20080729

Revision 3 as of 2008-08-06 16:18:59

Clear message

Agenda

Items we will be discussing:

  • Review ACTION points from previous meeting.
  • Review progress made on the specification listed on the Roadmap.

  • Open Discussion.
  • Agree on next meeting date and time.

Minutes

Clamav and spamassassin in main

ScottK mentioned that there are a couple of MIR left to be written in order to get clamav and spamassassin into main in time for Intrepid. Help is more than welcomed to make this happen. The list can be found in the wiki page.

Boot Support for Degraded RAID

kirkland's patches to mdadm and initramfs-tools have been uploaded. He's also updated the wiki page with testing instructions.

ACTION: kirkland to write a blog post about testing booting from a degraded raid array.

Review ServerGuide for Intrepid

sommer updated some sections of the Server Guide as listed on the wiki page. He will mark those that require a review. mathiaz updated the Server Team knowledge base with instructions to checkout the guide and start working on it.

Anyone interested in helping out should branch the server guide and send their review as a bzr branch.

Migrate new installs and upgrades of client and server packages to use SSL v3 or TLS

ivoks mentioned that openssl in intrepid is compiled without support for sslv2.

Add 'status' action to server init scripts

kirkland mentioned that a lot of work has been done in that area. Patches have been created and are waiting for sponsoring. mathiaz will take a shot at them.

james_w wondered if the patches had been sent to Debian and warned about the process to follow when filing a lot of bugs in Debian. kirkland and james_w will discuss this issue off-line.

Call for WP testers

nijaba stated he had a bunch of whitepapers he'd like to get feedback on: automated deployment, mysql clustering, redhat cluster. ivoks and sommer volunteered to do some reviews. Anyone interested in reviewing the mysql clustering and automated deployments whitepapers should send an email to nijaba.

Server survey, Intel interested to partner

nijaba reported that Intel was interested in joining the Server survey. A hardware-related question would be added and they would blog about it. No objections were raised. nijaba also added that RedMonk was the analyst firm we would partner with.

ACTION: nijaba to write up a question related to hw and submit it for review to the rest of the server team.

RAILS integration

NeilW gave a quick summary of the ruby problem: the current Rubygems package doesn't place gem installed binaries on the path. So the likes of Capistrano just doesn't work. He's been talking with the Debian maintainer and came up with different solutions.

He also uploaded a new version of mod_passenger to REVU. mathiaz will have a look at it.

Agree on next meeting date and time

Next meeting will be on Tuesday, August 5th at 15:00 UTC in #ubuntu-meeting.

Log

{{{[16:01] <mathiaz> #startmeeting [16:01] <MootBot> Meeting started at 10:04. The chair is mathiaz. [16:01] <MootBot> Commands Available: [TOPIC], [IDEA], [ACTION], [AGREED], [LINK], [VOTE] [16:01] <mathiaz> Today's agenda: https://wiki.ubuntu.com/ServerTeam/Meeting [16:01] <mathiaz> Previous meeting logs: https://wiki.ubuntu.com/MeetingLogs/Server/20080722 [16:01] <ivoks> and rest of the universe... [16:02] <mathiaz> [TOPIC] Clamav and spamassassin in main [16:02] <MootBot> New Topic: Clamav and spamassassin in main [16:02] <mathiaz> ScottK told me that there are some more MIRs that need to be written === ivoks_ is now known as ivoks [16:02] <mathiaz> so that clamav and spamassassin can be moved to main [16:02] <mathiaz> any help in this area is welcomed [16:03] * nijaba waves [16:03] <ivoks> i've promised some work, but didn't have time yet :/ [16:04] <mathiaz> [TOPIC] Boot Support for Degraded RAID [16:04] <MootBot> New Topic: Boot Support for Degraded RAID [16:04] <mathiaz> kirkland: got some feedback on testing ? [16:04] <kirkland> mathiaz: I have instructions in the wiki page [16:04] <kirkland> mathiaz: https://wiki.ubuntu.com/BootDegradedRaid [16:05] <kirkland> mathiaz: kees sponsored the patches to mdadm and initramfs-tools [16:05] <kirkland> mathiaz: there's a bit more to be done, to the installer and grub [16:05] <kirkland> mathiaz: but I have to shift my focus away from RAID at the moment [16:05] <kirkland> mathiaz: I'll write you a blog entry for testing it this week, though [16:06] <mathiaz> kirkland: great ! [16:06] <kirkland> mathiaz: testing went well enough for kees to sponsor the patches Wink ;-) [16:07] <mathiaz> kirkland: right - I'd mention in the post that we're looking for tests on real hardware [16:07] <mathiaz> kirkland: as most of your testing instructions have to do with kvm [16:07] <kirkland> mathiaz: hmm, perhaps [16:07] <kirkland> mathiaz: if it works in kvm, i'm pretty confident about real hw [16:07] <kirkland> mathiaz: but i'll mention real hw, with some caveats [16:08] <mathiaz> kirkland: right - but real hw is the real-life use case [16:08] <mathiaz> kirkland: anyway - it's just a suggestion for your post. [16:08] <kirkland> mathiaz: yes, but testers have to be willing to lose all data [16:09] <kirkland> mathiaz: stacked encryption and lvm would be something i'd like other people to help test [16:09] <kirkland> mathiaz: and open *new* bugs [16:09] <mathiaz> [ACTION] kirkland to write a blog post about testing booting from a degraded raid array [16:09] <MootBot> ACTION received: kirkland to write a blog post about testing booting from a degraded raid array [16:09] <kirkland> mathiaz: I don't really have time to write up complete, detailed instructions on that [16:09] <mathiaz> kirkland: isn't there such a test case in the iso testing pages (stacked encryption) ? [16:10] <nijaba> kirkland: If you do so, why not think of it as addition that could be put in the server guide? [16:10] <kirkland> mathiaz: yes, there are LVM and Encryption instructions in the ISO testing [16:11] <kirkland> mathiaz: those need to be augmented to talk about stacking on top of RAID [16:11] <kirkland> nijaba: it would be good info for the server guide [16:11] <kirkland> nijaba: i'm a bit over-taxed at the moment, though [16:11] <mathiaz> kirkland: so encrypted over lvm over raid arrays ? [16:11] <sommer> kirkland, nijaba degraded raid, etc is on my list Smile :) [16:11] <sommer> for documenting that is [16:12] <nijaba> sommer: so I think kirkland's testing instruction might be a good starting point [16:12] <kirkland> sommer: great, thanks [16:12] <kirkland> mathiaz: yes [16:12] <sommer> nijaba: definitely... I should be able to help test this week [16:13] * nijaba hugs sommer [16:13] <mathiaz> kirkland: right - and up to now, this setup is not supported in the installer [16:13] <mathiaz> kirkland: or a least with a simple recipe [16:13] <mathiaz> kirkland: you have to setup everything by hand IIRC [16:13] <kirkland> mathiaz: right [16:13] <kirkland> mathiaz: i figured it out, eventually [16:14] <kirkland> mathiaz: but some comprehensive documentation, with screenshots, would be a brilliant addition to the server guide [16:14] * mathiaz nods [16:14] <mathiaz> [TOPIC] Review ServerGuide for Intrepid [16:14] <MootBot> New Topic: Review ServerGuide for Intrepid [16:14] <kirkland> (assuming we test it and it works!!!) [16:14] <jdstrand> kirkland: (sorry for the late response) I'd ahve to agree with mathiaz about testing on real hardware. people have testing environments just like we use kvm, so I think you'll get some takers-- especially ones that care about this (they would likely test on real hardware anyway before deploying) [16:14] <mathiaz> sommer: are you tracking somewhere the section you'd like to add to the guide ? [16:15] <mathiaz> sommer: (eg the raid install we've just talked about) [16:15] <sommer> https://wiki.ubuntu.com/IntrepidServerGuide [16:15] <sommer> mathiaz: I think that's it... the wiki is being slow for me [16:16] <sommer> yep that's the link [16:16] <sommer> the "Confirmed" section has the list [16:16] <mathiaz> sommer: great [16:16] <mathiaz> sommer: could you mark the section that need to be reviewed ? [16:16] <sommer> mathiaz: sure will do [16:17] <mathiaz> sommer: so that they can be easily identified as a task for new contributors [16:17] <mathiaz> sommer: I've also updated the instructions to update the server guide [16:17] <mathiaz> sommer: and how to check out the bzr branch [16:18] <mathiaz> https://wiki.ubuntu.com/ServerTeam/KnowledgeBase#head-0ae127e06ffba31c94b458fbef6eb033e5d8461e [16:18] <sommer> mathiaz: looks good, I've just started using bzr as more than a svn clone Smile :) [16:19] <mathiaz> we should try to get reviews to be sent via bzr branches [16:19] <sommer> mathiaz: works for me... I also heard back from mdke, and he was going to try and update doc.u.c [16:19] <mathiaz> sommer: anything else new in the server guide for intrepid ? [16:19] <mathiaz> sommer: great [16:20] <sommer> mathiaz: I've updated the Likewise-open section based on the discussions at UDS [16:21] <sommer> mathiaz: other than that, the LDAP section has been updated somewhat [16:21] <sommer> I'll mark those as needed review on the intrepidserverguide page [16:21] <mathiaz> sommer: excellent ! [16:21] <mathiaz> [TOPIC] Migrate new installs and upgrades of client and server packages to use SSL v3 or TLS [16:21] <MootBot> New Topic: Migrate new installs and upgrades of client and server packages to use SSL v3 or TLS [16:21] <ivoks> that's done Smile :) [16:22] <zul> ivoks: so there is no need to change dovecot and apache? [16:22] <ivoks> zul: no [16:22] <mathiaz> ivoks: yop - with an openssl upload [16:22] <zul> ivoks: sweet thats two off my todo list [16:22] <mathiaz> that disable sslv2 [16:22] <ivoks> zul: none of our libraries knows wtf is sslv2 anymore Big Grin :) [16:22] <zul> heh [16:23] <mathiaz> that's all I have from last week meeting [16:23] <mathiaz> is there anything else related to last week items ? [16:23] <kirkland> mathiaz: I'd like to thank the community for a strong response on the call for help with the init script status actions [16:24] <mathiaz> [TOPIC] Add 'status' action to server init scripts [16:24] <MootBot> New Topic: Add 'status' action to server init scripts [16:24] <kirkland> mathiaz: you can see a ton of progress since last week on https://wiki.ubuntu.com/InitScriptStatusActions#head-75448d0a482c28a007476b3be2394d9db3213662 [16:24] <james_w> kirkland: hi, I have a question about this one. [16:24] <mathiaz> kirkland: right - the response has been great [16:24] <mathiaz> kirkland: thanks to your blog post IMO [16:24] <macd> mathiaz, NeilW still needs some reviews on his passenger package on REVU. [16:24] <kirkland> mathiaz: lots of patches still awaiting sponsorship, thanks to zul for sponsoring a goodly bucket of them [16:24] <macd> mathiaz, I think thats all that was left from last weeks points [16:24] <kirkland> james_w: what's up? [16:24] <mathiaz> kirkland: I'll dive into the bugs filed today and sponsor them [16:25] <kirkland> mathiaz: thanks [16:25] <mathiaz> macd: right - I haven't read neilw reply to my email - it's in my inbox though [16:25] <macd> mathiaz, and hes here as well [16:25] <james_w> I got pulled up on to debian-devel a couple of weeks ago when filing patches in Debian becuase I wasn't following Debian norms for this sort of thing. [16:26] * NeilW waves timidly. [16:26] <kirkland> james_w: how so? [16:26] <mathiaz> macd: ok - I'll add a point about rails to the agenda [16:26] <macd> mathiaz, thx [16:26] <james_w> I saw that you have forwarded a whole bunch of these patches to Debian, which is great, thank you, but I wondered if you were not following Debian conventions. [16:26] <kirkland> james_w: how so? [16:27] <kirkland> james_w: i'm not aware of conventions that I've violated.... [16:27] <kirkland> james_w: I haven't seen any response whatsoever from Debian on the individual init script patches I've submitted [16:27] <james_w> firstly, when filing a bunch of bugs on one subject you are requested to notify debian-devel of the fact. [16:27] <kirkland> james_w: the initial functional addition of status_of_proc() was accepted into the lsb-base package, after a few revisions [16:28] <kirkland> james_w: hmm, I've only filed 2 Debian bugs of this type, and one other person filed 1 [16:28] <james_w> secondly, slangasek pointed out in one of the bugs (not in response to your patch) that Debian might like to look at this as a whole, rather than individual packages acting unilaterally. [16:29] <kirkland> james_w: that would be great! [16:29] <james_w> kirkland: I don't wish to criticise you. I'm bringing this up mainly as I got stung with this recently, and I thought that you may not be aware of these conventions. [16:29] <kirkland> james_w: there is a Debian policy bug against this that is many years old, actually [16:30] <james_w> kirkland: I didn't know that. [16:30] <kirkland> james_w: okay, thanks, i'll ping you offline about the best way to push this en masse to Debian [16:30] <kirkland> james_w: let me get you a bug #.... [16:31] <mathiaz> james_w: thanks for the warning [16:31] <kirkland> james_w: 2 of them... one from 2003: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=208010 [16:31] <ubottu> Debian bug 208010 in debian-policy "change =A710.4 "set -e OR check return status" to=" [Wishlist,Open] [16:31] <james_w> that works for me, I didn't want you to get flamed for trying to do the right thing. I think it's one area in our relationship with Debian that may be a bit neglected. [16:31] <kirkland> james_w: and this one from 2005: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=291148 [16:31] <ubottu> Debian bug 291148 in debian-policy "change =A710.4 "set -e OR check return status" to=" [Unknown,Open] [16:31] <mathiaz> let's move on [16:32] <mathiaz> [TOPIC] Call for WP testers [16:32] <MootBot> New Topic: Call for WP testers [16:32] <kirkland> james_w: I have replied to the later, to no response in 1.5 months [16:32] <mathiaz> nijaba: ? [16:32] <ivoks> wp? [16:32] <nijaba> I have a few white papers that are coming out [16:33] <nijaba> There are on different subjects, such as automated deployment [16:33] <nijaba> or clustering mysql [16:33] <nijaba> before we go on publishing them, I'd love for some volunteers to proofread/test them [16:34] <mathiaz> nijaba: where/how will you publish the wp ? [16:34] <ivoks> nijaba: me Smile :) [16:34] <nijaba> they'll be published on the ubuntu web site as pdf once they are ready [16:34] * sommer volunteers [16:34] <nijaba> as they target corporate customer [16:34] <nijaba> whome have a hight tendancy not to trust wikis [16:35] <zul> do they need real hardware or can they use a vm for testing it [16:35] <ivoks> nijaba: i'd like mysql clustering and redhat cluster suite, if available Smile :) [16:35] <mathiaz> nijaba: ok [16:35] * nijaba blames his keyboard for typos [16:35] <nijaba> ivoks: RHCS: still in the works, but I'll write you down for mysql [16:36] <mathiaz> nijaba: for the testing phase, is there a list of available wp ? [16:36] <mathiaz> nijaba: how should potential testers notify their interest in doing a review ? [16:36] <ivoks> nijaba: if you need help with rhcs, ping me... [16:36] <nijaba> mathiaz: atm, I have 2, 3 mores are coming, I'll inform the commuity when they are ready [16:36] <nijaba> ivoks: I will point ShangWu to you then Wink ;) [16:37] <nijaba> ivoks: he is the one doing all the work [16:37] <ivoks> nijaba: we already established a contact Big Grin :) [16:37] <nijaba> ivoks: great Smile :) [16:37] <mathiaz> nijaba: so as of now, what do you expect ? [16:38] <mathiaz> nijaba: should people send you an email in they're interested ? [16:38] <mathiaz> nijaba: for the mysql cluster and redhat cluster wp [16:38] <nijaba> mathiaz: volunteers to test automated deployments and mysql clustering WP [16:38] <mathiaz> nijaba: ok - they should send you an email ? [16:38] <nijaba> yes, email would be perfect [16:38] <mathiaz> nijaba: great [16:39] <mathiaz> let's move on [16:39] <mathiaz> [TOPIC] Server survey, Intel interested to partner [16:39] <MootBot> New Topic: Server survey, Intel interested to partner [16:39] <mathiaz> nijaba: ? [16:39] <nijaba> soren, as we are still waiting for IS to provide a host for the survey [16:40] <nijaba> more people are getting aware that we are gearing up for it [16:40] <nijaba> Intel contacted me last week, asking if they could join in [16:40] <mathiaz> nijaba: in which format would this work out ? [16:41] <nijaba> as we already agreed for RedMonk as an analyst firm, I told them I would haveto ask for your opinion first [16:41] <mathiaz> nijaba: RedMonk - didn't know about it Wink ;) [16:41] <nijaba> mathiaz: I think they want to add a hardware related question [16:42] <nijaba> mathiaz: we did not know the name of them, but that's whom the analyst firm we talked about is [16:42] <mathiaz> nijaba: ok - so Intel would join in [16:42] <nijaba> mathiaz: I do not know, what do other think? does it make sense? [16:42] <mathiaz> nijaba: we'd add a question about hardware, and they would ... ? [16:43] <ivoks> it does make sense [16:43] <nijaba> they would blog about it, allow us to use their logo... whatever we me reasonably think of... [16:43] <nijaba> s/me/may [16:44] <mathiaz> nijaba: as long as they can help us getting more exposure and getting people to take the survey, I'm ok with it [16:44] * nijaba blames his keyboard even more [16:44] <macd> SOunds like very good exposure [16:44] <nijaba> any objections? [16:45] <ivoks> no objections, just be careful Wink ;) [16:45] <sommer> sounds like a good idea to me [16:45] <mathiaz> nijaba: nope - the more exposure we have, the better Big Grin :) [16:45] <nijaba> ivoks: I know, I worked for them Wink ;) [16:46] <ivoks> nijaba: there are people runing ubuntu-server on other chips too, so no one wants to create impression that ubuntu is great with intel, but not so great with others Smile :) [16:46] <nijaba> ivoks: sure, I'll make sure their question is "vendor neutral" [16:46] <mathiaz> nijaba: could you come up with the question you wanna add to the survey and ask for review here ? [16:46] <nijaba> and if amd wants to join, they are welcome Wink ;) [16:47] <nijaba> mathiaz: I will, for sure [16:47] <ivoks> nijaba: or sun Smile :) [16:47] <nijaba> ivoks: right [16:47] <mathiaz> [ACTION] nijaba to write up a question related to hw and submit it for review to the rest of the server team [16:47] <MootBot> ACTION received: nijaba to write up a question related to hw and submit it for review to the rest of the server team [16:47] <mathiaz> let's move on [16:47] <mathiaz> [TOPIC] RAILS integration [16:47] <MootBot> New Topic: RAILS integration [16:48] <mathiaz> macd: ? what's the status [16:48] <macd> mathiaz, waiting on passenger to get in [16:48] <mathiaz> macd: IIRC there was some more discussion about the issue on rubygems in after last week meeting [16:48] <macd> NeilW rolled a new package that contains the fixes we talked about last week [16:49] <mathiaz> macd: NeilW: I'll review the passenger package [16:49] <macd> mathiaz, I was around for most of that, NeilW can be more elaborate on what the exact issue is [16:49] <NeilW> There are a number of things that Ubuntu needs to do to get Rails working. [16:50] <NeilW> At the moment pratically everybody ignores the apt packages and just installs everything via RubyGems. [16:50] <NeilW> That's for a few reasons; [16:50] <NeilW> Firstly, the current Rubygems package doesn't place gem installed binaries on the path. [16:50] <NeilW> So the likes of Capistrano just doesn't work. [16:51] <macd> Most people use Cap to deploy/configure their apps [16:51] <NeilW> Secondly the apt packages in place don't integrate with the rubygems database. [16:51] <macd> So its a pretty important point [16:51] <NeilW> So gem will pull in a second copy of Rails on top of the apt package [16:51] <NeilW> and chaos ensues. [16:51] <persia> Could apt-installed packages put something in the gems DB in the postinst? [16:52] <NeilW> persia: That's how to fix it - but the packages we're getting from Debian don't do that. [16:53] <NeilW> You don't even need postinst. You just install a marker file in a particular place. [16:53] <NeilW> Gem will then see it as installed. [16:53] <NeilW> Finally there is no real decision on how Ubuntu/Debian is going to handle the Ruby1.8 to Ruby1.9 transition. [16:54] <NeilW> Gem is integral to Ruby1.9, but has been removed from the ruby package. [16:54] <NeilW> and a new gem package created that handles both versions. [16:55] <NeilW> unfortunately gem1.8 and gem1.9 don't really know about each other and they will clash if you try and point their binaries at the same place. [16:55] * ivoks waves... it's 35C in this room [16:56] * nijaba takes a slice of a well roasted ivoks Wink ;) [16:56] <mathiaz> NeilW: are you discussing this issue with the debian maintainers ? [16:56] <mathiaz> NeilW: if so where ? [16:57] <NeilW> I've discussed the point with Lucas who I think is trying to move Debian away from their current stance on Rubygems. [16:57] <mathiaz> NeilW: we try to follow debian whenever we can [16:57] <mathiaz> NeilW: great. [16:58] <NeilW> Yes I know. [16:58] <mathiaz> NeilW: however if things are totally broken we're ready to make changes [16:58] <NeilW> I think you may have to take the lead in this case. My feeling is that there is an impasse. [16:59] <mathiaz> NeilW: it may worth trying to describe what would be the proper way of doing things from a end user perspective [16:59] <macd> Is there a point where we decide to no longer merge/sync a package from upstream and decide to maintain it totally at this level? [16:59] <mathiaz> macd: well it depends on the situation and if we have the man power to do it [16:59] <mathiaz> macd: gnome is such an example [16:59] <NeilW> apt-get install rails-server [17:00] <NeilW> It's my dream. [17:00] <mathiaz> macd: but we have a gnome maintainer in the ubuntu project [17:00] <mathiaz> NeilW: right - if you could start a wiki page to outline what it would be like [17:00] <macd> I think we can stick that on the rails spec page? [17:00] <NeilW> I thought we already had one? The Rails spec. [17:00] <mathiaz> NeilW: especially the interaction between rubygems and the installed package [17:01] <mathiaz> macd: NeilW: yes - you can use the rails spec. [17:01] <macd> We have what changes need to happen, bugs filed against them, some packages rebuit to fix it [17:01] <macd> not sure what else needs to happen here [17:01] <mathiaz> macd: ok - so may *I* should go back to the Rails spec and have a look at it [17:01] <mathiaz> macd: *may be* I [17:01] <macd> I gotcha ;P [17:02] <mathiaz> macd: I'll have a look at the spec page [17:02] <NeilW> mathiaz: What would you do with such a description. How does the descision process work? [17:02] <mathiaz> NeilW: I'll reply to your email [17:02] <NeilW> ok? [17:02] <mathiaz> NeilW: we should probably move the discussion to a public mailing list to get more input on that [17:03] <mathiaz> NeilW: it would help me to grasp the problem - may be it's already there and I just haven't read the wiki page lately [17:03] <mathiaz> NeilW: I'll reply to your email and we'll take it from there. [17:03] <macd> can you CC server ML? [17:03] <mathiaz> macd: will do [17:03] <macd> and we can get it all on there [17:04] <NeilW> Thanks [17:04] <mathiaz> @schedule [17:04] <ubottu> mathiaz: Schedule for Etc/UTC: 31 Jul 14:00: Mentoring Reception | 31 Jul 16:00: Ubuntu Mobile | 02 Aug 13:00: Xubuntu Community | 03 Aug 18:00: Mozilla Team | 05 Aug 22:00: Community Council | 06 Aug 20:00: Maryland LoCo IRC [17:04] <macd> looks like were clear for almost a year Wink ;) [17:05] <mathiaz> [TOPIC] Open discussion [17:05] <MootBot> New Topic: Open discussion [17:05] <mathiaz> anyone wants to add something ? [17:06] <mathiaz> [TOPIC] Agree on next meeting date and time [17:06] <MootBot> New Topic: Agree on next meeting date and time [17:06] <mathiaz> same place, same time, next week ? [17:07] <sommer> sure [17:07] <nijaba> If anyone is in san francisco next week, come and say hi at our linuxworld booth! [17:07] <mathiaz> all right then [17:07] <nijaba> and please accept my apologies for next week meeting, as I'll be stuck there I beleive [17:08] <mathiaz> see ya all next week, same time, same place [17:08] <mathiaz> thanks all for coming [17:08] <sommer> nijaba: heh sounds awful Smile :) [17:08] <mathiaz> #endmeeting }}}