20100914

Actions

  • roaksoax to announce updated cluster packages when they land, ETA before finalfreeze
  • jib to check with kim0 if he can attend meetings regularly ro not

Minutes

Scribe & chair was jib

Topic: Review ACTION points from previous meeting

  • Daviey to contact upstream eucalyptus, regarding issues that seem to be blocked
    • Blocked issues are resolved, but new ones have come up. Hope to be fixed by FinalFreeze (637659)

  • jjohansen to follow up on 574910 with the current direction

    • Followed up with comments, but have not found all root causes yet
    • Patches are in -proposed for some causes
  • jjohansen and smoser to report on 613083 and 606373

    • Both have progress, but are with Amazon right now
  • Mathias to review the puppet server guide additions
    • They were done during the meeting
  • RoAkSoAx to email ubuntu server mailing list with update of cluster stack progress

    • Work wasn't completed yet, but expected this week

Topic: Maverick develoment

  • Very little development going on in this part of the cycle. It's now about bug fixes, testing and stabilization
  • Critical issues for release are listed here: https://wiki.ubuntu.com/ServerTeam/MaverickReleaseStatus

    • No red flags for release identified
    • 582963 requires upstrart/plymouth expertise and may not be fixable by release

Topic: Weekly Updates & Questions for the QA Team (hggdh)

  • hggdh is spending a lot of time on Eucalyptus testing, but things are looking up
  • A mail will be sent to -devel about the new way of handling regression-* tags on bugs

Topic: Weekly Updates & Questions for the Kernel Team (jjohansen)

  • 621175 (-virtual kernel contains too many modules) was closed but then re-opened recently. jjohansen will investigate if action is needed.

  • 634316 (include amazon EBS performance patch in -virtual kernel) will likely be applied to the Maverick kernel. Lucid will probably not see this patch.

Topic: Weekly Updates & Questions for the Documentation Team (sommer)

  • Sommer is happy about the server guide quality and there are no outstanding reviews at this point

Topic: Open discussion

  • Mathias asked if it's indeed sane to drop the Hadoop packages we sync from Debian and only offer the Cloudera ones. Consensus was yes, as the Debian Hadoop packages don't currently work and not even it's author recommends them.

Topic: Announce next meeting date and time

  • Tuesday 2010-09-21 at 1800 UTC - #ubuntu-meeting

Log

[19:00] <jiboumans> #startmeeting
[19:00] <MootBot> Meeting started at 13:00. The chair is jiboumans.
[19:00] <MootBot> Commands Available: [TOPIC], [IDEA], [ACTION], [AGREED], [LINK], [VOTE]
[19:00] <SpamapS> ._-=o/
[19:00]  * SpamapS is on fire
[19:00] <jiboumans> [TOPIC] Review ACTION points from previous meeting
[19:00] <MootBot> New Topic:  Review ACTION points from previous meeting
[19:00] <jiboumans> i'll be your scribe/chair for the day
[19:00] <kirkland> howdy
[19:00] <jiboumans> Daviey to contact upstream eucalyptus, regarding issues that seem to be blocked.
[19:00] <Daviey> done.. but a new issue arisen
[19:01] <jiboumans> do tell
[19:01] <Daviey> Hmm
[19:01] <Daviey> currently we are seeing a meta data service failure
[19:02] <Daviey> upstream have been notified by myself qand hggdh
[19:02] <Daviey> I have a trivial patch that seems to restore functinaility
[19:02] <Daviey> but really want upstream to dicuss it
[19:02] <jiboumans> ok, that's bug 637659
[19:02] <ubottu> Launchpad bug 637659 in eucalyptus (Ubuntu Maverick) "metadata service returns 500 error" [High,Confirmed] https://launchpad.net/bugs/637659
[19:02] <jiboumans> gotcha
[19:02] <Daviey> yes!
[19:02] <jiboumans> alright, let's move on then
[19:02] <jiboumans> jjohansen to follow up on bug 574910 with the current direction
[19:02] <ubottu> Launchpad bug 574910 in linux-ec2 (Ubuntu) "High load averages on Lucid while idling" [Undecided,In progress] https://launchpad.net/bugs/574910
[19:03] <jjohansen> jiboumans: added some comments, I haven't nailed down any specific load averages
[19:03] <jiboumans> so what are we doing right now?
[19:04] <jjohansen> jiboumans: basically we need to search down every problem that could be causing high load averages
[19:04] <jjohansen> there are a few known already
[19:04] <jjohansen> the write back patches are in proposed.
[19:05] <jjohansen> The high number of wakeups, is not properly address and last I saw still being examined upstream
[19:05] <jjohansen> but there is something else going on too, and I haven't been able to replicate
[19:06] <jiboumans> understood
[19:06] <jiboumans> moving on: jjohansen and smoser to report on bug 613083 bug 606373 and next meeting
[19:06] <ubottu> Bug 613083 on http://launchpad.net/bugs/613083 is private
[19:06] <ubottu> Launchpad bug 606373 in linux (Ubuntu Maverick) "cloud-init output does not get to console when booted with pv-grub and ramdisk" [High,Confirmed] https://launchpad.net/bugs/606373
[19:07] <jiboumans> kernel upgrade in the cloud bugs
[19:07] <jjohansen> 613083 has progress but we are not sure when the fix will hit
[19:07] <smoser> Amazon is looking into it.  Last i was aware, there were 2 bugs, one of them fixed on their side.  They indicated that no changes would be necessary in our kernels.
[19:07] <jiboumans> are tehy both on amazon?
[19:07] <jjohansen> yes
[19:07] <smoser> yes
[19:08] <jjohansen> 606373 I thought I had a working fix, but smoser is reporting it doesn't work for him and we haven't had a chance to compare yet
[19:08] <jiboumans> ok, so best we can do is track and provide information i suppose
[19:08] <smoser> bug 606373
[19:08] <ubottu> Launchpad bug 606373 in linux (Ubuntu Maverick) "cloud-init output does not get to console when booted with pv-grub and ramdisk" [High,Confirmed] https://launchpad.net/bugs/606373
[19:09] <smoser> jiboumans, yes. for 613083, we should have nothing to do.  I have reasonable confidence that it will be fixed.
[19:09] <jiboumans> sounds like we've done what we can from our side
[19:09] <jiboumans> moving on
[19:09] <jiboumans> sommer and mathiaz to sync up regarding puppet server guide
[19:10] <mathiaz> jiboumans: I haven't had to review the puppet section
[19:10] <mathiaz> jiboumans: it seems that it landed though
[19:10] <sommer> it's pretty simple so shouldn't take too long :)
[19:10] <mathiaz> sommer: ok - I'll give it a look
[19:10] <sommer> mathiaz: thanks I appreciate it
[19:10] <jiboumans> [ACTION] mathias to review puppet server guide entry
[19:10] <MootBot> ACTION received:  mathias to review puppet server guide entry
[19:11] <jiboumans> RoAkSoAx to email ubuntu server mailing list with update of cluster stack progress
[19:11] <RoAkSoAx> o/
[19:12]  * jiboumans waits for roaksoax
[19:12] <RoAkSoAx> well unfortunately I didn't email because we've been working on getting the packages done and I wanted to provide an update after the work was done given that right after last meeting we received further input on our issues that kinda helped us on see how to resolves our issues
[19:13] <jiboumans> roaksoax: so what's the current status?
[19:13]  * jiboumans also notes it's finalfreeze on thursday
[19:13] <RoAkSoAx> jiboumans: packages should be in shape to hit main before final freeze. ivoks has to do a final upload with the fixes
[19:14] <RoAkSoAx> final fixes
[19:14] <RoAkSoAx> hoefully for today
[19:14] <jiboumans> [ACTION] roaksoax to announce updated cluster packages when they land, ETA before finalfreeze
[19:14] <MootBot> ACTION received:  roaksoax to announce updated cluster packages when they land, ETA before finalfreeze
[19:14] <ttx> RoAkSoAx: on the release front it's tracked through bug 635884
[19:14] <ubottu> Launchpad bug 635884 in redhat-cluster (Ubuntu Maverick) "redhat-cluster fails to build from source in maverick" [High,Confirmed] https://launchpad.net/bugs/635884
[19:14] <RoAkSoAx> indeed
[19:14] <jiboumans> [TOPIC] Maverick develoment
[19:14] <MootBot> New Topic:  Maverick develoment
[19:15] <jiboumans> all quiet on the western front; most development items are now around process, documentation, etc. most of our time is spent on testing, bug fixing & stabilization
[19:15] <jiboumans> work item tracker looks all good from here and no real items left on the critical path
[19:15] <jiboumans> i'll hand over to ttx for the release critical issues
[19:15] <SpamapS> jiboumans: burnup is fixed btw. ;)
[19:15] <ttx> https://wiki.ubuntu.com/ServerTeam/MaverickReleaseStatus is almost up to date, smoser sneakily fixed bug 623609 15 minutes before the meeting
[19:15] <jiboumans> spamaps++
[19:15] <ubottu> Launchpad bug 623609 in grub2 (Ubuntu Maverick) "grub-pc needs some help in uec instances" [High,Fix released] https://launchpad.net/bugs/623609
[19:16] <smoser> it was fixed monday. silly launchpad just didn't know it.
[19:16] <ttx> right, is silly LP fixed now ?
[19:16] <ttx> silly LP did not close the bugs automatically anymore.
[19:17] <ttx> on Daviey's plate we have:
[19:17] <Daviey> :(
[19:17] <ttx> Bug 617053
[19:17] <ubottu> Launchpad bug 617053 in eucalyptus (Ubuntu Maverick) "on upgrade from 1.6.2, euca_upgrade should preserve DISABLE_ISCSI="Y"" [Medium,Triaged] https://launchpad.net/bugs/617053
[19:17] <ttx> Bug 637659
[19:17] <ubottu> Launchpad bug 637659 in eucalyptus (Ubuntu Maverick) "metadata service returns 500 error" [High,Confirmed] https://launchpad.net/bugs/637659
[19:17] <ttx> I think both are clsoe to get a fix
[19:17] <Daviey> (both would benefit from upstream reading)
[19:17] <ttx> 2 new bugs have been assigned to JamesPage, I'll review and sponsor the proposed fixes
[19:18] <Daviey> both seem to be fixed in bzr.. but waiting for them
[19:18] <ttx> the redhat-cluster one we already mentioned
[19:18] <ttx> that leaves Bug 632696, I'll sync with jdstrand later today
[19:18] <ubottu> Launchpad bug 632696 in libvirt (Ubuntu Maverick) "libvirt won't start a VM with serial or console when apparmor is enabled" [High,New] https://launchpad.net/bugs/632696
[19:19] <ttx> On the Other Critical/High Maverick bugs list, the only one with concern is, I think;
[19:19] <ttx> Bug 582963
[19:19] <ubottu> Launchpad bug 582963 in apache2 (Ubuntu Maverick) "SSL pass phrase dialog can't read input" [High,Triaged] https://launchpad.net/bugs/582963
[19:19] <jdstrand> ttx: hallyn mentioned he is tracking that down more since only he can reproduce it atm
[19:19] <ttx> jdstrand: ok
[19:19] <hallyn> i think it needs to be 'invalid' again, haven't reproduced it since
[19:20]  * hallyn wonders if there's an fd leak...
[19:20] <ttx> that passphrase bug is slightly complex since it's an interaction/plymouth thing, and we broke things last time we tried to fix it
[19:21] <ttx> zul: could you ask someone more knowledgeable on upstart how we could fix that, and if it's safe to do before release ?
[19:21] <zul> ttx: ack
[19:21] <cjwatson> ttx: LP fix is in progress, but in the meantime bugs need to be closed manually
[19:21] <jdstrand> hallyn: it wouldn't be the first time
[19:21] <ttx> cjwatson: thx for the update
[19:21] <ttx> Bug 628055 - i'll reproduce it before throwing it to Daviey again
[19:21] <ubottu> Launchpad bug 628055 in libvirt (Ubuntu Maverick) "Instances don't start correctly: Security Labeling error running aa_change_profile()" [High,Incomplete] https://launchpad.net/bugs/628055
[19:22] <ttx> Once you are done with those, please feel free to pick bugs in the last list
[19:22] <ttx> we haven't had that many extra bugs fixed over the last week :)
[19:22] <jiboumans> ttx: are we ready to move on?
[19:22] <ttx> yes
[19:23] <jiboumans> [TOPIC] Weekly Updates & Questions for the QA Team (hggdh)
[19:23] <MootBot> New Topic:  Weekly Updates & Questions for the QA Team (hggdh)
[19:23] <hggdh> life is good, with Eucalyptus swallowing all my time ;-)
=====jsalisbury_brb is now known as jsalisbury
[19:23] <hggdh> apart from that -- expect an email to -devel describing the changes we want to make on the regression-* tags
[19:24]  * SpamapS likes the simplification
[19:24] <hggdh> we propose to drop regression-potential, since it is pretty much only giving us more work
[19:24] <hggdh> but details later
[19:25]  * hggdh is done
[19:25]  * jiboumans waits 5 more seconds for questions
[19:25] <jiboumans> [TOPIC] Weekly Updates & Questions for the Kernel Team (jjohansen)
[19:25] <MootBot> New Topic:  Weekly Updates & Questions for the Kernel Team (jjohansen)
[19:25] <jiboumans> looks like we covered the first two already
[19:25] <jiboumans> but not yet: Bug 621175 - -virtual kernel contains too many modules
[19:25] <ubottu> Launchpad bug 621175 in linux (Ubuntu Maverick) "-virtual kernel contains too many modules" [High,In progress] https://launchpad.net/bugs/621175
[19:26] <jjohansen> should be fixed
[19:26] <ttx> jjohansen: was reopened recently, was it re-fixed ?
[19:26] <smoser> its reported to be re-fixed.
[19:26] <jjohansen> ah no, I hadn't noticed
[19:26] <smoser> a permissions on a maintainance script that would do the module filtering
[19:27] <jjohansen> I'll check, and if it is still broken refix
[19:27] <jiboumans> jjohansen: anything else from the kernel front?
[19:27] <jjohansen> yeah, just a sec
[19:28] <jjohansen> Bug #634316
[19:28] <ubottu> Launchpad bug 634316 in linux (Ubuntu) "include amazon EBS performance patch in -virtual kernel" [Undecided,New] https://launchpad.net/bugs/634316
[19:28] <jjohansen> I am planning on recommending this patch unless anyone has issues with it
[19:28] <jiboumans> ebs performance on ubuntu is something that's been lamented before
[19:29] <jiboumans> so if we have a fix, that'd be a good thing
[19:29] <jiboumans> jjohansen: can we get that one on lucid?
[19:29] <jjohansen> well it does come with risk as it hasn't been well tested by us
[19:29] <jjohansen> no
[19:30] <jiboumans> smoser: do we have ebs/fs regression tests for amis?
[19:30] <jjohansen> its against the pv-ops driver not the full xen
=====easter_egg is now known as easter_egg|off
[19:30] <smoser> we test ebs boot and general worked-ness.
[19:30] <smoser> not extensive disk test, though
[19:31] <jiboumans> i'd like to give it a shot jjohansen
[19:31] <jjohansen> okay, I will submit today
[19:33] <jiboumans> anything else from kernel?
[19:33] <jjohansen> not atm
[19:34] <jiboumans> thanks jjohansen
[19:34] <jiboumans> [TOPIC] Weekly Updates & Questions for the Documentation Team (sommer)
[19:34] <MootBot> New Topic:  Weekly Updates & Questions for the Documentation Team (sommer)
[19:36] <jiboumans> no updates from sommer?
[19:36] <sommer> things are good
[19:36] <sommer> string freeze is in effect, and I feel good about the maverick guide :-)
[19:36] <jiboumans> sommer: ok, all sections (bar the puppet one) are getting review?
[19:36] <jiboumans> excellent
[19:37] <jiboumans> no kim0, so skipping 'Weekly Updates & Questions for the Ubuntu Community Team (kim0)'
[19:37] <mathiaz> jiboumans: puppet review has actually been done :)
[19:37] <jiboumans> mathiaz++
[19:37] <jiboumans> multitasker
[19:37] <sommer> heh, just had that typed, thanks all for your help and feedback
[19:37] <SpamapS> jiboumans: maybe we should remove that section from the agenda.. thats about 4 meetings in a row with no attendance by kim0.
[19:37] <jiboumans> spamaps: point.. action for me to follow up with him
[19:37] <ttx> multitasker, or kernel questions skipper :P
[19:37] <jiboumans> [ACTION] jib to check with kim0 if he can attend meetings regularly ro not
[19:37] <MootBot> ACTION received:  jib to check with kim0 if he can attend meetings regularly ro not
[19:38] <jiboumans> alright
[19:38] <jiboumans> [TOPIC] Open discussion
[19:38] <MootBot> New Topic:  Open discussion
[19:38] <mathiaz> o/
[19:38] <mathiaz> one of the remaining item on the hadoop spec is whether the hadoop package currently in maverick should be dropped from the archive
[19:39] <mathiaz> I've focused on testing the cloudera packages from archive.cloudera.com
[19:39] <mathiaz> and the hadoop packages currently in maverick are not feature complete
[19:39] <mathiaz> I'd like to get your input on whether it makes sense to drop the current hadoop packages from maverick?
[19:39] <jiboumans> as the debian maintainer also not recommended them for production use, i'd be inclined to not have them show up
[19:39] <ttx> mathiaz: makes sense to drop the universe hadoop packages, but only if the cloudera ones can easily be found
[19:40] <SpamapS> mathiaz: we should collaborate on a session at UDS for this, I think we need to formalize the conduits, and put a policy in place for this exact sort of thing.
[19:40] <ttx> I'm not sure making hadoop disappear will help anyone. Removing it to make the others more easily found, yes
[19:40] <mathiaz> ttx: well - it's complicated to define "easily found"
[19:40] <ttx> mathiaz: I think we defined it as "apt-cache search finds it"
[19:40] <jiboumans> mathiaz: place holder package?
[19:41] <jiboumans> our 'hadoop' instead of debians
[19:41] <ttx> jiboumans: wasn't that your definition ?
[19:41] <jiboumans> ttx: yeah, that's my goal
[19:41] <jiboumans> but right now, the 'hadoop' in maverick is not one the maintainer even recommends you install
[19:41] <SpamapS> +1 for having something in apt-cache search that shows hadoop
[19:41] <mathiaz> would a place holder package be accepted given the current release cycle?
[19:41] <SpamapS> -1 for having it be broken
[19:41] <mathiaz> it would be a new empty package
[19:42] <mathiaz> ttx: should I discuss that with the release team
[19:42] <mathiaz> ttx: ?
[19:42] <Daviey> pursue FFe :)
[19:42] <SpamapS> postinst would be  'echo Please see http://xxx for info on running hadoop'
[19:42] <ttx> mathiaz: yes, we cna discuss that with skaet
[19:43] <ttx> mathiaz: I'm not sure what's the less worse solution here
[19:43] <Daviey> "less worse" :)
[19:43] <ttx> I'm just unconvinced by the "remove package" solution
[19:43] <mathiaz> ttx: ok - we shoudl discuss that with the release team then
[19:43] <SpamapS> Daviey: its more better than less worser
[19:43] <ttx> mathiaz: pick a moment when I'm around
[19:43] <mathiaz> ttx: remove + placeholder package would work for you?
[19:44] <ttx> mathiaz: yes
[19:44] <jiboumans> as would for me
[19:44] <mathiaz> ok - great
[19:44] <mathiaz> that's all for me
[19:44] <jiboumans> any other topics?
[19:44] <SpamapS> one
[19:44] <SpamapS> Related to that
[19:44] <ttx> SpamapS: if that's not too orthogonal with the effort you want to push for next cycle
[19:44] <SpamapS> I have stopped updating the Cassandra Ubuntu PPA that I created
[19:44] <SpamapS> There has been zero interest from the community.
[19:45]  * ttx mentions there has been zero interest from upstream in the Terracotta PPA, fwiw
[19:45] <SpamapS> i think until we have packages *in the release* that help you find the conduit/ppa/whatever .. the community will continue to google for it, find upstream's instructions, and ignore any packaging options
[19:46] <SpamapS> Unless we want to attack SEO for "Cassandra Ubuntu" or "Terracotta Ubuntu", then yes, the empty package w/ a pointer to our current status for whatever package we're addressing there, is probably the way to go.
[19:47] <SpamapS> My only reluctance would be if it has any reverse dependencies
[19:47] <ttx> SpamapS: one single package pointing to all of them, or one per project ?
[19:47] <SpamapS> As they'll break horribly.
[19:47] <SpamapS> ttx: one per project.
[19:47] <ttx> SpamapS: ok, then mathiaz's empty package can be reused in Natty
[19:47] <jiboumans> i think this is something we can take off line
[19:48] <SpamapS> Agreed
[19:48] <jiboumans> any other topics?
[19:48] <SpamapS> but suffice to say, the cassandra ubuntu PPA is dead unless somebody picks it up.
[19:48] <jiboumans> going once...
[19:48] <ttx> SpamapS: didn't you sign with your blood to maintain it forever ?
[19:48] <mathiaz> SpamapS: you may wanna delete it then
[19:48] <jiboumans> [TOPIC] Announce next meeting date and time
[19:48] <MootBot> New Topic:  Announce next meeting date and time
[19:48] <jiboumans> Tuesday 2010-09-21 at 1800 UTC - #ubuntu-meeting
[19:49] <jiboumans> #endmeeting
[19:49] <MootBot> Meeting finished at 13:49.
[19:49] <SpamapS> mathiaz: I'll email the cassandra users list before doing that.
[19:49] <ttx> SpamapS: I'd keep it and advertise it with whatever empty-package solution we find for natty
[19:49] <ttx> maybe it will gain momentum...

MeetingLogs/Server/20100914 (last edited 2010-09-16 17:36:55 by jib)