Items we will be discussing:

  • Review ACTION points from previous meeting.
  • Review high priority bugs related to the Server Team.
  • It has been suggested that a more concise Server Team Report would help the readability of the monthly TeamReports. I would like to discuss paraphrasing, or only using the highlights for the Server Team portion of the report. -- Asommer.

  • Review ServerTeam/Roadmap and ServerTeam/ReportingPage.

  • Agree on next meeting date and time.


mathiaz asked if someone ran into a bug that should be considered as a release blocker.

There was some discussion about the different kvm/libvirt/virt-manager bugs but the most critical ones have been fixed.

kirkland and jdstrand uploaded a fix for the ldap hang at boot bug.

nealmcb reported a failed install of postgresql - mathiaz suggested it was related to ubuntu-vm-builder instead of postgresql.

ServerTeam report and the monthly report

sommer heard complains that the Server team section in the monthly report was too big. mathiaz said he would further edit the section to make it less verbose. He also asked if the format of the weekly minutes were good: everybody agreed that the minutes were useful.

There was some discussion about the content of the monthly report, whether it should list ideas or projects that are worked on by the server team. mathiaz stated that the monthly report is about things that have been completed. It may not be the best place to talk about ongoing activities. He added that there should be a session about this topic at UDS and he would seek the input of the Community team.

ACTION: mathiaz to make the server section of the monthly report less verbose.

VMware release times in the partner repository

owh raised the issue of having vmware-server available from the partner repository the day hardy is released. It took a couple of weeks when gutsy was released. nijaba said he hoped that vmware-server would be available faster this time. However this matter is in the hands of VMWare and the partner repository team of Canonical. There isn't much we can do about it.

Ubuntu OpenWeek

mathiaz reminded that the Ubuntu Open Week [1] will be next week. There will be a session about the Server Team (given by mathiaz) and another one about virtualization (lead by soren). There are lots of other topics covered and the schedule is posted on the wiki page [1].

[1]: https://wiki.ubuntu.com/UbuntuOpenWeek/

Other Business

nxvl wondered if it's a good idea to start merging now. mathiaz said to wait until the intrepid repositories are opened, which should happen sometime next week. We'll go into Merging mode then. He also reminded that UDS is comming up and people should start to think about topics they wanna discuss.

Agree on next meeting date and time

Next meeting will be on Wednesday, April 30th at 21:00 UTC in #ubuntu-meeting.


Started logging meeting in #ubuntu-meeting
[23:00:45] <mathiaz> Welcome to the last meeting of the Ubuntu Server Team
[23:00:56] <jdstrand> \o
[23:00:58] * faulkes- raises an eyebrow
[23:01:01] <mathiaz> for the Hardy Release cycle
[23:01:07] <sommer> heheh
[23:01:08] <jdstrand> oh
[23:01:10] <jdstrand> *phew*
[23:01:18] <nxvl> i'm going for a quick smoke and brb
[23:01:27] <soren> dendrobates sends his apologies.
[23:01:34] <soren> He's on a road trip.
[23:01:57] <mathiaz> Unless you've dropped off the Ubuntu planet, you know that we're releasing hardy tomorrow
[23:02:00] * kirkland recommends UpIRC PalmOS application for IRC connectivity in the car ;-)
[23:02:36] <nijaba> kirkland: this is an add free channel ;)
[23:02:37] <mathiaz> so testing testing and testing are the magic words these days
[23:02:43] * owh recommends mIRGGI for an N95 :)
[23:03:05] <mathiaz> Today's agenda: https://wiki.ubuntu.com/ServerTeam/Meeting
[23:03:44] <mathiaz> Last meeting note (https://wiki.ubuntu.com/MeetingLogs/Server/20080416) lists no ACTION
[23:04:24] <mathiaz> [TOPIC] Review release-blocker/criticial bugs related to the Server Team.
[23:04:51] <mathiaz> Anyone knows about such a bug ?
[23:05:16] <zul> I hope not
[23:05:25] <nxvl> nop
[23:05:29] * nijaba shivers
[23:05:30] <nxvl> i haven't find any
[23:06:14] <mathiaz> we've been chasing some kvm bugs
[23:06:27] <mathiaz> but it was solved for hardy
[23:06:46] <soren> Yeah. They've been popping up like crazy for the last 5-6 days.
[23:07:08] <nijaba> soren: that's because we all did the iso testing using it, I guess
[23:07:16] <jdstrand> I don't know that kvm has ever seen such a workout
[23:07:17] <mathiaz> soren: for my understanding, you've disable virtio-blk in hardy ?
[23:07:24] <soren> nijaba: Oh, yeah. Sure.
[23:07:26] <soren> mathiaz: That's right.
[23:07:31] <soren> mathiaz: Well...
[23:07:55] <soren> mathiaz: The kernel supports it, and kvm has the code, too, but virtinst (and hence virt-manager) doesn't enable it for hardy guests.
[23:08:06] <soren> We still have virtio-net, which by far was the most important one.
[23:08:09] <kirkland> mathiaz: worth noting that jdstrand uploaded a fix for the nss/ldap login hang bug that's been around for a long while
[23:08:48] <mathiaz> soren: so if I use virsh and an xml file to define my guest, I still have access to virtio-blk ?
[23:08:54] <nijaba> a brillant fix by kirkland!
[23:09:10] <jdstrand> soren: you did remove a troublesome patch for virtio-blk that makes it really not very good for normal uses though
[23:09:14] <jdstrand> soren: correct?
[23:09:28] <soren> mathiaz: Right.
[23:09:38] <soren> jdstrand: Right.
[23:09:52] <soren> For rather extreme valus of "not very good" even.
[23:09:54] <soren> :)
[23:10:01] <jdstrand> heh
[23:10:04] <nealmcb> I had an odd postgres install problem on an ubuntu-vm-builder vm last night - I should try it again.... https://bugs.edge.launchpad.net/ubuntu/+source/postgresql-common/+bug/193696
[23:10:07] <ubotu> Launchpad bug 193696 in postgresql-common "Postgresql 8.3 not responding" [Low,Fix released]
[23:10:26] <mathiaz> nealmcb: I've commented on that bug - it's probably related to ubuntu-vm-builder
[23:10:36] <nxvl> soren: did you see the shotdown botton bug i reported?
[23:10:45] <nxvl> shutdown*
[23:10:45] <nealmcb> mathiaz: I saw that, but the postgres was installed via apt-get, not in the original build
[23:10:47] <mathiaz> nealmcb: postgresql works correctly in a guest
[23:11:23] <mathiaz> nxvl: bug number ?
[23:11:26] <soren> nxvl: I've not seen it, but I'm almost sure I know what it is :)
[23:11:34] * nxvl searches
[23:11:34] <soren> nxvl: ...and it's not a bug :)
[23:11:50] * nijaba bets for acpi
[23:11:58] <jdstrand> the libnss-ldap bug was tough to triage, but once it was triaged things fell into place. kirkland did most of the fix and it is working great. thanks kirkland!
[23:11:58] <kirkland> nxvl: soren: I reported one too, a kvm segfault, acpi related?
[23:12:08] <nxvl> Bug #190886
[23:12:12] <ubotu> Launchpad bug 190886 in virt-manager "virt-manager: pause/break key does not work" [Undecided,Confirmed] https://launchpad.net/bugs/190886
[23:12:16] <soren> kirkland: I haven't seen it. I'm not really caught up on launchpad mail.
[23:12:42] <soren> nxvl: They're different issues.
[23:12:44] <jdstrand> still, all these kvm issues aren't release critical
[23:12:54] <soren> nxvl: Your problem is that you have nothing that's responding to the acpi events the shutdown button sends.
[23:13:00] <soren> nxvl: solution: Install acpid in the guest.
[23:13:10] * nxvl tries
[23:13:15] <kirkland> soren: yeah, i haven't noticed any ill effects, besides the segfault message on /var/log/messages
[23:13:16] * jdstrand likes to 'Destroy' things
[23:13:41] <jdstrand> (in virt-manager that is)
[23:14:23] <mathiaz> so it seems like we're on track for hardy release
[23:14:27] <mathiaz> let's move on
[23:14:44] <mathiaz> [TOPIC] ServerTeam report and the monthly report
[23:15:10] <mathiaz> sommer: you've been told we should have shorter reports ?
[23:15:26] <sommer> mathiaz: ya, that was the gist of things
[23:15:40] <mathiaz> sommer: right - I agree with that
[23:16:02] <nijaba> does this mean we work too hard?
[23:16:05] <mathiaz> I'm still working out the whole status reporting thing
[23:16:20] <sommer> nijaba: heh, I think it means we report too hard
[23:16:23] <owh> nijaba: No, it means we talk too much.
[23:16:26] <kirkland> nijaba: nah, just talk too much :-)
[23:16:26] <mathiaz> nijaba: it means we should edit the monthly report a little bit more
[23:17:17] <mathiaz> for now, I just take the meeting minutes and shove them in the monthly report with some light editing
[23:17:33] <owh> Is there a specified format that we need to adhere to?
[23:17:45] <mathiaz> So I'd like to know if the format of the Weekly meeting minutes is good ?
[23:18:05] <mathiaz> Are the minutes too short, too long or verbose ?
[23:18:22] <nxvl> mathiaz: i found them good
[23:18:25] <faulkes-> I find them concise and to the point
[23:18:28] <sommer> mathiaz: I think they are too long and verbose
[23:18:28] <nxvl> mathiaz: they have all the information needed
[23:18:52] <owh> The current format shows what people would need if they were not at the meeting. Isn't that the whole point?
[23:18:57] <sommer> mathiaz: it was mentioned that we don't have to report on the "ideas" that people are working on
[23:18:58] <mathiaz> sommer: they == the minutes sent each week or the content of the monthly report ?
[23:19:06] <nxvl> so anyone can understand what has been talked if they weren't present
[23:19:13] <sommer> mathiaz: sorry, the monthly report... I think the minutes are fine
[23:19:33] <mathiaz> sommer: ok - I aggree with you that the monthly report is about what has been done
[23:19:50] <owh> But reporting on ideas is where progress comes from - it sparks something with someone and magic happens.
[23:19:56] <mathiaz> sommer: I've tried to edit out the section about what people are working on
[23:20:03] <mathiaz> sommer: may it's just too verbose
[23:20:22] <mathiaz> owh: right - but I think the monthly report is about the past
[23:20:25] <nealmcb> mathiaz: I think the minutes are great. the reports could just be a shorter list with references to the minutes
[23:20:35] <sommer> mathiaz: I think that's what they were getting at
[23:20:49] <owh> mathiaz: But the monthly report is visible to more people, so the ideas spread wider.
[23:21:22] <mathiaz> ok - so I'll try to make the monthly report less verbose on focus on what has been achieved
[23:21:31] <sommer> we could always link the team report to the more detailed server report
[23:21:44] <mathiaz> owh: agreed - I'm not sure that the monthly report is the correct way to do it
[23:21:45] <owh> That sounds like a good idea.
[23:22:09] <mathiaz> owh: I'd like to have a way to advertise what we're working on now (ie - the present and the futur)
[23:22:19] <mathiaz> owh: but the monthly report seems to be about the past
[23:22:44] <owh> mathiaz: But if that is the only such forum, then there is no real alternative location.
[23:22:46] <mathiaz> sommer: that means we'd have three documents to maintain
[23:22:48] <nealmcb> mathiaz: I also would like to see ideas for the future out there, and don't want another report series....
[23:23:00] <nealmcb> so put ideas for the future in the reports, but just in more summarized form?
[23:23:04] <mathiaz> sommer: the weekly minutes, the monthly server report and the server section of the monthly report
[23:23:23] <sommer> mathiaz: heh, and there's also the blog idea :)
[23:23:23] <faulkes-> if I have a list of current in-progress projects, I am happy to update the forum sticky on a regular basis
[23:23:38] <owh> Where there *specific* objections to the monthly report?
[23:23:42] <mathiaz> sommer: right - that's why I've tossed the idea of a blog
[23:23:50] <mathiaz> owh: too verbose I think
[23:23:58] <sommer> owh: ^^^ yes
[23:24:13] <nealmcb> I recall noticing that there are also some differences e.g. in formatting - lots of subheadings in our report and that is rare in others
[23:24:43] <nealmcb> so the table of contents looks like it is all server, all the time :-)
[23:24:59] <owh> Excellent :)
[23:25:00] <mathiaz> nealmcb: true - I'll try to fix that
[23:25:08] <sommer> nealmcb: it's pretty much a pay per view channel
[23:25:21] <owh> nealmcb: That's how it should be right :)
[23:25:35] <nealmcb> cool - we can even monetize it!!
[23:25:45] <mathiaz> [ACTION] mathiaz to make the server section of the monthly report less verbose
[23:25:58] <owh> I've just looked at https://wiki.ubuntu.com/ServerTeam/ReportingPage/2008-04, that's a single screen. Is that too verbose?
[23:26:05] <mathiaz> I'd add that I've asked for a session at UDS about this
[23:26:28] <mathiaz> I've got a couple of other ideas on improving the server team community and how to attract new contributors
[23:26:45] <mathiaz> we should get a session at UDS about this to discuss all of that
[23:27:40] <mathiaz> If you have ideas, add them to the IdeaPool page
[23:27:40] <nealmcb> mathiaz: a session on reporting for all teams, or just server?
[23:27:42] <mathiaz> https://wiki.ubuntu.com/ServerTeam/IdeaPool
[23:28:17] <mathiaz> nealmcb: just server - and it wouldn't be just about reporting
[23:28:31] <mathiaz> nealmcb: I'd also get the input from the community team
[23:28:41] <mathiaz> UDS will be a good place to discuss that
[23:29:20] <mathiaz> [TOPIC] VMware release times in the partner repository.
[23:29:34] <mathiaz> owh: ??
[23:29:39] <owh> Yup
[23:29:55] <owh> I deploy servers with clients and rely on vmware-server.
[23:30:07] <owh> The challenge is that the release dates are not synchronised.
[23:30:22] <owh> This means that I cannot deploy a current release if it's recent.
[23:30:35] <owh> Are there things we can do as a server team to mitigate this?
[23:30:42] <soren> Not really.
[23:30:57] <owh> Other than migrate to a different solution :)
[23:31:01] <soren> It's in the hands of the partner repository people and VMWare.
[23:31:22] <owh> There seems to be a disconnect there.
[23:31:31] <owh> That is, release === release.
[23:31:33] <faulkes-> ok, is it possible to speak to them about it or is this a one off issue?
[23:32:04] <owh> faulkes-: Well it has happened two releases running and it seems it will happen again with Hardy.
[23:32:39] <owh> I understand that we don't control our partners.
[23:32:54] <owh> But if they purchase the right to be one, that comes with responsibilities.
[23:32:59] <faulkes-> owh: will you send me a description of the issue/problem and I will see who we can approach @ vmware to discuss it?
[23:33:00] <owh> IMHO
[23:33:25] * faulkes- doesn't know but they do have offices just down the road from here
[23:33:38] <nealmcb> how long was the delay in previous releases?
[23:33:38] <nijaba> point is that vmware does not want to build the tools until we release. that will not change until they complete the floss version of their tool
[23:33:38] <owh> faulkes-: There was an email to U-D-D IRC about it for hardy where a developer was discussing it with them,
[23:33:47] <nijaba> we talk with them bi-weekly
[23:34:04] <owh> nijaba: But isn't that why we freeze stuff?
[23:34:14] <zul> faulkes-: burlington?
[23:34:25] <owh> nealmcb: I recall one delay of at least a month or more.
[23:34:30] <nijaba> owh: they don't care, that's there rule...
[23:34:44] <nijaba> their, even
[23:34:49] <owh> nealmcb: I think it may have been at least 2 months, but my memory is hazy.
[23:34:59] <faulkes-> zul: yes
[23:35:08] <owh> nijaba: Well, that shouldn't be how it works. If you are a partner then you have a commitment.
[23:35:09] <nijaba> owh: it should not be that long this time (I hope)
[23:35:16] <faulkes-> zul: well, oakville, just near the border, they are about 5 min away
[23:35:47] <owh> So, what is the process to raise this kind of issue?
[23:36:16] <owh> Other than faulkes- knocking on their door.
[23:36:55] <ivoks> sounds like politics
[23:37:11] <owh> Am I under the misunderstanding that deploying a server with a partner repository is not meant for mission critical deployment?
[23:37:30] <nijaba> owh: what do you mean exactly?
[23:37:41] <faulkes-> owh: I think you have made your point
[23:38:00] <faulkes-> nijaba: that if they commit to a partnership there is an inherent obligation to meet certain timeframes/responsibilities
[23:38:08] <nijaba> owh: show me one os that has certified partner tools the day of release and I'll start to understand
[23:38:09] <owh> faulkes-: Yes.
[23:38:25] <owh> nijaba: That's not my point.
[23:38:27] <faulkes-> nijaba: point
[23:38:46] <owh> nijaba: It should be that way. If we cannot work towards that, then why are we partnering at all?
[23:38:47] <faulkes-> however, the question remains, if we want to solve the issue, is there a way we can easily do so?
[23:39:08] * owh hunts for the email thread.
[23:39:27] <faulkes-> I'm happy to show up at there door with a case of beer to discuss it but thats a bit extreme
[23:39:32] <nijaba> faulkes-: the only real solution is for vmware to complete their oss implementation of their tool so that WE can compile them when needed
[23:39:51] <nijaba> faulkes-: and that's the plan we've been following for a while now
[23:40:03] <faulkes-> ok
[23:40:05] * faulkes- nods
[23:40:11] <nijaba> faulkes-: unfortunately they did not complte it in time for hardy as we would have hoped.
[23:40:24] <owh> Hold on, I'm not talking about vmware-tools.
[23:40:25] <jdstrand> I am not speaking authoritatively here, but I always just figured they were going to do all their regression testing, etc when they knew things weren't going to change-- which means release
[23:40:34] <owh> I'm talking about vmware-server
[23:41:05] <nijaba> owh: vmware server is another story, as they do not even support the thing
[23:41:58] <owh> nijaba: So, if it's not supported, why is it in the partner repository at all?
[23:42:20] <nealmcb> I hear there's this kvm/libvirt/ubuntu-vm-builder suite that is sweet and open-source and could use some polish :-)
[23:42:56] <nijaba> owh: partner repo does not equates to vendor supported, but to vendor provided, I gues
[23:42:57] <owh> nealmcb: As soon as I can run that on hardware that runs vmware, I'm in.
[23:43:27] <mathiaz> it seems that owh highlighted a misconception about the partner repository
[23:43:36] <owh> nealmcb: And as you well know, I've been putting my oar in there :)
[23:43:50] <nealmcb> I see value with folks providing stuff on ubuntu and figure it is up to them how and when they do so
[23:43:54] <mathiaz> This may need to be clarified by Canonical
[23:43:55] <owh> mathiaz: At least a non-shared understanding.
[23:44:02] <nealmcb> owh: right on - just didn't want that to get lost in the conversation
[23:44:17] <mathiaz> Let's move on
[23:44:23] <owh> nealmcb: Ah, the soren-pay-tv-channel :)
[23:44:30] * faulkes- has to head out
[23:44:44] <owh> Right, so what is the action for this?
[23:44:47] <mathiaz> [TOPIC] Other Business
[23:44:58] <owh> Who can talk to someone about the issue?
[23:44:59] <mathiaz> owh: there isn't much we can do unfortunately
[23:45:12] <owh> mathiaz: Well we identified several issues.
[23:45:15] <mathiaz> owh: it's up to vmware to figure it out
[23:45:42] <owh> mathiaz: But it's up to us to point issues out to the partner team and Canonical.
[23:45:44] <mathiaz> owh: I think that nijaba heard your concern and knows about it - it may take more time that expected to come to an proper solution
[23:46:01] <soren> owh: "soren-pay-tv-channel"?
[23:46:06] <nijaba> mathiaz: I did, indeed...
[23:46:25] <owh> mathiaz: nijaba: Thanks, all good.
[23:46:27] <mathiaz> owh: right - nijaba knows about it now - I think we cannot do much more
[23:46:51] <owh> soren: Just a weird reference to all ubuntu-server all the time, pay-tv and you being mr virtual :)
[23:47:18] <soren> owh: ok :)
[23:48:06] <mathiaz> As you may know, there will be an Ubuntu OpenWeek next week
[23:48:13] <mathiaz> https://wiki.ubuntu.com/UbuntuOpenWeek/
[23:48:42] <mathiaz> I'll give an intro about the Server Team on Tuesday
[23:49:08] <mathiaz> and soren will lead a session on virtualization
[23:49:22] <mathiaz> there are a lot more sessions scheduled
[23:49:28] <mathiaz> check the wiki page: https://wiki.ubuntu.com/UbuntuOpenWeek/
[23:50:16] <mathiaz> [TOPIC] #
[23:50:16] <mathiaz> Agree on next meeting date and time.
[23:50:23] <mathiaz> [TOPIC] Agree on next meeting date and time.
[23:50:38] <mathiaz> Same place, same time, next week ?
[23:50:38] <owh> Can I observe that the times associated with those sessions are heavily weighted to being while I'm asleep?
[23:50:51] <nxvl> mathiaz: next week do we need to have a weekind, or it will be "vacations" until ibex repositories are up?
[23:51:08] <mathiaz> owh: true - but there will be irc logs
[23:51:39] <mathiaz> nxvl: repositories for intrepid may open next week
[23:51:39] <owh> I'm just thinking, we've sort of gone global a while ago :)
[23:51:50] <nxvl> woohooo, merge time
[23:51:55] <mathiaz> nxvl: there is also UDS to prepare
[23:52:08] * nijaba won't be around next week
[23:52:14] <nxvl> btw, is it a good idea to start merging now? or do i want until the repositories are up?
[23:52:18] <mathiaz> nxvl: so if you wanna discuss something, it's also a good time to do some reasearch and get prepared
[23:52:30] <nxvl> mathiaz: yep, i was forgeting that
[23:52:31] <nxvl> :D
[23:52:35] <owh> mathiaz: Is there going to be a way for the team to contribute to UDS even if they cannot physically attend?
[23:52:35] <nxvl> mathiaz: you are right
[23:52:43] <nxvl> next week, same time, same place
[23:52:50] <owh> WFM
[23:52:56] <ivoks> owh: yes, irc, and voip
[23:52:56] <mathiaz> owh: yes - there are VOIP sessions
[23:53:01] <mathiaz> owh: you can dial in
[23:53:08] <mathiaz> owh: OTOH UDS is in Prague
[23:53:12] <owh> Hmm, VoIP across satellite, yum.
[23:53:22] <mathiaz> owh: so there may be some time zone issue for you ;)
[23:53:26] <sommer> is there going to be icecast as well?
[23:53:45] <mathiaz> sommer: icecast -> voip IIRC
[23:53:48] <owh> mathiaz: Time I can fix, distance not so much.
[23:54:02] <mathiaz> people will be able to hear the discussions in the rooms
[23:54:04] <sommer> mathiaz: ah, gotcha
[23:54:08] * owh observes satellites in a geostationary orbit :)
[23:54:38] <owh> Right, so meeting time then :)
[23:54:51] <owh> Same bat channel works for me.
[23:55:15] <mathiaz> Allright - so next week, same time same place
[23:55:21] <nijaba> thanks mathiaz
[23:55:25] <nealmcb> :-)
[23:55:29] <mathiaz> thanks all for your participation
[23:55:33] <kirkland> adios
[23:55:37] <owh> mathiaz: Merci Beaucoup
[23:55:39] <mathiaz> and we've done some great work for hardy
[23:55:39] <sommer> thanks mathiaz, later on all
[23:55:39] <jdstrand> thanks mathiaz!
[23:55:55] <mathiaz> #endmeeting

MeetingLogs/Server/20080423 (last edited 2008-08-06 16:59:38 by localhost)