20080812

Agenda

Items we will be discussing:

Minutes

Remove multiuser options and updating init scripts

zul reminded that the multiuser option is no longer supported in Ubuntu. james_w filed bugs to keep track of the packages that need to be fixed. They're pretty easy to fix, and any help will be appreciated. mathiaz suggested to write up a blog post about it.

ACTION: james_w to write a blog post about removing multiuser options.

New eBox packages

foolano gave an update on the ebox packages: ebox-mail, ebox-mailfilter, ebox-trafficshaping and ebox-webserver are good candidate to be included in intrepid. These packages are available in the ebox-unstable PPA. He asked for sponsoring to get them in universe before FeatureFreeze. soren suggested to check if the per-package upload ACL system could be used to grant foolano upload privileges just of the ebox packages.

ACTION: foolano to send an email to the MC to get upload rights to the ebox package

ISV's - VMware on Gutsy and Hardy

owh asked about the state of the vmware-server package available from Canonical's partner archive. A version for hardy hasn't been released yet. dendrobates stated that kernel abi changes are problematic and we are trying to reduce them to make it easier.

Ubuntu VM builder

soren finally got it to build a Xen based image. He'll put out an e-mail when it's ready for actual use. Developers are very welcome to pitch in. The "--help" option is quite verbose and should be a good starting point. The code can be found in the python-rewrite branch.

Migrate openldap configuration to cn=config

mathiaz uploaded version 2.4.11 to intrepid. The default configuration backend has been switched to the cn=config tree. Migration from slapd.conf to cn=config is being done during the upgrade.

Tomcat6 server stack support

Koon announced that tomcat6 had been uploaded to intrepid. It's currently sitting in the NEW queue. Testing is welcome.

He's also working on reducing openjdk-6-jre-headless dependencies so that it doesn't pull half of a desktop install in servers just to run tomcat6. He filed bugs and attached debdiffs to it. He is waiting for sponsoring.

Boot Support for Degraded RAID

kirkland has been working on adding RAID support to grub-install. His patch is waiting for sponsoring. He hopes it will make it for alpha4 scheduled to be released on Thursday. If so mathiaz suggested that this new feature should be documented in the Release Notes in order to get more testing.

The last item he's working on is adding a question during the installation on whether the system should be set to boot from a degraded array (defaulting to No, which matches the current behaviour).

ACTION: kirkland to document BootDegradedRaid in the release notes if it's included in alpha4.

Agree on next meeting date and time

nealmcb aggreed to run the meeting next week as mathiaz and Koon won't be available.

Next meeting will be on Tuesday, August 19th at 15:00 UTC in #ubuntu-meeting.

Log

[16:01] <mathiaz> #startmeeting
[16:01] <MootBot> Meeting started at 10:06. 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> Last week meeting minutes: https://wiki.ubuntu.com/MeetingLogs/Server/20080805
[16:02] <mathiaz> I don't see any actions marked there
[16:03] <mathiaz> Koon: is there anything special to report on ?
[16:03] <mathiaz> I think kirkland published most of his blog posts
[16:03] <Koon> mathiaz: no, nijaba still has to post his hw question, an action from the previous meeting
[16:04] <Koon> that's about it :)
[16:04] <mathiaz> Koon: ok - he is not around for this meeting
[16:04] <mathiaz> let's move on then
[16:04] <mathiaz> [TOPIC] Remove multiuser options and updating init scripts
[16:04] <MootBot> New Topic:  Remove multiuser options and updating init scripts
[16:04] <mathiaz> zul: ^ ?
[16:05] <zul> so yeah pitti had a list of packages that was still using multiuser for fast tear down that ssems to be depreciated in debian no one has started to fix them yet and I just wanted to bring them to peoples attention
[16:06] <soren> Where's the list?
[16:06] <zul> its a great way for people to get started in package development
[16:06] <Koon> zul: someone posted bugs about them, right ?
[16:06] <zul> http://lists.ubuntu.com/archives/ubuntu-devel-announce/2008-June/000430.html
[16:06] <MootBot> LINK received:  http://lists.ubuntu.com/archives/ubuntu-devel-announce/2008-June/000430.html
[16:06] <zul> Koon: yep james_w I believe
[16:06] <james_w> https://lists.ubuntu.com/archives/ubuntu-devel/2008-August/025945.html
[16:06] <james_w> https://bugs.edge.launchpad.net/ubuntu/+bugs?field.tag=multiuser is the list of bugs
[16:07] <zul> oh hello james_w :)
[16:07] <james_w> they're pretty easy to fix, and any help will be appreciated
[16:07] <james_w> hey zul
[16:08] <mathiaz> ok - are there any detailed instructions on how to do this ?
[16:09] <zul> in pitti's original email yes, but not for cdbs
[16:09] <mathiaz> zul: how many of these package use cdbs ?
[16:10] <zul> vsftpd and net-snmp as I recall
[16:10] <zul> I would do them myself but I have more pressing things
[16:10] <james_w> cdbs uses DEB_UPDATE_RCD_PARAMS (I think that's the right variable), which will be "mutliuser" currently, that needs the same substitution
[16:10] <mathiaz> zul: ok
[16:10] <james_w> I think there are a couple more, but most are plain debhelper
[16:10] <mathiaz> james_w: have you written a blog post about it ?
[16:10] <james_w> mathiaz: no, I haven't.
[16:11] <mathiaz> james_w: kirkland wrote a blog post about the status action in init script - the response has been very good so far
[16:11] <james_w> perhaps I will do that then
[16:11] <mathiaz> james_w: outlining detailed instructions on how to do that is helpfull
[16:12] <mathiaz> james_w: http://ubuntuserver.wordpress.com/2008/07/24/ubuntu-init-script-status-actions/
[16:13] <james_w> nice
[16:14] <mathiaz> james_w: the key point being to list the packages to fix and how-to do it
[16:15] <mathiaz> james_w: are you willing to write up a blog post about this ?
[16:15] <james_w> yeah, I think so
[16:15] <mathiaz> james_w: IIRC you're on the planet now
[16:15] <james_w> yeah, just last week
[16:16] <mathiaz> ACTION: james_w to write a blog post about removing multiuser options
[16:16] <james_w> I need to write up a form letter for forwarding to Debian as well, so that we make sure that we get that bit correct
[16:16] <mathiaz> [ACTION] james_w to write a blog post about removing multiuser options
[16:16] <MootBot> ACTION received:  james_w to write a blog post about removing multiuser options
[16:16] <mathiaz> james_w: great - thanks !
[16:16] <mathiaz> let's move on
[16:16] <nealmcb> james_w: thanks for your attention to good relations with debian!
[16:16] <mathiaz> [TOPIC] New eBox packages
[16:16] <MootBot> New Topic:  New eBox packages
[16:16] <foolano> hi
[16:16] <mathiaz> foolano: ^ ?
[16:17] <foolano> I've been packaging the latest eBox version for intrepid.
[16:17] <foolano> I've packaged a few new modules which could be interesting to have in intrepid:
[16:17] <foolano> ebox-mail (postfix + courier -dovecot support is not ready yet
[16:17] <foolano> ebox-mailfilter (amavis + clamav + spamassassain)
[16:17] <foolano> ebox-trafficshaping (to be used when the machine is working as a gateway), ebox-jabber (users in ldap can connect to the jabber server)
[16:17] <foolano> ebox-webserver
[16:18] <foolano> The latter is just a simple module to manage a couple of things of apache:
[16:18] <foolano> enable public_html directories for users stored in ldap and create virtual domains.
[16:18] <foolano> Right now, modules using ldap are broken beacuse of the new ldap config backend in intrepid.
[16:18] <foolano> I'll fix this today or tomorrow.I just have to change the way we add schemas and acls to ldap
[16:18] <foolano> the packages are available in my ppa: http://launchpad.net/~ebox-unstable/+archive
[16:19] <nealmcb> foolano: what jabber server does that use?
[16:19] <foolano> jabberd2
[16:19] <foolano> basically, that's what i wanted to say. the packages are in good shape and I wonderes if someone could upload them
[16:20] <foolano> s/wonderes/wondered/
[16:20] <mathiaz> foolano: they're all in your ppa ?
[16:20] <foolano> mathiaz: yep
[16:21] <mathiaz> foolano: what is in universe now ? the same version as in hardy ?
[16:21] <soren> Here's a thought... The minutes from the last TB meeting says that Launchpad supports giving per-package uploads rights to people. Perhaps eBox would be a good test for this as well. The eBox guys could be allowed to upload just that?
[16:21] <soren> it'd have to be approved by... Er... someone.
[16:21] <foolano> mathiaz: yep the same version as in hardy
[16:21] <mathiaz> soren: approved ? like sponsored ?
[16:22] <soren> No, like..
[16:22] <soren> Er..
[16:22] <mathiaz> soren: or just the acl set correctly ?
[16:22] <soren> Someone has to approve letting foolano upload ebox packages to launchpad.
[16:22] <soren> Like when the MC approves new MOTUs.
[16:22] <mathiaz> soren: right - mdz was looking into doing this
[16:22] <soren> Right. For something in main, right?
[16:22] <soren> i forget what it was.
[16:23] <mathiaz> soren: correct - something like that
[16:23] <soren> foolano: Would you be interested in this, or do you prefer going through a proxy+
=== asac_ is now known as asac
[16:23] <foolano> i think ebox-mail + ebox-mailfilter  can be useful to lower the barries for users who want deploy those services using a GUI...
[16:23] <mathiaz> foolano: would you be comfortable doing this (uploading directly to intrepid) ?
[16:23] <soren> ?
[16:24] <foolano> yep, i would be comofortable doing that
[16:24] <foolano> cuz otherwise is a pain in the ass for the one who uploads them
[16:24] <soren> Heh :)
[16:24] <foolano> right now there are 21 packages
[16:24] <foolano> :)
[16:24] <mathiaz> soren: well - I guess we'd have to ask the motu-council about this
[16:24]  * zul keeps quiet
[16:24] <nealmcb> where does ebox stand with augeas?
[16:24] <mathiaz> soren: or the technical-board
[16:24] <soren> mathiaz: i think that's the correct forum, yes.
[16:24] <soren> mathiaz: MC, that is.
[16:24] <nealmcb> oops - I bet still waiting on more augeas support...
[16:25] <mathiaz> soren: ok - could you take care of this ?
[16:25]  * foolano says hi to zul :P
[16:25] <soren> mathiaz: MC gives out MOTU membership. Since these are universe packages, it seems like they should do it.
[16:25] <soren> foolano: Could you send sort of an application for this to motu-council@lists.ubuntu.com? That would be very helpful.
[16:26] <soren> foolano: The MC has a meeting tomorrow afternoon, so if you could get it in before then, that would be perfect.
[16:26] <foolano> alright then
[16:26] <foolano> i guess i gotta explain this situation, right?
[16:26] <soren> Fantastic.
[16:26] <soren> Yes.
[16:27] <mathiaz> foolano: I'm glad to help you out on your application
[16:27] <owh> Just as an aside, does a package in universe have a different support length than in main?
[16:27] <soren> I'm on the MC, by the way, so I can fill in the gaps, if there's doubt.
[16:27] <soren> owh: Yes.
[16:27] <foolano> ok thanks, i'll probably nag you a little bit later while im writing it
[16:27] <soren> owh: 0 days.
[16:27] <mathiaz> foolano: if you could send me a draft of your application email, I'd be happy to review it and give you feedback
[16:27] <soren> owh: As opposed to 5 years.
[16:27] <owh> But isn't eBox the ubuntu-server recommended gui?
[16:28] <foolano> mathiaz: ok thanks
[16:28] <soren> owh: Yes.
[16:28] <owh> As-in, aren't we supposed to support it?
[16:28] <nealmcb> do we have any sense for how many people are using ebox at this point?
[16:29] <mathiaz> foolano: we'll see how the MC reacts to it
[16:29] <soren> owh: It's not guaranteed to be supported. If someone wants to support it, they're welcome and invited to.
[16:29] <mathiaz> foolano: it may need some time until you've got the right to upload
[16:29] <owh> soren: But doesn't that leave server administrators wide-open?
[16:29] <nealmcb> yeah - and giving upstream the ability to easily upload new packages will help with that
[16:29] <soren> owh: meaning?
[16:29] <mathiaz> foolano: let's wait a week to see how things are moving
[16:29] <foolano> mathiaz: the only thing is the FF is close...
[16:29] <mathiaz> foolano: exactly - FF is in three weeks now
[16:30] <owh> soren: Well, you deploy a server that is supported for 5 years but not all components are.
[16:30] <soren> owh: It's the "wide open" part, I don't understand.
[16:31] <mathiaz> foolano: well - a bit less than three weeks
[16:31]  * soren grumbles about English punctuation
[16:31] <zul> umm...can you guys continue that somewhere else?
[16:31] <owh> soren: Well, as an administrator, I follow the ubuntu-server recommendation to install eBox, but then it isn't supported like the rest of my ubuntu-server is. I'm left wide-open with a server that isn't working.
[16:31] <owh> zul: Well, it ties into the next agenda point.
[16:31] <soren> owh: how is it not working?
[16:31] <mathiaz> foolano: depending on the MC reaction, I'll have a look at sponsoring you're ebox packages
[16:32] <foolano> mathiaz: ok
[16:32] <owh> soren: Well, if eBox is not supported and a bug is found then I have a problem.
[16:32] <foolano> owh: upstream supports it
[16:32] <mathiaz> [ACTION] foolano to send an email to the MC to get upload rights to the ebox package
[16:32] <MootBot> ACTION received:  foolano to send an email to the MC to get upload rights to the ebox package
[16:32] <soren> owh: Well, yes. That doesn't mean that it's not working or wide open or whatever.
[16:32] <mathiaz> let's move on -
[16:33] <mathiaz> [TOPIC] ISV's - VMware on Gutsy and Hardy
[16:33] <MootBot> New Topic:  ISV's - VMware on Gutsy and Hardy
[16:33] <mathiaz> owh: ?
[16:33] <owh> That's me.
[16:33] <owh> This point is not intended as a bug report or as a way to whinge about vmware-server, it's about supporting a production server that relies on an ISV to provide what they committed to...
[16:33] <soren> owh: universe does not mean unsupportable. You are free to submit patches (like anyone else).
[16:33] <owh> vmware-server is in the gutsy partner repository. It's supposed to be supported for 12 months, only it isn't, that is, there are no kernel modules for the current gutsy kernel and hasn't been for some time...
[16:33] <owh> vmware-server was supposed to be released a few days after hardy was released, and as far as I know it still has not - though if someone has other information I'd love to hear that...
[16:33] <owh> As I see it, the current situation leaves the ubuntu-server team in a place where one hand is working independently from the other and the ubuntu-server product becomes compromised from a reliability and support perspective...
[16:33] <owh> So, my question is this: If we're working hard to roll out a reliable server platform, how is it possible that an ISV is not required to meet their obligations? Or have I just got the wrong end of the stick?
[16:34] <owh> This ties in with the eBox support question.
[16:34] <nealmcb> we have a virtualization solution in main - kvm.  I think at some point we do need a gui admin solution in main
[16:35] <nealmcb> the sooner the better :)
[16:35] <kirkland> sorry I'm late mathiaz, the intrepid kernel is panic'ing on me
[16:35] <mathiaz> owh: I think we've already discussed the vmware issue a couple of months ago
[16:35] <owh> nealmcb: Sure, if I could ditch vmware I would.
[16:36] <owh> mathiaz: Yes, on the 23rd of April.
[16:36] <nealmcb> what are the outages for you at this point of kvm?
[16:36] <owh> mathiaz: The problem is that no progress was made as far as I could see.
[16:36] <owh> nealmcb: Huh?
[16:36] <soren> nealmcb: virt-manager is in main..
[16:37] <soren> nealmcb: Oh..
[16:37] <soren> nealmcb: ignore me. I misunderstood.
[16:37] <owh> nealmcb: Ah, you mean, why can I not ditch it?
[16:37] <nealmcb> yeah
[16:37]  * soren guesses hardware support
[16:37] <owh> nealmcb: For one, I don't have dedicated CPU support, so I need to use qemu and it's verry slooow.
[16:38] <owh> Just changing hardware is not feasible in the middle of a products life.
[16:38] <nealmcb> any hope for getting qemu in main?  still need a fix for the gcc version isues?
[16:38] <nealmcb> how slow is kqemu for servers?
[16:38] <owh> From playing, not actual measurement, 50% slower.
[16:38] <owh> At least.
[16:39] <soren> Oh, it's much more than that.
[16:39] <owh> soren: Slower or faster?
[16:39] <soren> It would have to be at least... 85% slower.
[16:39] <soren> On a good day.
[16:39] <owh> Hmm, so I was being kind then :)
[16:39] <nealmcb> owh: it would seem that you need to work directly with vmware if thats what you need.  maybe get them to open source the whole thing :)
[16:39] <owh> The CPU load is ridiculous.
[16:40] <soren> nealmcb: kvm includes qemu, so there shouldn't be any need to move qemu to main.
[16:40] <owh> nealmcb: The issue is not that, the issue is that Canonical has signed VMware up as an ISV, but it seems that no actual progress is being made.
[16:40] <mathiaz> owh: I think that we came to a similar conclusion last time we talked about that issue - it's up to VMWARE to deal with this
[16:40] <soren> nealmcb: The reason for keeping qemu around is to keep all the non-x86 emulators it provides, but we have no interest in supporting those.
[16:41] <nealmcb> soren hmm I though qemu was in universe - can you run qemu on non-virt hardware using just main?
[16:41] <soren> Quite so. We're not trying to be difficult (I would like nothing more than for this to be fixed), but it's simply out of our hands.
[16:41]  * nealmcb remembers some gcc version issue with qemu
[16:41] <mathiaz> owh: I'd get in touch with nijaba  - he is the best point of contact to resolve this issue
[16:41] <soren> nealmcb: Just install kvm and use that.
[16:41] <owh> mathiaz: While I understand what you're saying, I don't see a world where vmware-server is used on a workstation.
[16:41] <soren> nealmcb: It'll fall back to "being qemu" if it doesn't find the hardware support it wants.
[16:42] <nealmcb> soren: ok - I get it now....
[16:42] <owh> mathiaz: As in, it's a ubuntu-server product.
[16:42] <soren> owh: I'm not sure I follow?
[16:42] <owh> We are the server platform for ubuntu. vmware-server is deployed on our platform.
[16:43] <owh> Canonical signed a contract to this effect with VMware.
[16:43] <owh> It's not being maintained.
[16:43] <owh> Our server product suffers.
[16:43] <owh> Apart from leaving administrators in the cold.
[16:44] <soren> Ok. You mentioned workstations, too?
[16:44] <owh> This is a fundamentally broken thing.
[16:44] <soren> I think that's what confused me.
[16:44] <dendrobates> owh: I am not aware of the language of the contract between vmware and canonical, but we are trying to work within their processes to get this fixed.
[16:44] <owh> soren: Yes, as in, vmware-server won't be used on a workstation.
[16:45]  * soren tries to make the connection
[16:45]  * soren fails
[16:45] <owh> dendrobates: Excellent.
[16:45] <owh> soren: I mean that the vmware-server product is a product not generally related to ubuntu, it's specific to us.
[16:45] <dendrobates> owh: they have difficulty with kernel abi changes and we are trying to reduce them to make it easier.
[16:45] <soren> owh: Ah, I see.
[16:46] <nealmcb> I think "us" is ubuntu.  vmware stuff is a vmware/canonical issue, not an ubuntu issue
[16:46]  * sommer agrees with nealmcb 
[16:46] <owh> nealmcb: Don't you think that that relates directly to us
[16:47] <nealmcb> I'd like to see it all resolved, but this meeting can't help you, other than to make libvirt/kvm or other open source solutions  better
[16:47] <owh> Sure, I was trying to find out if there was a way to have this process not break in the future, which is why I started asking about eBox too.
[16:48] <owh> If we as a team roll out and support a server and the things we rely on have a different support structure, we have a problem.
[16:48] <mathiaz> owh: ebox is bit different - it's opensource, so it can be fixed by anyone
[16:48] <nealmcb> ubuntu is free forever.  vmware isn't in that boat.  though perhaps you're asking if ubuntu wants to take over the open source part of vmware - is that what you're getting at, soren?
[16:48] <owh> nealmcb: That's only the toolbox AFAIK.
[16:48] <soren> nealmcb: Was that really a question for me?
[16:49] <owh> dendrobates: Thank you for your clarification.
[16:49] <nealmcb> soren: I don't understand your point about vmware-server and workstations
[16:49] <mathiaz> ok - let's move on then
[16:49] <soren> nealmcb: I had none.
[16:49] <soren> nealmcb: I think you got me mixed up with someone else..
[16:49] <mathiaz> The current Roadmp: https://wiki.ubuntu.com/ServerTeam/Roadmap
[16:50] <mathiaz> [TOPIC] Ubuntu VM builder
[16:50] <MootBot> New Topic:  Ubuntu VM builder
[16:50] <soren> \o/
[16:50] <mathiaz> soren: anything ready for wider testing ?
[16:50] <nealmcb> soren: oops - right....
[16:50] <soren> I've finally gotten it to build a Xen based image. It's very rough around the edges, but I'm working hard on getting it shaped up.
[16:50] <owh> And yay to mvo for sandbox testing!
[16:51]  * mvo blushes
[16:51] <soren> I'll put out an e-mail when it's ready for actual use. Developers are very welcome to pitch in.
[16:51] <mathiaz> soren: is there a minimal set of documentation ?
[16:51] <owh> soren: Don't be shy, give out the goss!
[16:51] <mathiaz> soren: or some basic instructions somewhere ?
[16:52] <zul> soren: link again?
[16:52] <soren> mathiaz: It's "--help" option is quite verbose, actually.
[16:52] <owh> Should we be using the mailing list for in-progress stuff like what soren is doing, or is that really intended more for support?
[16:52] <soren> I'd love to see more developmenty sort of stuff on the list.
[16:53] <owh> Ditto
[16:53] <soren> I should set an example to that effect.
[16:53] <mathiaz> as we're approaching FF we'll have more things happening
[16:53] <mathiaz> anyone is free to post call for testing on ubuntu-server@
[16:54] <mathiaz> writing a blog post of the ubuntuserver blog is also possible
[16:54] <mathiaz> kirkland wrote such a post and I published it on the blog
[16:55] <mathiaz> I can create an contributor account for anyone interested in publishing a post on the ubuntuserver blog
[16:55] <mathiaz> [TOPIC] Migrate openldap configuration to cn=config
[16:55] <MootBot> New Topic:  Migrate openldap configuration to cn=config
[16:56] <mathiaz> I uploaded openldap 2.4.11 yesterday with the default config backend using cn=config instead of slapd.conf
[16:56] <mathiaz> migration from slapd.conf to cn=config is automatically done
[16:57] <mathiaz> although slapd.conf support is still enabled in slapd, it's no longer supported by the packaging scripts.
[16:57] <kirkland> I have been *very* impressed by the community response to these posts
[16:58] <mathiaz> bottom line: if you're not migrating to cn=config and stay with slapd.conf, you're on your own
[16:58] <kirkland> fwiw, we have gotten some good init script patches, and I received some well written bugs against ecryptfs-utils
[16:59] <mathiaz> [TOPIC] Tomcat6 server stack support
[16:59] <MootBot> New Topic:  Tomcat6 server stack support
[16:59] <mathiaz> Koon: ^ ?
[16:59] <Koon> So I packaged a Tomcat6 server stack, it was uploaded yesterday
[16:59] <Koon> (doesn't show up in the archives yet, not sure if that's normal)
[17:00] <Koon> The goal is to have it as a tasksel in the end, but the road is still long
[17:00] <mathiaz> Koon: it's in the NEW queue - waiting for an archive admin to push it in the archive
[17:00] <Koon> ok
[17:00] <Koon> In parallel I'm working to reduce openjdk-6-jre-headless dependencies
[17:00] <mathiaz> Koon: https://launchpad.net/ubuntu/intrepid/+source/tomcat6/6.0.16-1ubuntu1
[17:01] <Koon> so that it doesn't pull half of a desktop install in your servers just to run tomcat6 ;)
[17:01] <mathiaz> Koon: are you tracking what is needed to be done ?
[17:02] <mathiaz> Koon: have you filed bugs about it ?
[17:02] <Koon> yes, all but one
[17:02] <mathiaz> Koon: IIUC, it's mainly dependency issues - it shouldn't be to hard to prepare debdiff
[17:02] <Koon> mathiaz: the last one is a Recommend, but I want to have doko opinion on why he added it in the first place
[17:03] <mathiaz> Koon: ok - do you have a list of the filed bugs ? Are debdiffs already done for them ?
[17:03] <mathiaz> @schedule
[17:03] <ubottu> Schedule for Etc/UTC: 14 Aug 12:00 UTC: Ubuntu Mobile Team | 14 Aug 14:00 UTC: Ubuntu Java Team | 21 Aug 12:00 UTC: Ubuntu Mobile Team | 21 Aug 14:00 UTC: Ubuntu Java Team | 22 Aug 12:00 UTC: Ubuntu MOTU | 23 Aug 13:00 UTC: Xubuntu Community
[17:03] <Koon> yes
[17:03] <mathiaz> Koon: so you're just waiting for sponsoring ?
[17:04] <Koon> yes.
[17:04] <Koon> bug 249178
[17:04] <ubottu> Launchpad bug 249178 in ecj "libecj-java shouldn't recommend java2-runtime" [Low,Triaged] https://launchpad.net/bugs/249178
[17:04] <Koon> bug 256096 (multiple packages)
[17:04] <ubottu> Launchpad bug 256096 in libcommons-dbcp-java "Common Java libraries should depend on a headless runtime" [Low,Triaged] https://launchpad.net/bugs/256096
[17:05] <soren> I'd love to see more developmenty sort of stuff on the list.
[17:05] <soren> Whoops...
[17:05] <mathiaz> Koon: great - anything else to add on tomcat6 ?
[17:05] <owh> DeJaVue all over again there soren :)
[17:05] <Koon> no, please test it :)
[17:06] <soren> Yeah, that was some pretty lucky cutting and pasting.
[17:06] <mathiaz> Koon: great - thanks for your work
[17:06] <Koon> when I come back from vacation I want a full page of bug reports
[17:06] <mathiaz> [TOPIC] Boot Support for Degraded RAID
[17:06] <MootBot> New Topic:  Boot Support for Degraded RAID
[17:06] <mathiaz> kirkland: ^ ?
[17:07] <kirkland> mathiaz: yessir
[17:07] <kirkland> mathiaz: i'm waiting slangasek to sponsor my latest round of patches
[17:07] <kirkland> mathiaz: these affect grub-install
[17:08] <kirkland> mathiaz: i've gotten part of them in to grub-installer (used in the installer)
[17:08] <kirkland> mathiaz: these patches write GRUB to the MBR of each disk in the RAID
[17:08] <kirkland> mathiaz: slightly complicated by the automatic detection and calculation of the disk/partition offsets
[17:09] <mathiaz> kirkland: great - is there anything to test for now ?
[17:09] <kirkland> mathiaz: i'm hoping that it makes it into Alpha 4
[17:09] <mathiaz> kirkland: or this is still very much a WIP now ?
[17:09] <kirkland> mathiaz: no, it's really wrapping up
[17:09] <kirkland> mathiaz: the last thing I really want to complete is to add a debconf question to the installer
[17:10] <kirkland> mathiaz: asking the user if they want to default to booting or not booting if in degraded mode
[17:10] <mathiaz> kirkland: ok - so if it makes it into alpha4, we should defintely document it on the RelaseNotes
[17:10] <mathiaz> kirkland: so that people can test it
[17:10] <owh> kirkland: What default are you proposing for that question?
[17:11] <kirkland> mathiaz: agreed.  i added ecryptfs-utils to the release notes too
[17:11] <kirkland> owh: the same behavior we have always had... "no"
[17:11] <owh> kirkland: You mean Not booting?
[17:12] <sommer> also, if anyone wants to do some reviewing: http://doc.ubuntu.com/~mdke/test/serverguide/C/advanced-installation.html
[17:12] <owh> kirkland: Uh do you mean, "Not booting if in degraded mode?"
[17:12] <kirkland> owh: it will drop you to a busy box prompt if you are missing a disk
[17:12] <mathiaz> sommer: isn't doc.ubuntu.com up-to-date anymore ?
[17:13] <owh> kirkland: Hmm, I'm sure that people with remote server farms will disagree with that default.
[17:13] <kirkland> owh: the default is that the initramfs will detect when a disk has gone missing and drop you to a busy box prompt to take the actions you deem necessary
[17:13] <kirkland> owh: you can override that either a) in a config file in /etc, or b) on the kernel boot line
[17:13] <sommer> mathiaz: no, there are some changes to the layout and there's some issue with the versions of xml... or something, I'm not sure I read the email that close
[17:13] <owh> kirkland: But doesn't that defeat the purpose of having RAID at all?
[17:13] <mathiaz> sommer: right - I remember now
[17:14] <nealmcb> what priority would the debconf question be?
[17:14] <kirkland> i suspect owh will be one of many people (myself included) to switch "BOOT_DEGRADED=yes"
[17:14] <owh> Which is why I asked :)
[17:15] <kirkland> but this has been the expected default for a very long time, i have been strongly strongly advised not to change that
[17:15] <owh> kirkland: Aren't we supposed to provide best practice?
[17:15] <nealmcb> perhaps the text for the question could advise server folks to change to yes
[17:16] <owh> I'm not in a position to disagree with "strongly strongly advice", but I was just wondering.
[17:16] <kirkland> the text of the question will be very well crafted, and I'll draft sommer to review the text ;-)
[17:16] <owh> :)
[17:16] <kirkland> it will explain both the pros and cons
[17:17] <kirkland> the cons are very dangerous, booting a server in an unprotected mode
[17:17] <nealmcb> owh: on a desktop, the best answer is probably "no" - since folks will be there and fix it right away and others don't care
[17:17] <kirkland> we don't want to surprise users with that
[17:17] <mathiaz> kirkland: great - thanks for the update - we'll see if it can make it to alpha4
[17:17] <kirkland> nealmcb: right, i agree with you
[17:17] <owh> nealmcb: Sure.
[17:17] <owh> So, is the default different for the server?
[17:17] <owh> :)
[17:17] <mathiaz> kirkland: be sure to document it in the Release notes if it's included in alpha4
[17:18] <nealmcb> but having things default differently for server vs desktop would be more confusing than it would be worth
[17:18] <kirkland> mathiaz: i will....  just waiting on slangasek to sponsor
[17:18] <mathiaz> kirkland: great - thanks
[17:18] <owh> nealmcb: Not sure I agree with that.
[17:18] <mathiaz> [ACTION] kirkland to document BootDegradedMode in the release notes if it's included in alpha4
[17:18] <MootBot> ACTION received:  kirkland to document BootDegradedMode in the release notes if it's included in alpha4
[17:19] <mathiaz> [TOPIC] # Open Discussion.
[17:19] <MootBot> New Topic:  # Open Discussion.
[17:19] <mathiaz> something to add ?
[17:19] <owh> There is a discussion about fsck underway on u-d-d.
[17:20] <owh> Should we as u-s wade in?
[17:20] <mathiaz> owh: if you an opinion on the matter you're more than welcome to express it in the thread
[17:20] <mathiaz> owh: if you *have* and opinion
[17:21] <owh> I was more asking from a perspective as the team.
[17:21] <owh> I mean, aren't we the collective of server people here?
[17:22] <owh> At what point do we act as a team to help further the server platform?
[17:22] <mathiaz> owh: true - but we're also part of the developer community - we shouldn't discuss these matters in our own corner and then express it in the thread
=== johnc4511 is now known as johnc4510
[17:22] <mathiaz> owh: if you have something to add to the discussion just do it
[17:22] <owh> You don't think that there is value in discussing it here?
[17:23] <sommer> owh: I think the discussion is going to be the same as the main thread
[17:23] <mathiaz> owh: if you have an opinion about how fsck should be handled in the server use case, express it in the thread - that's the best place to do so
[17:23] <owh> sommer: That's a fair point.
[17:24] <owh> That's my only Open Discussion item :)
[17:26] <mathiaz> [TOPIC] Agree on next meeting date and time
[17:26] <MootBot> New Topic:  Agree on next meeting date and time
[17:26] <mathiaz> I won't be available next week to run the meeting, neither Koon
[17:26] <mathiaz> anyone interested to step up ?
[17:27]  * owh would volunteer, but the likelyhood is that I'll be fast asleep.
[17:27] <owh> And if that's an incentive for someone else to step up, all the better :)
[17:27] <mathiaz> kirkland: are you up to the task ?
[17:27] <kirkland> mathiaz: I will be in London next week...  perhaps another time, when I know my surroundings?
[17:28] <nealmcb> If you need someone, I can do it.
[17:28] <mathiaz> kirkland: ok - nealmcb just voluntered :)
[17:28]  * Koon cheers
[17:28] <mathiaz> nealmcb: thanks for your proposal
[17:28] <kirkland> mathiaz: thanks, keep me on the short list though, i'm happy to help
[17:28] <mathiaz> so same time, same place, next week ?
[17:29] <owh> Can we make it two hours earlier?
[17:29] <owh> Or is that too much to ask?
[17:29] <owh> owh == 00:29:46
[17:30] <owh> It's cool if not.
[17:30] <mathiaz> owh: that would be a bit to early for the people on the west coast
[17:31] <mathiaz> owh: unfortunately, we're all in different timezone
[17:31] <owh> mathiaz: West Coast of where?
[17:31] <mathiaz> owh: north america
[17:31] <owh> Isn't that like noon?
[17:31] <sommer> owh: 6:00am
[17:31] <mathiaz> owh: nop - it's 9:31 AM on the west coast
[17:31] <zul> that would be east coast
[17:31] <owh> Ah.
[17:32] <owh> Fair enough.
[17:32] <sommer> is there anyone on the west coast, attending?
[17:32] <mathiaz> all right then - same place, same time, next week
[17:32] <mathiaz> your host will be nealmcb !
[17:32] <mathiaz> #endmeeting

MeetingLogs/Server/20080812 (last edited 2008-08-13 00:06:05 by mathiaz)