Ubuntu Wiki

Agenda

Items we will be discussing:

• Review ACTION points from previous meeting.

• Review progress made on the specification listed on the Roadmap.

• Rapache presentation and discussion.

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

Minutes

Add 'status' action to server init scripts

kirkland generated a list of init scripts that could have a status action added to them. The wiki page also outlines how to add a status action and file a bug in LP to ask for inclusion in the relevant package.

Anyone interested in helping out should head to the wiki page.

Clamav and spamassassin in main

ScottK reminded that there are a couple of MIRs that still need to be written before clamav and spamassassin can be included into main. The wiki page tracks all the remaining packages that should have a MIR written for. ScottK can be contacted to get an initial review of the MIR. ivoks and sommer volunteered to write some of the MIRs.

Rapache presentation and discussion

tacone gave an overview of the rapache project emgent and him started after UDS. Rapache is an Apache configurator gui. The goal is to lower the entry barrier for former windows system administrators used to configure IIS with a GUI.

Review ServerGuide for Intrepid

sommer reported that he almost finished a new section about Kerberos. He also asked for a review of the Samba section.

Ubuntu VM builder

soren is still working on the rewrite in python of ubuntu-vm-builder. He is currently adding Xen support. mathiaz asked if there was some documentation available. soren said there wasn't any - any help in this area is welcomed.

ACTION: soren to write a short README file to help people get started with the new ubuntu-vm-builder.

Migrate openldap configuration to cn=config

mathiaz spent most of his time last week on implementing cn=config support to the openldap package. Both new installs and upgrades are supported. He sent a patch to the Debian maintainers and is waiting for their feedback.

He also plans to look into FreeIPA once the cn=config migration is completed.

Boot Support for Degraded RAID

kirkland has some working code in his ppa. More testing of the mdadm package is welcome.

ACTION: kirkland to update the wiki page BootDegradedRaid with some testing instructions.

RAILS integration

macd reported that mod_rails has been packaged and uploaded to REVU. mathiaz reviewed it and sent his feedback to Neil (the packager). Overall it looks good.

macd also mentioned the discussions he's having with the Debian maintainers for ruby/rubygems that is taking place in bug 145267. There was some discussion about the issue which boils down to a PATH issue. The debian gems and the source installed gems don't end up in the same place. Rails apps looks in a specific place and rails isn't capable of looking in more than one place. soren, macd and persia discussed the path issue and deferred it to #ubuntu-server after the meeting.

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

ivoks prepared patches for a couple of packages to disable sslv2 in their configuration. He also sent an email on ubuntu-devel about disabling sslv2 directly in the openssl package. Discussion is ongoing, with a proposal to create an openssl-sslv2 package in universe that would be built with sslv2 enabled.

ACTION: ivoks to prepare a patch for the openssl package to disable sslv2.

Agree on next meeting date and time

Next meeting will be on Tuesday, July 29th 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]
[16:02] <mathiaz> Today's agenda: https://wiki.ubuntu.com/ServerTeam/Meeting
[16:02] <mathiaz> The previous meeting minutes: https://wiki.ubuntu.com/MeetingLogs/Server/20080715
[16:02] <mathiaz> [TOPIC] Add 'status' action to server init scripts
[16:02] <MootBot> New Topic:  Add 'status' action to server init scripts
[16:03] <mathiaz> kirkland: ^ ?
[16:03] <kirkland> mathiaz: howdy, yessir
[16:03] <mathiaz> kirkland: I've noticed you've generated a first list of init scripts
[16:03] <kirkland> mathiaz: Right, so Debian has accepted, in toto, the status_of_proc() functionality, and we've merged the updated lsb-base package into Intrepid
[16:03] <mathiaz> kirkland: however this list should be trimmed as some ot them don't need to have status action added to them
[16:04] <kirkland> mathiaz: and I did generate a rough list of init scripts, as found in default installs of Intrepid server/ubuntu-desktop/kubuntu/xubuntu
[16:04] <kirkland> mathiaz: that's absolutely true.  i'm hoping we can leverage the nature of the wiki to make that happen.
[16:04] <kirkland> mathiaz: i'd encourage people to add/remove scripts to/from that page, as they review it
[16:04] <ScottK2> mathiaz: I have to go to a meeting very shortly.  I'd appreciate just a minute or two early in the meeting before I have to go.
[16:04] <kirkland> https://wiki.ubuntu.com/InitScriptStatusActions
[16:05] <mathiaz> kirkland: I'd filter the list and reduce it to the package that have a daemon running so that the status_of_proc can be used in there
[16:05] <kirkland> mathiaz: I'd like to see links to bug numbers, if and when people work on those
[16:05] <mathiaz> ScottK2: go ahead
[16:05] <ScottK2> Thanks
[16:05] <kirkland> mathiaz: cool, would you do the first pass of filtering, then?
[16:05] <ScottK2> A lot of people said they wanted clamav/spamassassin in Main.
[16:06] <mathiaz> kirkland: I'll find someone to do it - I've got a couple of mentoring requests
[16:06] <ScottK2> Looking at https://wiki.ubuntu.com/ClamavSpamassassinInMain, it's been sommer and I so far doing the work.
[16:06]  * ScottK2 does not have time to do all these MIR due to $WORK, so this is a call (again) for people who want this to invest some time in it.
[16:06] <ScottK2> MIR are not that hard to do and I'll be glad to answer questions/review stuff.
[16:07] <ScottK2> mathiaz: That's it.
[16:07] <mathiaz> ScottK2: ok - great !
[16:07] <mathiaz> [ACTION] mathiaz to ask prospective mentors to trim down the list of init scripts.
[16:07] <MootBot> ACTION received:  mathiaz to ask prospective mentors to trim down the list of init scripts.
[16:07] <ivoks> ok, i'll help, but don't expect miracles... me does the $WORK too :D
[16:08] <sommer> and I should have more time this week/weekend for some mir action
[16:08] <mathiaz> ivoks: are you refering to MIR or status init script ?
[16:08] <ivoks> MIR
[16:08] <mathiaz> ivoks: ok.
[16:08] <mathiaz> I'll mention the MIR in the minutes again
[16:08] <mathiaz> [TOPIC] Integration of SASL and Postfix
[16:09] <MootBot> New Topic:  Integration of SASL and Postfix
[16:09] <mathiaz> ivoks: did you have time to discuss the issues with lamont ?
[16:09] <ivoks> i've sent an email, but didn't get response
[16:09] <lamont> mathiaz: I expect he has, I haven't had time...
[16:09] <ScottK2> ivoks: Thanks.
[16:09] <mathiaz> ivoks: lamont ok - great
[16:10] <mathiaz> That's all from the last meeting - anyone wants to add something ?
[16:10] <lamont> ivoks: I expect I'll have time tomorrow evening or so to at least look at thing
[16:10] <lamont> s
[16:10] <ivoks> lamont: ok
[16:11] <mathiaz> emgent: around ?
[16:11] <tacone> mathiaz: no
[16:11] <tacone> I am here to talk in his place.
[16:11] <mathiaz> tacone: are you here for rapache ?
[16:11] <tacone> he will try to get here later if he can
[16:11] <mathiaz> [TOPIC] Rapache presentation and discussion
[16:11] <MootBot> New Topic:  Rapache presentation and discussion
[16:11] <tacone> omg, already, ok
[16:11] <mathiaz> tacone: we can postpone if you want
[16:11] <tacone> no no:)
[16:11] <tacone> after the UDS emgent came back and told me that
[16:12] <tacone> people felt the need to have more server-related GUIs
[16:12] <tacone> from that born rapache.
[16:12] <tacone> Rapache is an Apache configurator gui
[16:13] <tacone> I am unsure about the scope of it. If it falls in ubuntu-server or not, because is a desktop application.
[16:13] <tacone> our goal is to lower the entry barrier for former windows system administrators used to configure IIS with a program
[16:14] <mathiaz> tacone: does it support administrating remote servers ?
[16:14] <mathiaz> tacone: it seems that it's requires a desktop environment to run it
[16:14] <tacone> mathiaz: currently NOT. that's in the road map as a very high priority
[16:14] <tacone> that's a nice point, let me elaborate
[16:14] <mathiaz> tacone: screen shots somewhere ?
[16:14] <tacone> LOL
[16:14] <tacone> too fast :)
[16:14] <emgent> hello
[16:15] <mathiaz> tacone: I don't see any on the rapache website
[16:15] <mathiaz> http://www.rapache.org/
[16:15] <tacone> http://www.stefanoforenza.com/blog/wp-content/uploads/2008/07/rapache-main.png
[16:15] <tacone> http://www.stefanoforenza.com/blog/wp-content/uploads/2008/07/rapache-edit-window.png
[16:15] <tacone> http://www.stefanoforenza.com/blog/wp-content/uploads/2008/07/rapache-modules.png
[16:15] <emgent> [LINK] https://launchpad.net/rapache
[16:15] <tacone> http://www.stefanoforenza.com/blog/wp-content/uploads/2008/07/rapache-syntax-highlightining.png
[16:15] <tacone> I am not sure also how much time I have to talk here, so feel free to stop or warm at any time
[16:16] <tacone> we put together a blueprint: https://wiki.ubuntu.com/EasyGuiApacheSetup
[16:16] <mathiaz> tacone: ok
[16:16] <tacone> Supporting remote servers: that's in the roadmap but not trivial
[16:17] <ivoks> does augeas support remote configuration?
[16:17] <tacone> so we just started on localhost
[16:17] <emgent> true, anyway 0.5. version is available on rapache-devel PPA
[16:17] <ivoks> if yes, then moving to augeas would solve that problem :)
[16:17] <tacone> ivoks: not on my knowledge
[16:17] <mathiaz> tacone: ok - great - I guess most of us are server people - so we're more interested in non-GUI environment
[16:17] <tacone> we thought about augeas but we'd like to see how things develop with it.
[16:17] <mathiaz> ivoks: not that know of
[16:17] <emgent> https://launchpad.net/~rapache-devel/+archive
[16:17] <soren> I think Func is meant to add remoteness to augeas. I'm not sure, though.
[16:18] <mathiaz> tacone: ok - seems like a good start - you'd probably need to update the wiki page
[16:18] <tacone> mathiaz: actually I am not sure were to talk about rapache. in ubuntu-server, ubuntu-desktop, lol
[16:18] <tacone> mathiaz: we have too much stuff around, it's falling outdated :-)
[16:18] <mathiaz> tacone: it seems that this is a case where you can discuss on both channels
[16:18] <mathiaz> tacone: probably ubuntu-server is a good place to start with
[16:18] <tacone> mathiaz: we're starting with ubuntu-server :)
[16:19] <ivoks> right, we'll help on backend :)
[16:19] <mathiaz> tacone: that's where you'll find most of your users
[16:19] <tacone> nice.
[16:19] <tacone> well rapache won't require sysadmin to install anything on remote servers
[16:19] <mathiaz> anything else on rapache ?
[16:19] <tacone> but an ssh certificate, I think that could be quite valuable
[16:20] <tacone> as not every sysadmin wants to have estraneous programs on their machine
[16:20] <tacone> I guess we're done for now, unless you have more questions
[16:20] <mathiaz> tacone: not for me
[16:20] <tacone> emgent: around ? anything other to say ?
[16:21] <mathiaz> tacone: thanks for you presentation ! keep up the good work
[16:21] <tacone> as the last thing one question from me
[16:21] <tacone> where should we continue the discussion ? shall we talk with ubuntu-desktop people ?
[16:22] <mathiaz> tacone: I think #ubuntu-server is the best place to keep discussing
[16:22] <mathiaz> tacone: that's where you'll find most of your end-users
[16:22] <tacone> nice, you'll see us there often then :)
[16:22] <mathiaz> let's move on
[16:22] <tacone> thank you very much everyone for listening.
[16:22] <mathiaz> tacone: you're welcome :)
[16:22] <mathiaz> [TOPIC] Review progress made on the specification listed on the  Roadmap.
[16:23] <Moot2> New Topic:  Review progress made on the specification listed on the  Roadmap.
[16:23] <mathiaz> Our roadmap: https://wiki.ubuntu.com/ServerTeam/Roadmap
[16:23] <mathiaz> [TOPIC] Review ServerGuide for Intrepid
[16:24] <mathiaz> sommer: what's going on this front ?
[16:24] <sommer> mathiaz: good, almost have Kerberos wrapped up
[16:24] <Moot2> New Topic:  Review ServerGuide for Intrepid
[16:24] <mathiaz> sommer: anything that could be reviewed/tested ?
[16:24] <mathiaz> sommer: I hadn't had time to look into the bzr changes
[16:24] <sommer> samba stuff, definitely needs some review
[16:25] <mathiaz> sommer: it's still on my todo list
[16:25] <mathiaz> sommer: ok
[16:25] <sommer> mathiaz: cool, there's still time
[16:26] <mathiaz> [TOPIC] Ubuntu VM builder
[16:26] <Moot2> New Topic:  Ubuntu VM builder
[16:26] <mathiaz> soren: ^ ?
[16:26] <soren> Ah, yes.
[16:26] <soren> I'm working on a complete rewrite in Python.
[16:27] <soren> This is mainly to be able to use it as more of a library (to be able to embed it in various things).
[16:27] <soren> Another side effect is that we can add lots more logic to it.
[16:27] <soren> Of course, we could have done that in bash too, but it's loads easier in Python.
[16:28] <ivoks> and more cool
[16:28] <soren> It's shaping up rahter nicely, but I'm adding Xen support to it, and that involves changing some core things about the disk handling stuff.
[16:28] <mathiaz> soren: is there anything ready for more widespread consumption/testing ?
[16:29] <soren> People are very welcome to grab the code off of launchpad and play around with it. It should be able to build dapper, feisty, gutsy, hardy, and intrepid, kvm VM's.
[16:29] <soren> ...everything else will be added over the next couple of weeks.
=== Moot2 is now known as MootBot
[16:29] <mathiaz> soren: is there a package in intrepid and/or hardy ?
[16:29] <mathiaz> soren: some documentation somewhere ?
[16:30] <soren> Not right now, no.
[16:30] <soren> I'll probably upload it to Intrepid as a new package, and then yank the old one out when they're on par.
[16:31] <macd> Am I too late for server?
[16:31] <mathiaz> soren: ok - once there is some basic documentation written, testing by other can start
[16:31] <soren> The cli is rather self-documenting, I think.
[16:31] <mathiaz> macd: not really - we're just hald-way through
[16:31] <soren> Documentation contributions are also very, very welcome :)
[16:32] <macd> mathiaz, great I just want to hit on the ruby stuff going on at the end if thats ok
[16:32] <mathiaz> soren: at least a README file or Getting started even if it's 10 lines
[16:32] <soren> Good point.
[16:32] <mathiaz> macd: ok - I'll add a point to the agenda
[16:32] <macd> ty
[16:32] <soren> I'll whip something up.
[16:33] <mathiaz> soren: just so that people can get started with it and play around
[16:33] <soren> Right.
[16:34] <mathiaz> [ACTION] soren to write a short README file to help people get started with ubuntu-vm-builder
[16:34] <MootBot> ACTION received:  soren to write a short README file to help people get started with ubuntu-vm-builder
[16:34] <mathiaz> [TOPIC] Migrate openldap configuration to cn=config
[16:34] <MootBot> New Topic:  Migrate openldap configuration to cn=config
[16:34] <mathiaz> I'
[16:34] <soren> You'
[16:34] <soren> :p
[16:34] <mathiaz> I worked on that during last week
[16:35] <mathiaz> I got a patch that supports new install and upgrade and sent it to the debian maintainers
[16:35] <mathiaz> I'm waiting for feedback on it
[16:36] <mathiaz> Once it's there, it should be easier to implement things like schema/module loading
[16:36] <mathiaz> I may look into FreeIPA
[16:36] <mathiaz> to see if we can get something in ubuntu
[16:37] <mathiaz> [TOPIC] Boot Support for Degraded RAID
[16:37] <MootBot> New Topic:  Boot Support for Degraded RAID
[16:37] <mathiaz> kirkland: ?
[16:37] <kirkland> mathiaz: I have a test that works on KVMs
[16:37] <kirkland> mathiaz: s/test/patch/
[16:38] <kirkland> mathiaz: i'm going to test on real hardware, now that I'm back home, and submit
[16:38] <kirkland> mathiaz: there's more to be done in the installer
[16:38] <mathiaz> kirkland: do you know to whom it should be submit for review ?
[16:39] <kirkland> mathiaz: kees said kees -> cjwatson -> evand
[16:39] <kirkland> mathiaz: the first in the series of patches is relatively minor
[16:39] <kirkland> mathiaz: i'll attach it to the bug today, assuming the real-hw test goes as expected.
[16:40] <mathiaz> kirkland: ok. Is it worth asking for more testing  now ?
[16:40] <mathiaz> kirkland: or we'd better wait for the inclusion in intrepid ?
[16:40] <kirkland> mathiaz: well, i have a pkg in my PPA
[16:40] <kirkland> mathiaz: i'm certainly open to any feedback on that (mdadm)
[16:40] <mathiaz> kirkland: what is required to do testing ?
[16:40] <kirkland> mathiaz: huah...  well, in KVM it's extraordinarily complicated
[16:41] <soren> why?
[16:41] <kirkland> mathiaz: unless soren has some magic to make that easier
[16:41] <mathiaz> kirkland: let's assume on real hw then
[16:41] <kirkland> mathiaz: really hardware, you just need 2 physical disks (can be flash media > 1GB)
[16:41] <kirkland> soren: i'm open to advice as to providing a new KVM machine to-be-installed with 2 physical disks from the start
[16:42] <kirkland> soren: at least through virt-manager, that's not easy
[16:42] <kirkland> soren: so for my testing, i've been doing an install with 1 machine, raiding two partitions on the same disk together
[16:42] <soren> kirkland: Well, it's a GUI, so there's a lot of clicking to do anything.. :) Is it more difficult than expected?
[16:42] <kirkland> soren: then after install, adding another physical disk to the VM, and then to the RAID
[16:42] <kirkland> soren: and then, I have a setup I can work with
[16:43] <mathiaz> kirkland: so you'd install intrepid on a two-disk system that boots from a raid array on /, install your ppa adm package, poweroff, remove one disk, power on and it should still be booting ?
[16:43] <soren> Huh?
[16:43] <soren> Oh, because you can't add two disks to begin with?
[16:43] <kirkland> mathiaz: yes assuming you have either a) edited the config file to tell it to do so, or b) pass it the appropriate kernel option
[16:43] <soren> I highly recommend not using virt-manager for the initial install then. :)
[16:43] <kirkland> soren: right...  i'd like to start off an install with two .img disks
[16:44] <mathiaz> soren: kirkland: could you figure out a way to simplify the testing instructions using KVM ?
[16:44] <kirkland> mathiaz: the default behavior should be as it always has...  on missing disk, drop to initramfs prompt
[16:44] <stgraber> kvm -hda blah.img -hdb blah1.img -m 512 -cdrom ubuntu-server.iso
[16:44] <mathiaz> kirkland: could you add some testing instructions to the wiki page ?
[16:44] <kirkland> mathiaz: sure
[16:45] <mathiaz> kirkland: great
[16:45] <kirkland> mathiaz: i'm actually a little slammed at the moment
[16:45] <kirkland> mathiaz: but I'll see what I can do
[16:45] <mathiaz> [ACTION] kirkland to update the wiki page BootDegradedRaid with some testing instructions
[16:45] <MootBot> ACTION received:  kirkland to update the wiki page BootDegradedRaid with some testing instructions
[16:46] <kirkland> stgraber: thanks, i'll run with that
[16:46] <soren> qemu-img create disk1.img 5G && qemu-img create disk2.img && kvm -drive file=disk1.img,if=ide,index=0 -drive file=disk2,fi=ide,index=1 -drive file=intrepid.iso,if=ide,media=cdrom -boot d
[16:46] <mathiaz> kirkland: well it may not be the best moment then
[16:46] <soren> Whoops.
[16:46] <soren> That wasn't supposed to go here :)
[16:46] <mathiaz> kirkland: I'm interested in figuring out how to test this thing
[16:46] <kirkland> soren: that's your login passphrase, right?
[16:46] <kirkland> :-P
[16:46] <mathiaz> kirkland: and document it
[16:46] <kirkland> mathiaz: okay
[16:47] <soren> kirkland: sssh... don't tell anyone.
[16:47] <soren> :)
[16:47] <ivoks> what? where?
[16:47] <ivoks> :)
[16:47] <sommer> hah
[16:48] <mathiaz> [TOPIC] RAILS integration
[16:48] <MootBot> New Topic:  RAILS integration
[16:48] <mathiaz> macd: /
[16:48] <mathiaz> macd: ?
[16:48] <macd> yes
[16:48] <macd> sorry, was on another screen
[16:48] <macd> First off, mod_rails is packaged, mathiaz looked at it, we'll be making those changes as soon as Neil (the packager) gets back form vac.
[16:49] <macd> mathiaz, we need some notes on REVU/LPP bug for that
[16:49] <mathiaz> macd: right - I've sent my notes to neil as I wasn't able to login in REVU :/
[16:49] <macd> and if we could get a few other people looking at it for some more acks, we can be done with that
[16:49] <macd> mathiaz, your not the only one I cant either
[16:50] <macd> Weve been talking with the debian packager for ruby/rubygems though its becoming a heated discussion
[16:50] <mathiaz> macd: yes - are these discussions public ?
[16:50] <macd> Hes not a rails user at all so he dosnt really understand the paths for gems
[16:50] <macd> yes on bug #145267
[16:50] <ubottu> Launchpad bug 145267 in libgems-ruby "Add rubygems bin to PATH" [Low,Confirmed] https://launchpad.net/bugs/145267
[16:51] <mathiaz> macd: right - gems are problematic from the distribution POV
[16:51] <macd> are they ever :/
[16:52] <macd> In my mind the decision just needs to be made, either we rip gems from  universe, and rails people install from source (which most do) or fix it to behave
[16:52] <mathiaz> macd: IIUC there a some similar issue with python eggs
[16:53] <soren> gems to Ruby what Pear is to PHP, right?
[16:53] <mathiaz> soren: IIUC - yes
[16:53] <soren> Ok.
[16:53] <macd> Neil and I have been talking about fixing gems, and having something like apt-build build gems as they become updated
[16:53] <macd> soren, yes, pretty much
[16:53] <mathiaz> macd: and CPAN for perl
[16:53] <soren> macd: So it downloads some source code, builds it and installs it somewhere, right?
[16:54] <macd> soren, yep
[16:54] <macd> the problem being the debian gems, and the surce installed gems dont end up in the same place
[16:54] <mathiaz> macd: it's a general issue with scripting langages - it may be interesting to look at what other langages are doing
[16:54] <soren> Well, depending on how they designed it, it could be either really easy to shove the compiled stuff somewhere under /var or it could be really, really difficult. :)
[16:55] <macd> soren, I almost suggested the whole thing just be one package in /opt ;P
[16:55]  * soren cringes
[16:55] <ivoks> apt-gem?
[16:55] <macd> ivoks, thats a long term idea for sure
[16:55]  * soren wacks ivoks
[16:55] <macd> but I dont see that being in intrepid ;'P
[16:56] <macd> Does anyone have any other ideas?
[16:56] <soren> It's hard to say without knowing more about gems.
[16:56] <macd> Im about out, short of what I mentioned above with just ripping gems out for now
[16:56] <macd> soren, you can look at that bug I posted with gems, and it covers it very well
[16:57] <soren> Why is it a problem that ruby "things" from apt get installed in a different place than the ones isntalled by gems?
[16:57] <macd> Its seriously just a path
[16:57] <mathiaz> macd: https://bugs.launchpad.net/ubuntu/+source/libgems-ruby/+bug/145267 ?
[16:57] <ubottu> Launchpad bug 145267 in libgems-ruby "Add rubygems bin to PATH" [Low,Confirmed]
[16:57] <macd> soren, the rails app looks in a specific place
[16:57] <macd> soren, and there is no way short of creating symlinks that it can look somewhere else
[16:57] <soren> macd: So add another place to look?
[16:57] <macd> soren, rails isnt capable of looking in more than one place
[16:57] <ivoks> jesus...
[16:57] <macd> thats the caveat that the gems maintainer doesnt seem to grab either
[16:58] <soren> Fix it.
[16:58] <Koon> macd: I suppose it could be patched to do it
[16:58] <soren> :)
[16:58] <macd> I bet it could, and we filed a bug upstream with rails
[16:58] <ivoks> all languages support multipath
[16:58] <macd> but those guys arent budging.
[16:58] <Koon> we have kinda the same problem with maven, though it's a little worse (it downloads already-built blobs)
[16:58] <mathiaz> macd: do you have the upstream rails bug url ?
[16:59] <macd> Most gems are binary already, but some do compile against ruby1.8 or ruby1.9 which presents another problem with debian gems
[16:59] <macd> mathiaz, take a quick laugh I dont have my bookmarks sync'd
[16:59] <macd> ;)
[16:59] <macd> but I'll add it to the rails spec wiki page later today
[17:00] <soren> macd: Clueless question alert: Google mentions $LOAD_PATH which seems to be a list of places to look for "stuff".
[17:00]  * soren doesn't know ruby lingo, so "stuff" will have to do.. :)
[17:00] <mathiaz> macd: great - if you can add links to the relevant ressources on the web that talk about the issue, it would be very helpful to get up-to-speed on the issue
[17:00] <soren> Why is that variable insufficient?
[17:00] <mathiaz> @schedule
[17:00] <ubottu> mathiaz: Schedule for Etc/UTC: Current meeting: Server Team | 22 Jul 18:00: LoCo Council | 22 Jul 20:00: Ubuntu Web Presence Team | 22 Jul 21:00: EMEA membership meeting | 23 Jul 17:00: QA Team | 23 Jul 22:00: Platform Team
[17:00] <macd> soren, one sec
[17:01] <macd> # Add additional load paths for your own custom dirs
[17:01] <macd>   # config.load_paths += %W( #{RAILS_ROOT}/extras )
[17:01] <macd> that is for plugins
[17:01] <mathiaz> Ok - let's move on - as we're running out of time
[17:02] <soren> And gems are... what?
[17:02] <soren> Ok, we can talk in #ubuntu-server in a bit.
[17:02] <persia> soren: It's binaries, not libraries: it's perfectly possible for one gem to depend upon another and run binaries included in the gem depended upon.  This doesn't work in SUDO without mangling /etc/environment, which is not indended for automated modification post-install.
[17:02] <macd> sure
[17:02] <macd> persia, thats pretty spot on as to whats happenning
[17:02] <soren> persia: Oh, things have to be in $PATH, not ruby's library path or whatnot?
[17:02] <mathiaz> [TOPIC] Open Discussion
[17:02] <MootBot> New Topic:  Open Discussion
[17:03] <macd> soren, yes, exactly
[17:03] <mathiaz> Anyone wants to add something ?
[17:03] <persia> soren: Depending on the gem, possibly both (but I'm not continuing more, given the desire of the chair).
[17:03] <soren> -> #ubuntu-server
[17:03] <macd> we can continue in -server in about 20?
[17:04] <mathiaz> Anyone wants to add something ?
[17:04] <ivoks> openssl patches
[17:04] <mathiaz> [TOPIC] Migrate new installs and upgrades of client and server packages to use SSL v3 or TLS
[17:04] <MootBot> New Topic:  Migrate new installs and upgrades of client and server packages to use SSL v3 or TLS
[17:04] <ivoks> i've been working on them; only left are proftpd and lighttpd (both universe)
[17:05] <ivoks> so, this will work, but we would be far better of with disabling sslv2 in openssl
[17:05] <mathiaz> ivoks: you've started a thread on -devel about removing support for sslv2 in openssl
[17:05] <mathiaz> ivoks: what's the current situation ?
[17:05] <ivoks> smaller delta with debian and a sure thing we disabled it everywhere
[17:05]  * nealmcb1 waves from oscon
[17:05] <ivoks> there's a proposal that we should really disable sslv2
[17:06] <ivoks> so, atm, we are waiting for really good argument against that action :)
[17:06] <ivoks> (disable in openssl)
[17:06] <ivoks> as it is disabled in NSS and gnutls
[17:06] <mathiaz> ivoks: ok
[17:06] <soren> I like that idea, but I'm sort of worried about 3rd party stuff still requiring it.
[17:07] <ivoks> soren: openssl-sslv2 package in universe?
[17:07] <soren> That's what I'm thinking. That was a good idea.
[17:07] <ivoks> we already have packages like that
[17:07] <zul> im sitting on dovecot and apache patches until we get a resolution
[17:07] <soren> Built from the same source, so we don't have much of a maintenance overhead.
[17:07] <ivoks> apache is one example; same source, multiple binaries with different options
[17:08] <ivoks> zul: there's vsftpd patch too
[17:08] <zul> ivoks: ah ok
[17:08] <mathiaz> it may be worth preparing a patch to the openssl package for the -sslv2 option in universe
[17:08] <ivoks> mathiaz: i'm ready to do that
[17:09] <mathiaz> ivoks: great - it may help in the discussion: to see the code
[17:09] <mathiaz> [ACTION] ivoks to prepare a patch for the openssl package to disable sslv2
[17:09] <MootBot> ACTION received:  ivoks to prepare a patch for the openssl package to disable sslv2
[17:09] <ivoks> this is the only right way to do it
[17:10] <mathiaz> ivoks: thanks for your help
[17:10] <mathiaz> Anything else to add ?
[17:10] <mathiaz> [TOPCI] Open Discussion
[17:10] <ivoks> anyone using redhat-cluster-suite, please report any problems :)
[17:11] <ivoks> it turns out we have lots of redhat-ism in that package
[17:12] <mathiaz> [TOPIC] Agree on next meeting date and time
[17:12] <MootBot> New Topic:  Agree on next meeting date and time
[17:12] <zul> ivoks: not surprising :)
[17:12] <mathiaz> same place, same time, next week ?
[17:12] <ivoks> sure
[17:12] <kirkland> yup
[17:13] <Koon> ok
[17:13] <mathiaz> excellent then - see yu all next week here
[17:13] <mathiaz> keep up the good work
[17:13] <macd> thanks!
[17:13] <mathiaz> #endmeeting