of ideas about what to do ith the new buildsystem and the package db.
15:03 < warren> given the requirements of packagedb to handle the release management stuff that Jesse talked about brew currently doing...
15:03  * iWolf_ here, moreso in another minute or two
15:04 < warren> has Jesse effectively described what those requirements are, or reviewed the proposed schema?
15:04 -!- You're now known as iWolf
15:04 < abadger1999> This is going to be essential to release management in the merged FE+FC
15:04 < dgilmore> f13: ping
15:04 < abadger1999> I don't think the current schema encompasses everything that as talked about.
15:04 < abadger1999> But it can probably be extended to handle it.
15:04 < warren> nod
15:05 < dgilmore> abadger1999: yeah the buildsys, packagedb  and update system are the most important pieces we need to work on
15:05 < abadger1999> I have an importer for extra's owners.list and cvs modules almost written so I'm going to finish that up and get a prototype running.
15:05 < abadger1999> Then Jesse and others can take a look and we can fix the schema to handle all the tasks we need to do.
15:06 < abadger1999> I'm hoping to have that running by Monday night so Jesse can take a look at it early next week.
15:06 < warren> cool
15:06 < dgilmore> abadger1999: cool
15:07 < dgilmore> let us know if you need anything
15:07 < abadger1999> At that time we should be able to break off little subprojects as well -- Get a web designer to help work on the front end, people to write other improters, etc.
15:07 < dgilmore> awsome
15:08 < dgilmore> abadger1999: your the VCS guy.  this has kinda been pushed back a little  but still is a high prioritory thing
15:08 < dgilmore> it seems  that we have made some progress on it
15:08 < dgilmore> where are we at  other then dwmw2 and alan cox  bitching at each other?
15:09 < mmcgrath> I get the sense that f13's pretty happy with mercurial.
15:09 < abadger1999> :-)
15:09 < abadger1999> f13's mercurial work is awesome.
15:09 < dgilmore> mmcgrath: i get that also
15:09 < warren> dist-hg using the existing system could be deployed very quickly
15:09 < warren> the questions remaining have to do with... do we want to remain with tarball + patch
15:09 < abadger1999> I'd be happy to use dist-hg.
15:09 < abadger1999> although there's still more work to be done to finish it.
15:09 < warren> blizzard wants exploded source
15:10 < abadger1999> Exploded source has plenty of advantages.
15:10 < halfline_> as long as the srpms are still tarball plus patches
15:10 < abadger1999> But it could have disadvantages as well.
15:10 < mmcgrath> warren: ?
15:10 < halfline_> (and not just a new tarball, or a pristine tarball and megapatch)
15:10 < mmcgrath> oh, nm
15:10 < warren> halfline_, yeah, I have no clue how that is supposed to work while maintaining the tarballs as being identical to upstream.
15:10 < mmcgrath> ignore me :D
15:10 < abadger1999>
15:10 < abadger1999> That is a proposal that went to the ubuntu lists.
15:11 -!- cyril_FC6 []  has joined #fedora-admin
15:11 < dgilmore> warren: im against exploded tarballs
15:11 < warren> ubuntu addressed the split vs. megapatch issue?
15:11 < warren> Ubuntu hasn't been so friendly toward upstream as us
15:11 < dgilmore> i like the current approach of pristine tarballs, + small patches
15:11 < warren> I am against exploded source if that makes it difficult to logically separate patches and feed them upstream one at a time.
15:11 < dgilmore> small patches == easier route to upstream
15:12 < abadger1999> Ubuntu, being debian based, has to live with megapatches at the root level (within the dpkg) but the proposal I posted makes use of DRCS branching to simulate small patches to the developer.
15:12 < warren> where did you post that proposal?
15:13 < abadger1999> Those will either be made into a megapatch when it gets turned into a .deb or they won't have a source package.
15:13 < dgilmore> ubuntu AFAIKT  only cares about  themselves  and  doesnt care to get things upstream  weather that is debain  or the upstream project
15:13 < abadger1999>
15:13 < warren> DRCS = Directory of Radioactively Contaminated Sites
15:13 < warren> I'm Feeling Lucky!
15:13 < abadger1999> The Ubuntu proposal.
15:13 < warren> oh
15:14 < abadger1999> Note that the loomify plugin to bzr doesn't exist yet.
15:14 < abadger1999> It's a specification and the developer is needign to get others to buy into the idea before going forward.
15:14 < dgilmore> the current system i feel works well.
15:14 < warren> loomify is comparable to quilt?
15:15 < warren> Maybe we need to be debating this in the yet to be created fedora-scm group
15:15 < abadger1999> dgilmore: I am unhappy with it -- I run quilt on everything I check out of our cvs.
15:15 < abadger1999> I shouldn't have to do that.
15:15 < dgilmore> abadger1999: hmm ok i would ask what you gain by that but this is not the place or time for it
15:16 < dgilmore> lets move onto DB1 upgrade
15:16 < abadger1999> warren: Most likely.  Like I said in fesco, doing exploded sources is exciting but also much different than what e do now.
15:16 < dgilmore> iWolf: status
15:16 < iWolf> dgilmore: waiting for the hardware swap
15:16 < warren> dgilmore, from a packager POV the current tarball + patch system works well, but many software developers feel it is an ugly layer of overhead.
15:16 < iWolf> the proxies getting taken out of rotation so we get the new Dells in place again.
15:16 < iWolf> Then I'll be rolling forward on that (beyond what I did for prep to a xen guest we worked with).
15:16 < dgilmore> iWolf: ok cool i guess we will have to bug pick  see where he is at
15:17 < iWolf> dgilmore: I think he was in meetings this week.
15:17 < dgilmore> iWolf: are you using the current db2 xen test
15:17 < iWolf> dgilmore: he said if he didn't get it this week he would get it next.
15:17 < dgilmore> iWolf: i believe so
15:17 < iWolf> dgilmore: not at the moment no.  I had already restored the dbs to play with them and that went well.
15:17 < iWolf> dgilmore: if we need to re-use it feel free.
15:18 < dgilmore> iWolf: i dont know if we do or not  but its there if you need to do anything
15:18 < dgilmore> lmacken: about ?
15:18 < iWolf> dgilmore: Cool.  I won't keep anything that can't be replaced on it if people find the need to re-use.
15:18  * mmcgrath notes we should be getting the new dells back soon.
15:19 < dgilmore> i guess lmacken is not here
15:19 < dgilmore> lmacken: when you get time can you email the list with an update of firewall stuff
15:20 < dgilmore> warren: xen boxes are up  i guess we can mark that as doen
15:20 < dgilmore> done
15:20 < dgilmore> warren: did you hear back on a ip for xen3?
15:20 < warren> hm...
15:20  * lmacken is here (but on the phone)
15:20 < warren> apparently he lost it
15:20 < dgilmore> warren: ok
15:20 < warren> he... doens't tend to read e-mail carefully enough
15:21 < warren> I don't remember what this was for anymore
15:21 < dgilmore> warren: well i started building xen3  and used which is the ip after xen2
15:21 < dgilmore> warren: it is for tha xen host for xen builder replacing hammer1
15:21 < dgilmore> if need be ill change it hammer1's ip
15:22 < lmacken> dgilmore: yeah, I'm going to finish the deployment of pyroman next week.  I'll keep the list up to date.
15:22 < dgilmore> lmacken: cheers
15:22 < warren> dgilmore, ok, so just ask to point xen3 (internal only) to
15:23 < dgilmore> warren:  yes please
15:23 < warren> k
15:23  * warren sending ticket
15:23 < dgilmore> warren: lets also ask for xen4  for the new build box on its way
15:23 < dgilmore> and a xenbuilder2   to use on it
15:24 -!- daMaestro [n=jon@fedora/damaestro]  has joined #fedora-admin
15:24 < warren> and I'm guessing
15:24 < daMaestro> any info on this -> ??
15:24 < dgilmore> warren: yeah
15:24 < warren> dgilmore, where is the new build box coming from?
15:24 < dgilmore> warren: not sure Dell donated it
15:24 < warren> cool
15:24 < dgilmore> its a dual dual core xen box
15:24 < dgilmore> with 4gb ram
15:25 < dgilmore> mdomsch  said it should be there any day
15:25 < dgilmore> s/sen/seon/
15:25 < dgilmore> gahh its got xeon cpus
15:26 < warren>
15:26 < warren>
15:26 < warren> (new external IP)
15:26 < warren> (new internal IP)
15:26 < warren> ?
15:26 < warren> dgilmore, xeon x86_64?
15:26 < dgilmore> warren: as far as i know  yeah
15:26 < warren> these requests look good?
15:26 < mmcgrath> daMaestro: you didn't include the stream ;-)
15:27 < dgilmore> warren: yeah that looks good
15:27 < warren> k
15:27 < daMaestro> mmcgrath, ok.. so i just need to submit the stream url?
15:27 < dgilmore> daMaestro: yep
15:27 < daMaestro> dgilmore, ok.. thanks.. i will get it setup
15:27 < daMaestro> we are just in the testing phase now
15:27 < daMaestro> i'm working on getting the metaweblog api working
15:28 < warren> xenbuilder2 needs 8887-8889, anything else?
15:28 < warren> (same ports open as hammer1)?
15:28 < dgilmore> warren: thats it
15:28 < dgilmore> warren: same as the other builders
15:28 < warren> k
15:28 < dgilmore> warren: we should get hammer1 and 3 taken out
15:29 < warren> OK.
15:29 < dgilmore> Legacy buildsys  is setup and ready
15:29 < dgilmore> it needs push scripts
15:30 < warren> dgilmore, f13 needs to decide what happens there
15:30 < dgilmore> its future depends on what legacy wants to do
15:30 < dgilmore> warren: :)  i know
15:30 < warren> given that he proposed discontinuing legacy during the summit
15:30 < daMaestro> is atom, rss or rdf better?
15:30 < warren> I didn't fully agree with it, and the legacy team is debating it on their list
15:30 < dgilmore> EPEL buildsys is up.  i have succesful builds for EL-4 and EL-5
15:30 < daMaestro> *for you guys
15:30 < dgilmore> daMaestro: rss
15:30  * lmacken has to run to another meeting.
15:30 < skvidal> daMaestro: atom has a spec
15:31 < daMaestro> ok
15:31 < lmacken> no updates from my end.. finals are done tomorrow night, and I'm going to start hacking on the update system and get the rest of the firewalls deployed next week
15:31  * lmacken &
15:31 < dgilmore> lmacken: thanks and good luck
15:31 < lmacken> dgilmore: thanks
15:31 < warren> dgilmore, hammer1 hardware is becoming xen3?
15:31 < dgilmore> warren: yes
15:31 < warren> k
15:32 < dgilmore> we will need to rename consoles. etc
15:32 < warren> that was part of my ticket
15:33 < dgilmore> warren: :D  thanks
15:33 < dgilmore> OK  anyone  have anything to add?
15:34 < dgilmore> floor is open for whatever
15:35 < mmcgrath> Damnit.
15:35 < mmcgrath> Ok, this isn't good
15:35 < mmcgrath>
15:36 < dgilmore> no thats not
15:36 < dgilmore> do we need a robots.txt file on admin
15:37 < mmcgrath> I'm going to enable authentication again
15:37 < mmcgrath> the issue is that some stuff breaks when I do that, we'll have to create a fake user for the scripts.
15:38 < dgilmore> mmcgrath: :)  yeah   accounts to the buildsys was the biggest
15:38 < abadger1999> Can you limit by ip instead?
15:38 < dgilmore> should be able to in the apache configs
15:38 -!- daMaestro [n=jon@fedora/damaestro]  has quit ["Leaving"] 
15:39 < abadger1999> Right.  So then the buildsys can get access but not google.
15:39 < warren> robots.txt is the only way you'll get that out of google's cache in the future.
15:39 < dgilmore> we will need a robots.txt  and allow only from certain ips without auth
15:39 < mmcgrath> abadger1999: not easily because we're behind a balancer and reverse proxy.
15:39 < warren> nod
15:40 < abadger1999> hmmm
15:43 < dgilmore> anyone got anything else?
15:44 < dgilmore> if not ill close the meeting in 60 seconds
15:44 < mmcgrath> I've got one thing
15:44 < dgilmore> mmcgrath: sure
15:44 < mmcgrath> warren: whats the easiest way for us to communicate with the developers to find out what is most important from infrastructure as far as releases go?
15:45 < dgilmore> mmcgrath: yes  that is important
15:45 < mmcgrath> we won't always get everything done but I'd hate for us to be working on something that isn't important to the project while something tha
t was important didn't get done.
15:45 < dgilmore> mmcgrath, warren:  what can we do so that non RH people can file tickets for stuff we need RH IS to do for us?
15:45 -!- paulobanon [i=keefejoh@gateway/web/cgi-irc/]  has joined #fedora-admin
15:45 < warren> mmcgrath, in the case of fedora merge, f13 needs to communicate effectively with infrastructure.
15:46 < warren> and when we make the hire for full-time fedora infrastructure lead, that role becomes accountable to all communication with developers
15:46 < paulobanon> howdy
15:46 < mmcgrath> sounds good.
15:46 < mmcgrath> hey paulo, we're just wrapping up
15:46 < dgilmore> warren: what was the word on that position?
15:46 < paulobanon>
15:46 < paulobanon> for everyone behind fw
15:46 < paulobanon> :D
15:47 < warren> dgilmore, mspevack and I are still working on it.
15:47 < dgilmore> warren: ok cool keep us in the loop
15:47 < warren> there is a huge and burning need for the role to happen
15:47 < warren> but no guarantees in anything =(
15:47 < mmcgrath> yep
15:48 < mmcgrath> alrighty, thats all I've got dgilmore, feel free to close whenever ;-)
15:48 < dgilmore> ok  closing meeting in 30
15:48 < dgilmore> ===  mark closed meeting ===