Agenda

Scribe will be: smoser

Items we will be discussing:

Minutes

Beta2 milestone release (ttx)

RC/Final preparation (ttx)

Weekly Updates & Questions for the QA Team (hggdh)

Weekly Updates & Questions for the Kernel Team (jjohansen)

Weekly SRU review: (mathiaz)

Call for Maverick blueprints (ttx)

Open Discussion

Agree on next meeting date and time

Next meeting will be on Wednesday, April 14th at 14:00 UTC in #ubuntu-meeting.

Log

[15:01] <ttx> #startmeeting
[15:01] <MootBot> Meeting started at 09:01. The chair is ttx.
[15:01] <MootBot> Commands Available: [TOPIC], [IDEA], [ACTION], [AGREED], [LINK], [VOTE]
[15:01] <ttx> Agenda for today at https://wiki.ubuntu.com/ServerTeam/Meeting
[15:01] <ttx> Scribe is smoser
[15:01] <ttx> [TOPIC] Review ACTION points from previous meeting
[15:01] <MootBot> New Topic:  Review ACTION points from previous meeting
[15:02] <smoser> i was hoping my name was left off the scribe list
[15:02] <ttx> smoser: nice try
[15:02] <ttx> can't find any action from last week
[15:02] <ttx> [TOPIC] Beta2 milestone release
[15:02] <MootBot> New Topic:  Beta2 milestone release
[15:03] <ttx> So Beta2 should be out tomorrow, server so far looking good
[15:03] <ttx> No more beta2-targeted bugs to fix, which is good
[15:03] <ttx> A few work items left for Beta2 @ http://people.canonical.com/~pitti/workitems/canonical-server-ubuntu-10.04-beta-2.html
[15:04] <ttx> mathiaz, kirkland, hggdh, ttx: please comment on feasibility of those last items by the end of the day
[15:04] <mathiaz> ttx: I'll resume publishing the blog post today
[15:04] <kirkland-mobile> okay
[15:04] <ttx> server-lucid-apt-mirror-ec2: one work item left, blocked on IS delivery, no update that I know of, will have to be postponed
[15:05] <ttx> kirkland: you have "Sign off on B2 candidate tests"
[15:05] <ttx> kirkland: and "    announce why kqemu "went away" in Ubuntu"
[15:05] <kirkland-mobile> will do both today
[15:05] <ttx> hggdh: you have "Perform tests on B2 candidate", is it under way ?
[15:06] <ttx> jiboumans: no news from IS on the new EC2 mirrors ?
[15:06] <jiboumans> ttx: not yet; sent a follow up to James Troup, but nothing yet
[15:06] <hggdh> ttx: I have been unable so far, battling kernel oopses on my laptoip
[15:07] <mathiaz> hggdh: you can use the hardware in the DC
[15:07] <jiboumans> kirkland-mobile: clearly 'sign off' depends on 'perform tests' :)
[15:07] <mathiaz> hggdh: I'm not using the UEC test rig for now
[15:07] <ttx> hggdh: those B2 tests are about ruynning them on the DC testrig anyway
[15:07] <hggdh> mathiaz: I will try it now, laptop stayed on all night without a oops
[15:08] <ttx> hggdh: I suspect your laptop doesn't have to stay up for the duration of the tests ?
[15:08] <ttx> hggdh: please coordinate with mathiaz/kirkland on that, those tests need to be run today (or never)
[15:08] <kirkland-mobile> byobu/screen FTW
[15:08] <hggdh> ttx: the problem I was having to whole day yesterday was for the laptop to stay on. period
[15:08] <ttx> heh
[15:09] <ttx> Last thing about beta§2 we have to discuss is "Ramdisk vs. no-ramdisk for cloud images"
[15:09] <ttx> smoser: please explain the new situation quickly
[15:09] <smoser> well, we're hoping for no ramdisks for beta-2.  In all my tests (and ttx), the images boot successfully without ramdiks.
[15:10] <smoser> yesterday i removed ramdisk from publishing of EC2 images (20100107), but failed to do so for UEC images.  20100107.1 should be out in 90 minutes or so with that change inside.
[15:10] <ttx> So the idea is to have candidates up with "noramdisk", test the heck out of them
[15:10] <ttx> on every single cloud you can put your hands on
[15:10] <ttx> if the results are satisfying by the end of the day, we'll keep them, otherwise we'll revert
[15:11] <ttx> (late in smoser's night)
[15:11] <smoser> if you want to test with no ramdisk, just to make sure you don't hit the previous hang (bug 531494) you can do so easily with 'uec-publish-image --ramdisk=none'
[15:11] <ubottu> Launchpad bug 531494 in Ubuntu Lucid "cloud-init job sometimes not running in cloud images without ramdisk" [High,Fix released] https://launchpad.net/bugs/531494
[15:11] <ttx> smoser: does that sound like a plan ?
[15:11] <smoser> that sounds like a plan to me.
[15:11] <ttx> comments ?
[15:11] <smoser> so far it is infinitely more stable than it was in beta-1 without ramdisk.
[15:11] <ttx> Basically we want to give noramdisk a last chance
[15:12] <kirkland-mobile> just registering todays image will give no ramdisk?
[15:12] <ttx> kirkland: no
[15:12] <kirkland-mobile> or are there special instructions?
[15:12] <ttx> the current one you have to use smoser's command (see #ubuntu-server)
[15:13] <ttx> but new ones should nbe up in 90min
[15:13] <smoser> kirkland-mobile, registering todays image 20100107 will register a ramdisk
[15:13] <smoser> registering 20100107.1 will not (that is due in 90minutes)
[15:13] <kirkland-mobile> ttx: thought it might be nice to have in the log here
[15:13] <smoser> or, alternativley use '--ramdisk=none'
[15:13] <smoser> on your uec-publish-tarball command line
[15:13] <kirkland-mobile> smoser: right, thanks
[15:14] <ttx> uec-publish-image --ramdisk=none image.tar.gz lucid-20100407-noramdisk amd64
[15:14] <ttx> ok, next up
[15:14] <kirkland-mobile> uec-publish-tarball is awesome, btw
[15:14] <ttx> [TOPIC] RC/Final preparation
[15:14] <MootBot> New Topic:  RC/Final preparation
[15:14] <ttx> Release is at the end of the month
[15:14] <ttx> Here are the priorties until then:
[15:14] <ttx> 1/ Release-critical  bugs
[15:15] <ttx> 2/ # Continued specs (eucalyptus-merging, uec-testing, puppet-ec2)
[15:15] <ttx> 3/ Lucid targets of opportunity
[15:15] <ttx> You can find a list of release-critical bugs and targets of opportunity at:
[15:15] <ttx> https://wiki.ubuntu.com/ServerTeam/ReleaseStatus
[15:16] <ttx> The RC bugs in that list take precedence over the spec work.
[15:16] <ttx> Note that FinalFreeze is next week
[15:16] <ttx> So that leaves little room for non-critical bugfixes :)
[15:16] <ttx> Please pay special attention during triage days to catch release-critical issues
[15:17] <ttx> We'll have a sudden influx of server lucid bugs after beta2
[15:17] <ttx> we need to catch regressions and upgrade bugs early
[15:17] <ttx> don't hesitate to subscribe me to those with a note, when you stumble on one
[15:17] <jiboumans> +1 on that; we have our triage days set. don't miss them, don't prioritise over those
[15:18] <ttx> Also we'll have to spend a common effort on openldap, we know there are a few issues around it, and mathiaz won't have the time to fix everything by himself
[15:18] <ttx> so once beta2 is out we'll do some triage/upgrade-bugfix o that
[15:18] <zul> whats wrong with openldap again?
[15:18] <ttx> zul: go triage the openldap bugs and find out :P
[15:19] <mathiaz> zul: we need to make sure that upgrade from hardy to lucid are working correctly
[15:19] <mathiaz> zul: the sticky point is the conversion from slapd.conf cn=config
[15:19] <ttx> Is everything clear on those priorities ? feel free to catch up with me or jib in case anything is unclear
[15:19] <zul> clear
[15:20] <ttx> let's make tyhis a stellar release, we enter the last run
[15:20] <ttx> [TOPIC] Weekly Updates & Questions for the QA Team (hggdh)
[15:20] <MootBot> New Topic:  Weekly Updates & Questions for the QA Team (hggdh)
[15:20] <ttx> hggdh: o/
[15:21] <hggdh> no new from me, execpt that stress testing was fine until day before yesterday ;-)
[15:21] <ttx> hggdh: ISO testcases look better now
[15:21] <hggdh> thanks to you, BTW
[15:22] <ttx> smoser, hggdh: the EC2 testcases still show "'multiple instances" and "single instance" rather than "multiple instances" and "userdata"
[15:22] <smoser> oh? i'll get that fixed.
[15:22] <mathiaz> ttx: are the iso test cases covered now?
[15:22] <ttx> We should fix that before the next respin
[15:22] <mathiaz> ttx: ie should set aside some time to run my iso testing scrips on the candidates?
[15:23] <mathiaz> ttx: should *I* set
[15:23] <ttx> mathiaz: I think we can achieve good coverage without you
[15:23] <mathiaz> ttx: great! thanks
[15:23] <hggdh> mathiaz: I have been running them on the rig
[15:23] <Daviey> hggdh: Would there be merit in blogging for community How to QA'ing + Test cases, and the tracker info.
[15:23] <ttx> Daviey: there would
[15:23] <hggdh> Daviey: it is a good idea, this is the correct moment
[15:23] <ttx> ara is usually blogging about that
[15:24] <mathiaz> Daviey: I think ara made a call for testing already
[15:24] <ttx> anything else for QA ?
[15:24] <hggdh> not from me
[15:24] <ttx> [TOPIC] Weekly Updates & Questions for the Kernel Team (jjohansen)
[15:24] <MootBot> New Topic:  Weekly Updates & Questions for the Kernel Team (jjohansen)
[15:24] <ttx> hggdh: thanks !
[15:24] <ttx> jjohansen: hi
[15:24] <jjohansen> hi
[15:25] <ttx> I don't think we have any issues open with kernel now, everything was fixed
[15:25] <jjohansen> so I have Bug #532553 targeted for post beta2
[15:25] <ubottu> Launchpad bug 532553 in linux-ec2 "linux-image-2.6.32-302-ec2 is missing iptables module xt_recent" [Low,Confirmed] https://launchpad.net/bugs/532553
[15:25] <jjohansen> and that is it atm
[15:25]  * kirkland-mobile points hggdh at jjohansen about his kvm oopses
[15:26] <jjohansen> ah, yeah kirkland-mobile would
[15:26] <kirkland-mobile> :)
[15:26] <jjohansen> yes I will be looking into kvm oops, and we will see what we can find
[15:26] <mathiaz> jjohansen: when is kernel final freeze/upload?
[15:27] <ttx> I think it's passed already
[15:27] <jjohansen> next week
[15:27] <jjohansen> it hits a couple days before everyone elses
[15:27] <ttx> Ah you don't mean kernelfreeze
[15:27] <jjohansen> no
[15:27] <ttx> finalkernelfreeze ;)
[15:28] <ttx> anything else for kernel ?
[15:28] <smoser> jjohansen, kirkland-mobile is there a bug for kvm oops ?
[15:29] <ttx> I think hggdg has one
[15:29] <ttx> hggdh, even
[15:29] <kirkland-mobile> smoser: yes, ask hggdh for it
[15:29]  * kirkland-mobile can't look it up atm
[15:29] <ttx> bug 556919
[15:29] <ubottu> Launchpad bug 556919 in linux "running KVM almost always hard-freezes the host" [High,New] https://launchpad.net/bugs/556919
[15:29] <kirkland-mobile> I suspect that it's cpu model specific
[15:30] <jjohansen> smoser: there are a couple Bug #444365, Bug #458201, Bug #556919
[15:30] <hggdh> <sigh/>
[15:30] <ubottu> Launchpad bug 444365 in linux "BUG: unable to handle kernel NULL pointer dereference while stopping apparmor, kvm related?" [Undecided,Incomplete] https://launchpad.net/bugs/444365
[15:30] <ubottu> Launchpad bug 458201 in linux "kernel stacktrace on volume detach in kvm guest" [Medium,In progress] https://launchpad.net/bugs/458201
[15:30] <mathiaz> hggdh: while doing uec testing you may run into bug
[15:30] <mathiaz> hggdh: bug 556833
[15:30] <ubottu> Launchpad bug 556833 in eucalyptus "System fails to reboot after eucalyptus preseeded instlation" [Undecided,New] https://launchpad.net/bugs/556833
[15:31] <ttx> ok, moving on
[15:31] <mathiaz> ttx: ^^ this is the fails to reboot
[15:31] <mathiaz> bug
[15:31] <ttx> mathiaz: was wondering if it wasn't the grub location preseeded value
[15:31] <ttx> but had no time to look into it
[15:31] <mathiaz> ttx: may be
[15:31] <mathiaz> ttx: we can discuss it later
[15:32] <ttx> since it seems not to be able to find grub .. and netbooting showed some swapping of /dev/sd[ab]
[15:32] <ttx> [TOPIC] Weekly SRU review: https://wiki.ubuntu.com/ServerTeam/KnowledgeBase#SRU%20weekly%20review  (mathiaz)
[15:32] <MootBot> New Topic:  Weekly SRU review: https://wiki.ubuntu.com/ServerTeam/KnowledgeBase#SRU%20weekly%20review  (mathiaz)
[15:32] <mathiaz> http://qa.ubuntu.com/reports/ubuntu-server-team/fixedbugs.ubuntu-server.latest.html
[15:32] <MootBot> LINK received:  http://qa.ubuntu.com/reports/ubuntu-server-team/fixedbugs.ubuntu-server.latest.html
[15:33] <mathiaz> anything worth SRUing on this list?
[15:33] <zul> nope
[15:34] <ttx> nope
[15:34] <ttx> mathiaz: sounds like an outdated list ?
[15:34] <ttx> "fixed from 2010-03-20 to 2010-03-29"
[15:35] <mathiaz> ttx: yes indeed
[15:35] <mathiaz> http://qa.ubuntu.com/reports/ubuntu-server-team/fixedbugs.ubuntu-server.2010-04-05.html
[15:35] <MootBot> LINK received:  http://qa.ubuntu.com/reports/ubuntu-server-team/fixedbugs.ubuntu-server.2010-04-05.html
[15:35] <mathiaz> ^^ this one is empty
[15:35] <mathiaz> so I guess the script broke :/
[15:35] <Daviey> Last modified: Mon, 29 Mar 2010 00:07:48 GMT
[15:36] <mathiaz> There aren't any bugs nominated for the dapper,hardy,intrepid,jaunty,karmic
[15:36] <mathiaz> That's all for the SRU process then
[15:36] <ttx> cool
[15:36] <mathiaz> [ACTION] mathiaz to fix the fixedbugs script
[15:36] <mathiaz> ttx: ^
[15:36] <ttx> [ACTION] mathiaz to fix the fixedbugs script
[15:37] <MootBot> ACTION received:  mathiaz to fix the fixedbugs script
[15:37] <ttx> [TOPIC] Call for Maverick blueprints
[15:37] <MootBot> New Topic:  Call for Maverick blueprints
[15:37]  * mathiaz suggest improving MootBot to allow users to record actions for themselves
[15:37] <ttx> So in a little more than one month we'll have UDS again !
[15:37] <ttx> So it's time to start thinking about what you want to work on for the Maverick cycle
[15:38] <ttx> And filing blueprints about those
[15:38] <mathiaz> ttx: what's the naming scheme for the blueprint name?
[15:38] <ttx> (That's mostly a community call, the Canonical server team will work on a wikipage before creating those)
[15:38] <ttx> server-maverick-*
[15:38] <ttx> and set Jos Boumans as the approver
[15:39] <Daviey> ttx: w.u.c or w.c.com?
[15:39] <mathiaz> Daviey: blueprint.lp.net
[15:39] <Daviey> (no, reference to wikipage0
[15:39] <mathiaz> Daviey: w.u.com - as usual
[15:40] <ttx> Daviey: the brainstorm of what we want to do as a team will probably be on a private wiki, but we'll file the results just after
[15:40] <mathiaz> Daviey: we don't write any specs in w.c.com
[15:40] <ttx> on wiki.u.com and blueprint.lp.net
[15:40] <Daviey> great.
[15:40] <mathiaz> ttx: why use w.c.c?
[15:40] <mathiaz> ttx: and not follow the blueprint.lp.net process?
[15:40] <ttx> Daviey: we usually have plenty of ideas but reality bites us :)
[15:41] <kirkland-mobile> desktop has theirs in w.u.c
[15:41] <ttx> mathiaz: sync with jib, he set up the page
[15:41] <mathiaz> ttx: ok
[15:41] <Daviey> oh sure, if it's commercially related blueprints - keep them on the prvate wiki :)
[15:41] <ttx> maybe I misread the link
[15:41]  * ttx checks
[15:41] <jiboumans> ttx: you did not
[15:41] <jiboumans> i'm scheduling a brainstorm call next week
[15:41] <jiboumans> the results will be published on w.u.c
[15:42] <jiboumans> i requested the input to be on w.c.c
[15:42] <mathiaz> jiboumans: why the input needs to be on w.c.c?
[15:42] <jiboumans> Daviey: as you noted, some are commerical or have commercial implications
[15:43]  * kirkland-mobile has to drop...later.
[15:43] <Daviey> jiboumans: ok.
[15:44] <ttx> [TOPIC] Open Discussion
[15:44] <MootBot> New Topic:  Open Discussion
[15:44] <ttx> Please shoot
[15:45] <RoAkSoAx> Cluster Packages are getting into Debian. Those include ivoks changes... they should be hitting Debian unstable soon, though I'm not sure if they are going to do so in time for us to sync/merge
[15:45] <mathiaz> RoAkSoAx: what's the state of the cluster stack in lucid then?
[15:45] <ttx> RoAkSoAx: they are still in universe ?
[15:46] <mathiaz> I also had a suggestion: drop the dovecot-postfix package
[15:46] <mathiaz> for lucid
[15:46] <RoAkSoAx> mathiaz, ttx cluster-agents, cluster-glue, heartbeat are still in universe
[15:46] <ttx> mathiaz: why?
[15:46] <RoAkSoAx> due to library split need. This is done already in debian
[15:46] <mathiaz> the implementation is not so great (using a different configuration file) to the poin that upstream filed a bug to modify dovecot to check for our alternate configuration file
[15:46] <RoAkSoAx> but as I just checked, they have been uploaded to unstable yesterday :).
[15:47] <Daviey> mathiaz: Is dovecot-postfix broken?  It seems nuts to drop it otherwise?
[15:47] <mathiaz> a lot of user were complaning that changing dovecot.conf doesn't work
[15:47] <ttx> RoAkSoAx: what's the current state of the cluster stack in lucid ?
[15:47] <mathiaz> which makes sense as dovecot-postfix is using dovecot-postfix.conf instead
[15:47] <ttx> ivoks, ScottK: any opinion on what mathiaz proposes ?
[15:48] <mathiaz> so now Ubuntu is known in the upstream community as: are you using Ubuntu? if so change dovecot-postfix.conf instead of dovecot.conf
[15:48] <RoAkSoAx> ttx, Well currently, 3 packages are still in universe (cluster-glue, cluster-agents, heartbeat). cluster-glue needed library split, which is done in debian
[15:48] <mathiaz> bug 511295
[15:48] <ubottu> Launchpad bug 511295 in dovecot "dovecot -n silently reports wrong configuration when using dovecot-postfix" [Low,Triaged] https://launchpad.net/bugs/511295
[15:48] <mathiaz> is the bug from upstream
[15:48] <RoAkSoAx> so we would need to sync them
[15:49] <RoAkSoAx> cluster-agents, and heartbeat were waiting for this library split to be able to upload them to main
[15:49] <Daviey> considering dovecot-postfix was a target for jaunty.. it seems a shame to throw away the work.
[15:49] <mathiaz> LTS being maintainedfor 5 years I don't think it would be a good idea to provide a package that uses a *different* configuration file for 5 years
[15:50] <ttx> mathiaz: yes, an email thread about that might be a good idea
[15:50] <mathiaz> Daviey: well - the idea is good - the implementation proved to be not so great
[15:50] <ttx> [ACTION] mathiaz to propose removal of dovecot-postfix on MLs
[15:50] <MootBot> ACTION received:  mathiaz to propose removal of dovecot-postfix on MLs
[15:50] <mathiaz> we tried - it's an LTS - let's back it off and give another try
[15:51] <RoAkSoAx> ttx, I'll review debian packages, to file sync bugs and apply proper Ubuntu changes if any
[15:51] <mathiaz> another point: zul how is mysql-server-5.0 removal going?
[15:51] <ttx> RoAkSoAx: is it currently broken ? As in, main packages depending on universe packages ? Critical bugs ?
[15:51] <zul> mathiaz: should be ready to go after the freeze is lifted
[15:51] <RoAkSoAx> ttx, it is not broken. Is up an running without library split and latest fixes
[15:51] <mathiaz> and third point
[15:51] <mathiaz> another proposal: drop the ubuntu-serverguide from the archive
[15:52] <ttx> RoAkSoAx: ok. Because it's quite late for non-critical fixes
[15:52] <ttx> mathiaz: rationale ?
[15:52] <mathiaz> sommer: hadn't had time to keep it up-to-date
[15:52] <ttx> mathiaz: we reviewed it recently, it's not that bad
[15:52] <mathiaz> and not having an up-to-date server guide is confusing at least for our users
[15:52] <Daviey> ttx: Does ANYBODY use it?
[15:52] <sommer> mathiaz: I did find time to get most of the major updates in
[15:52] <mathiaz> (see the ldap section for karmic)
[15:53] <ttx> I think the ldap section is the notable exception, rather than the rule
[15:53] <sommer> mathiaz: I'm working on an SRU for that
[15:53] <ttx> and Iwouldn't throw away the baby with the bath water
[15:53] <sommer> mathiaz: and the Lucid instrustions work for karmic
[15:53] <RoAkSoAx> ttx, well the packages that just hit Debian have many fixes, both in upstream and packaging. On the other hand, MIR for cluster-glue was rejected due to library split, which is now in Debian. Cluster-agents, Heartbeat MIRs were rejected too due to its dependency with cluster-glue.
[15:53] <mathiaz> ttx: may be - if so then my  proposal has a *lot* less weight
[15:53] <zul> ttx: i have to update landscape-client next week as well
[15:54] <mathiaz> I still think that the server guide is a great peace of work
[15:54] <ttx> RoAkSoAx: no time now, but we'll continue this discussion
[15:54] <mathiaz> and we should keep maintaining it
[15:54] <mathiaz> I'm just not sure if it should made available as a package
[15:54] <ttx> [ACTION] ttx to follow up with ivoks/RoaKSoax on cluster stack state
[15:54] <MootBot> ACTION received:  ttx to follow up with ivoks/RoaKSoax on cluster stack state
[15:54] <mathiaz> it may be worth discussing it at the next UDS
[15:54] <ttx> mathiaz: sounds late to change direction on that
[15:55] <mathiaz> the content is great - I'm not convinced about the form
[15:55] <ttx> mathiaz: but agreed for future release
[15:55] <sommer> mathiaz: I agree, and should have more time this cycle to do update... day job had a lot of projects coming to a close during the lucid cycle :-)
[15:55] <mathiaz> ttx: right - for lucid dropping the package from the archive seems the only solution for now
[15:56] <mathiaz> ttx: but if the guide is up-to-date
[15:56] <mathiaz> ttx: then it may be worth keeping in the archive
[15:56] <ttx> mathiaz: I think it's good enough.
[15:56] <mathiaz> ttx: ok
[15:56] <ttx> and if not bugs should be filed against it :)
[15:56] <ttx> ok, let's wrap up
[15:57] <Daviey> On another note, are people keeping an eye out to sponsor other peoples work into the archive
[15:57] <ttx> Next meeting, next week, same UTC time, same place
[15:57] <ttx> Daviey: we should all have at least 1 hour of sponsoring time  per week
[15:58] <mathiaz> I'm going through the sponsoring queue every Thursday
[15:58] <mathiaz> http://qa.ubuntu.com/reports/sponsoring/
[15:58] <MootBot> LINK received:  http://qa.ubuntu.com/reports/sponsoring/
[15:58] <ttx> ok, closing the meeting now, feel free to continue discussion on #ubuntu-server
[15:58] <ttx> #endmeeting
[15:58] <MootBot> Meeting finished at 09:58.
[15:58] <Daviey> I was just looking over that
[15:58] <mathiaz> ^^ for packages listed here under the -server seed
[15:58] <Daviey> super.

MeetingLogs/Server/20100407 (last edited 2010-04-08 13:35:05 by d14-69-66-169)