20100127

Agenda

Minutes

Review Action Points from Previous Meeting

The actions were assigned at the previous meeting:

  • Action: smoser to publish karmic cloud image refresh (done)
  • Action: Everyone: update status for your specs before the meeting starts (few left to do)
  • Action: ttx, zul to blog about papercuts, make sure UWN gets the word (done)
  • Action: ttx to send email about criteria and nomination to ubuntu-devel, ubuntu-server
  • Action: smoser to raise thread about the no-ramdisk / -virtual config tradeoff (done)
  • Action: zul, kirkland to unassign themselves from "maybe working on one day" bugs (done)

Alpha3 subcycle planning

Work Items for Alpha 3

Everything that's marked 'medium' or high is fairly sure to make it into 'alpha3'. 'Low' priority specs are targets of opprotunities and we'll do our best to get as many of those done as possible. Please note there's not been much change on this week but we pretty much expect it to be stable now.

No questions were raised in the discussion

Spec Review

A general status review was performed for specifications that were considered to be a "high" priority that are < 30%. The specifciations that were reviewed were:

  • server-lucid-apport-hooks
  • server-canonical-application-support
  • server-lucid-seeds
  • server-lucid-uec-testing
  • server-lucid-ec2-config

jib raised the point that if you have anything you need f2f time with the team or others on the platform team, make sure your code's in a shape to get cracking staight away on monday.

Server Papercuts

Now that project has been announced, we have 4 external papercut members, and 18 nominations so far. The next step is to discuss measurable objectives on how we should approach the papercuts. After much discussion, Thierry proposed that we review the papercuts proposals in the next meeting. It should be quick +1 round.

ACTION: ttx to announce papercuts nominations

Apport Hooks

zul provided an update on the current status and the next steps. zul mentioned that the php5 apport hook is complete, and the eucalyptus apport hook has been updated. However, there has not been any community contributions yet.

ACTION: zul to do another call to action for the apport hooks involvement

PHP 5.3 for Luicd

zul brought up in the meeting that Debian is moving to php version 5.3.1 for Debian Stable. It is currently in experimental and in the process to moving to Debian Unstable. zul mentioned that his preference is to stay with 5.2 since PHP applications that we care about (such as media wiki and drupal 5 still dont support PHP 5.3). Daivey asked if I have thought about what to with the "short tag" depreciation and mentioned that upstream is shipping with it by default. After much discussion about security, what it breaks, and to have it in backports the idea was to send an email to ubuntu-devel to get a wider variety of opinnons.

ACTION: zul to raise php5.3 update on -server and -devel mailinglist, outline which apps are currently not working with 5.3

Weekly Updates & Questions for the QA Team

Soren had perfrormed a UDW class on server automation which was great! Soren also mentioned that the PPA builds have been moved to a team PPA, so more people are being notified of build failures. He is also working on getting more of the qa-regression-testing run on a daily basis. As well as automated ISO testing will be moved to the DC

No other questions were raised for Soren.

Weekly Updates & Questions for the Kernel Team

A summary of launchpad bug 494565 was done by jjohansen. Jjohansen mentioned that the kernel team has changed the suggest configs to builtin (virto, virtnet, and 1 scsi). It booted in limited testing. Test images are available or you could have wait for the next kernel upload, (which is expected to be on friday). There is some concern that configs change will change the behavior so we need wider testing. It was noted that these changes were reasonably risk free but there is still some concern.

Assigned and to-be-assigned bugs

List at http://qa.ubuntu.com/reports/team-assigned/canonical-server-assigned-bug-tasks.html were cleaned up by kirkland and zul.

However there are new bugs there that has to be confirmed and triaged.

Weekly SRU review

List at http://qa.ubuntu.com/reports/ubuntu-server-team/fixedbugs.ubuntu-server.latest.html

The following bugs were nominated: #462172, #480152, #502878, #500457, #504897

Open Discussion

Nothing was raised in the open discussion.

Agree on next meeting date and time

Next meeting will be on Tuesday, Feb 11 at 15:00 UTC in #ubuntu-meeting.

Log

<MootBot> Meeting started at 08:00. The chair is jiboumans.
[14:00] <MootBot> Commands Available: [TOPIC], [IDEA], [ACTION], [AGREED], [LINK], [VOTE]
[14:00] <jiboumans> today's scribe is zul
[14:00] <jiboumans> [TOPIC] last week's action review
[14:00] <MootBot> New Topic:  last week's action review
[14:00] <jiboumans> Action: smoser to publish karmic cloud image refresh
[14:01] <soren> o/
[14:01] <smoser> done.
[14:01] <jiboumans> fwiw, the last meetings logs are here: https://wiki.ubuntu.com/MeetingLogs/Server/20100120 i'll be calling action points in order. please prep your answers
[14:02] <jiboumans> smoser: ta
[14:02] <jiboumans> Action: Everyone: update status for your specs before the meeting starts
[14:02] <jiboumans> looking here, a few aren't updated yet: http://macaroni.ubuntu.com/~pitti/workitems/canonical-server-lucid-alpha-3.html
[14:02] <ttx> jiboumans: bad link
[14:02] <jiboumans> please have them updated a few hours before this meeting at the latest, so they show up there
[14:02] <ttx> http://people.canonical.com/~pitti/workitems/canonical-server-lucid-alpha-3.html
[14:02] <MootBot> LINK received:  http://people.canonical.com/~pitti/workitems/canonical-server-lucid-alpha-3.html
[14:02] <jiboumans> ttx: right, it's moved.. my bad
[14:03] <jiboumans> [ACTION] everyone to update their specs status before the meeting. srsly!
[14:03] <MootBot> ACTION received:  everyone to update their specs status before the meeting. srsly!
[14:03] <jiboumans> Action: ttx, zul to blog about papercuts, make sure UWN gets the word
[14:03] <ttx> you still have one minute to complete before we go to spec review mode
[14:03] <ttx> jiboumans: done
[14:03] <zul> blogged about it need to talk to the UWN people
[14:03] <ttx> zul: showed up in UWN already
[14:03] <ttx> zul: no need to pester them
[14:03] <jiboumans> excellent
[14:04] <zul> ttx: ah sweet
[14:04] <jiboumans> zul++ ttx++
[14:04] <jiboumans> Action: ttx to send email about criteria and nomination to ubuntu-devel, ubuntu-server
[14:04] <ttx> done
[14:04] <jiboumans> excellent; details later on in the agenda
[14:04] <jiboumans> Action: smoser to raise thread about the no-ramdisk / -virtual config tradeoff
[14:04] <smoser> done, and kernel configs changed
[14:05] <smoser> don't know if there is a build available or not, but it wen to fix-commited early this week
[14:05] <jjohansen> they will be in the next kernel upload
[14:05] <jiboumans> ok, i'd like to discuss the details later on in the kernel section
[14:05] <jiboumans> smoser++ jjohansen++
[14:05] <jiboumans> Action: zul, kirkland to unassign themselves from "maybe working on one day" bugs
[14:05] <zul> done
[14:05] <kirkland> jiboumans: done
[14:05] <jiboumans> excellent
[14:05] <jiboumans> moving on
[14:05] <jiboumans> [TOPIC] Alpha3 subcycle planning
[14:05] <MootBot> New Topic:  Alpha3 subcycle planning
[14:06] <jiboumans> What's in the WI tracker is what we're committed to: http://people.canonical.com/~pitti/workitems/canonical-server-lucid-alpha-3.html
[14:06] <jiboumans> everything that's marked 'medium' or higher is faily sure to make it into alpha3. 'low' prio specs are targets of opportunity
[14:06] <jiboumans> we'll do our best to get as many of those done as possible
[14:06] <jiboumans> note there's not been much change on this since last week, but we expect it to be pretty much stable now, barring any external factors messing with it
[14:07] <jiboumans> any questions about the alpha3 plans?
[14:07] <zul> nope
[14:07] <soren> no
[14:08] <jiboumans> excellent, moving on
[14:08] <jiboumans> [TOPIC] spec quick review (ttx)
[14:08] <MootBot> New Topic:  spec quick review (ttx)
[14:08] <ttx> OK, general status is on the same magic page
[14:08] <ttx> I wanted to look into those "high" specs that are < 30 %
[14:08] <ttx> and team members that also are below that threshold
[14:09] <ttx> server-lucid-apport-hooks: 22%
[14:09] <zul> will be updating it today
[14:09] <ttx> server-lucid-canonical-application-support: 20%
[14:09] <zul> will be working on it today
[14:09] <ttx> I'd like to see those completed, as they are spillover from alpha2
[14:09] <ttx> ok
[14:09] <ttx> server-lucid-seeds: 13%
[14:10] <zul> ctdb needs to be worked on
[14:10] <ttx> mathiaz ^
[14:10] <mathiaz> ttx: more WI showed up to figure out why packages haven't been dropped from main
[14:10] <ttx> is that analysis taking long ?
[14:10] <mathiaz> ttx: nope
[14:10] <ttx> ok, so thaey should be burnable quite fast, ok
[14:10] <mathiaz> ttx: yes
[14:10] <ttx> server-lucid-uec-testing: 0%
[14:11] <mathiaz> ttx: was blocked on server access - which is done now
[14:11] <mathiaz> ttx: blocked on a installer bug now
[14:11] <mathiaz> ttx: I need to sync up with kirkland
[14:11]  * kirkland waves at mathiaz 
[14:12] <mathiaz> ttx: all the remaining WI rely on having a working UEC cloud
[14:12] <ttx> mathiaz: hmm, ok. Please update status to reflect the blockage
[14:12] <jiboumans> mathiaz: anything myself or ttx can do to move this along?
[14:12] <mathiaz> I don't think so - for now
[14:12] <ttx> or maybe refactor the WI to make it more obvious where we block
[14:13] <ttx> looks like this is blocking on one or more work yitems that are piori to all others
[14:13] <ttx> prior
[14:13] <ttx> but that doesn't appear in the work item for that spec
[14:13] <ttx> server-lucid-ec2-config: 0%
[14:13] <mathiaz> ttx: and how do you express that a WI is blocking others?
[14:14] <ttx> you don't, but having one INPROGRESS shows what you're working on
[14:15] <ttx> mathiaz: my understanding is that what you're doing (clearing out the basic infra stuff) isn't reflected in any WI currently
[14:15] <mathiaz> ttx: ah - I though INPROGRESS was equal to TODO
[14:15] <ttx> mathiaz: it's equal for the machine
[14:15] <zul> same
[14:15] <ttx> the report considers it the same. But /I/ understand better what you're blocking on
[14:15] <smoser> server-lucid-ec2-config: i'm still working on getting all the config stuff pulled in.  the cloud-config is largely functional in the released images, and more in ec2-init bzr
[14:16] <smoser> there are a few items left to pull in: EBS mounts, ephemeral storage RAID or mount points, runurl
[14:17] <mathiaz> ttx: ok
[14:17] <ttx> smoser: when do you think you can have boothooks/ec2-config completed ?
[14:17] <smoser> before next week.
[14:17] <ttx> smoser: cool, that will give us time to discuss XC2, then ;)
[14:17] <smoser> yeah
[14:17] <ttx> Moving to people under 30%
[14:18] <ttx> mathiaz - 10%
[14:18] <ttx> I think we know where its blocking
[14:18] <ttx> hm, everyone is under 30%
[14:18] <ttx> Let's burn some work items, everyone :)
[14:19] <ttx> jiboumans: anything you wanted to add ?
[14:19] <jiboumans> ttx: to focus on getting the specs in shape for next weeks sprint
[14:19] <jiboumans> if you have anything you need f2f time with the team, or others of the platform team, make sure your code's in a shape to get cracking straight away on monday
[14:20] <jiboumans> that is all, thank you :)
[14:20] <jiboumans> [TOPIC] server-lucid-papercuts (ttx)
[14:20] <MootBot> New Topic:  server-lucid-papercuts (ttx)
[14:20] <ttx> So the effort was announced, and we now have...
[14:21] <ttx> 4 external papercutters members
[14:21] <ttx> and...
[14:21] <ttx> 18 nominations so far
[14:22]  * alexm started nominating today, more to come
[14:22] <ttx> Please continue to blog about it, we are not overwhelmed yet :)
[14:22] <jiboumans> alexm++
[14:22] <jiboumans> ttx: remind us, what's the next step?
[14:22] <ttx> Today I wanted to discuss measurable objectives, if any
[14:22] <ttx> Should we have a papercut-fixing target goal ?
[14:23] <ttx> The desktop papercutters had to fix 10bugs per week, during 10 weeks
[14:23] <ttx> Should we also have a plan, or just people taking up accepted bugs from that bug list when they can ?
[14:23] <mathiaz> ttx: well - it depends on how many submission we have
[14:24] <zul> well we have 18 nominations so far so maybe 2 a week so far and adjust according as time goes by
[14:24] <mathiaz> ttx: have bugs already been accepted?
[14:24] <ttx> no. We can't commit much time to fixing them before FF anyway
[14:24] <mathiaz> ttx: I know I've already declined one bug (about mysql)
[14:25] <ttx> We could have a session during the meeting where we approve them
[14:25] <jiboumans> i do like hte idea of taking x bugs for y weeks (in our case the betas probably)
[14:25] <ttx> though that would overload the meeting
[14:25] <mathiaz> ttx: well - the list of nomination need to be approved at some point
[14:26] <mathiaz> ttx: *then* people can start working on a fix
[14:26] <ttx> ok, I propose that we review the papercuts proposals in next meeting... but lets make it quick, people should review them in advance
[14:26] <ttx> and we have a quick +1 round
[14:26] <jiboumans> agreed
[14:26] <ttx> we'll do it at the end of the meeting, on the remaining time
[14:26] <jiboumans> ttx: want an action of some sort?
[14:27] <ttx> jiboumans: yes, action me on announcing that
[14:27] <jiboumans> [ACTION] ttx to announce papercuts nominations
[14:27] <MootBot> ACTION received:  ttx to announce papercuts nominations
[14:27] <ttx> we want as many papercutters as we can to participate in that review
[14:27] <ttx> so I'll announce externally
[14:28] <ttx> and we'll try to do an x bugs / y weeks thing.
[14:28] <ttx> I'm done
[14:28] <jiboumans> excellent
[14:28] <jiboumans> [TOPIC] server-lucid-apport-hooks update (zul)
[14:28] <MootBot> New Topic:  server-lucid-apport-hooks update (zul)
[14:28] <jiboumans> same for you: what's the current status and what'st he next steps?
[14:28] <zul> so php5, eucalyptus is done for the high priority ones, samba is being worked on now, and vmbuilder hasnt started yet
[14:29] <zul> i havent gotten any contributions from the community yet
[14:29] <ttx> Doing apport hooks is easy ! and fun !
[14:29] <jiboumans> zul, mathiaz: any suggestions where we might find interested parties? at the UDS session it seemed lots of people had input at least
[14:29] <zul> yes it is!
[14:30] <zul> jiboumans: i prodded some people on the security team to help out as well
[14:30] <alexm> ttx: too bad they're not in Perl :P
[14:30] <zul> jiboumans: maybe send an email to ubuntu-server
[14:30] <Daviey> It might be a good idea to have a blog post + link to "how to".
[14:30] <ttx> alexm: if you can do perl, you can do apport hooks, python is not very complex :)
[14:30] <alexm> :D
[14:31] <jiboumans> daviey: agreed
[14:31] <ttx> Daviey: we sorta have that
[14:31] <jiboumans> alexm: from one perl guy to another: think of python as the llama book and you'll be fine ;)
[14:31] <ttx> Daviey: zul blogged about it and the server/apporthooks wikipage has links to doc
[14:32] <ttx> Daviey: but yes, zul should have done a DeveloperWeek session on server apport hooks :)
[14:32] <Daviey> ttx: Without actively searching, i can't think of a  "how to write apport hooks" post or wiki page i've seen linked to
[14:32] <jiboumans> let's do another call then on the lists and point people at the blog/page
[14:32] <Daviey> really!?  /me wonders how he missed that
[14:32]  * Daviey hides.
[14:32] <jiboumans> [ACTION] zul to do another call to action for the apport hooks involvement
[14:32] <MootBot> ACTION received:  zul to do another call to action for the apport hooks involvement
[14:32] <zul> acked
[14:32] <ttx> Daviey: https://wiki.ubuntu.com/ServerTeam/ApportHooks
[14:32] <jiboumans> zul: anything else on apport hooks?
[14:32] <zul> nope
[14:33] <jiboumans> [TOPIC] PHP 5.3 for lucid (zul)
[14:33] <MootBot> New Topic:  PHP 5.3 for lucid (zul)
[14:33] <zul> so alot of people have been asking about php 5.3 for lucid
[14:33] <zul> mathiaz has bugged me a couple of times about it as well
[14:33] <zul> and i checked with debian and they want to move to 5.3 for squeeze
[14:34] <zul> my preference is to stay with 5.2 for lucid but that isnt popular
[14:34] <Daviey> zul: Have you thought about what to do with the short tag deprecation.. AIUI PHP are now shipping with it disabled, but it will break a whole lot of websites.
[14:34] <zul> the reason why to stay with 5.2 for lucid is that php 5.3 breaks applications such as mediawiki and drupal 5 doesnt support it (but we do have drupal )
[14:35] <zul> Daviey: i think debian is disabling it as well
[14:35] <zul> i just want to know what poeple think
[14:36]  * Daviey agrees with zul that 5.2 is safer for an LTS.. However, -security might feel differently.
[14:36] <zul> the discussion for php 5.3 in debian is at http://lists.alioth.debian.org/pipermail/pkg-php-maint/2010-January/006602.html
[14:36] <ttx> zul: so the choice is between old 5.2 or incompatible 5.3 ?
[14:36] <mathiaz> zul: do you have a list of other important webapps that break /don't support 5.3?
[14:36] <ttx> zul: any indication of upstream support lifetime for both versions ?
[14:37] <zul> mediawiki and drupal is the one i know about right now
[14:37] <Daviey> mathiaz: ANY code snippet using <? rather than <?PHP , with the settings shipped as default.
[14:37] <zul> ttx: 5.3 will get security updates faster than 5.2
[14:37] <jmdault> Question: is it possible to ship 5.2 but to have the possibility to have 5.3 in backports?
[14:37] <mathiaz> jmdault: yes
[14:37] <zul> 5.2 and 5.3 dont install nicely at the same time right now
[14:37] <mathiaz> zul: faster - how about longer?
[14:38] <zul> mathiaz: no idea
[14:38] <nijaba> Magento and Joomla seem to have issues as well
[14:39] <ttx> I'd follow the jmdault suggestion
[14:39] <jiboumans> i dont think it's realistic to repackage all those for php5.3 before ff
[14:39] <ttx> if it breaks several apps we care about
[14:39] <ttx> we need to think "platform"
[14:39] <ttx> and ship what is required upper in the stack to run
[14:40] <ttx> not necessarily what the developers want to use
[14:40] <ttx> they can use backports for that
[14:40]  * nijaba fully agrees
[14:40] <ttx>  /but/ I'd throw a discussion on ubuntu-devel about it
[14:40] <ttx> just to take the wolves temperature.
[14:41] <mathiaz> zul: ^^ yes - I'd request for feedback on -devel and -server
[14:41] <jiboumans> if there's no other breakage other than configuration, we may have an option to ship a different default config
[14:41] <mathiaz> zul: outlining which application are *currently* not working with 5.3
[14:41] <ttx> zul: how feasible is it to ship both ? (/me shows PHP ignorance)
[14:41] <zul> ttx: it would be alot of work
[14:42] <soren> They can't coexist.
[14:42] <ttx> zul: ok, it was never done before
[14:42] <mathiaz> ttx: I don't think we wanna do that - as it means one would in be universe the other in main
[14:42] <mathiaz> ttx: and then which one in main?
[14:42] <mathiaz> we're back to square one then
[14:42] <ttx> mathiaz: my question was about how did we handle similar transitions in the past. If there never was 2 PHPs, its not the timt to introduce that
[14:42] <mathiaz> I'd suggest we go with 5.2 in main for lucid, and ship 5.3 in -backports if there is a lot of demand
[14:42] <jiboumans> [ACTION] zul to raise php5.3 update on -server and -devel mailinglist. outline which apps are currently not working with 5.3
[14:42] <MootBot> ACTION received:  zul to raise php5.3 update on -server and -devel mailinglist. outline which apps are currently not working with 5.3
[14:43] <jmdault> Two versions of PHP don't work
[14:43] <zul> acked
[14:43] <ttx> jmdault: cool, one more reason to hate it
[14:43]  * Daviey seems to remember we had a similar discussion for 5.3 inclusion in karmic.
[14:43] <jmdault> I tried that in the past for Mandriva, and hell broke loose
[14:44] <jiboumans> anything else on the php front zul?
[14:44] <zul> nope
[14:44] <jiboumans> [TOPIC] Weekly Updates & Questions for the QA Team (soren)
[14:44] <MootBot> New Topic:  Weekly Updates & Questions for the QA Team (soren)
[14:44] <jmdault> Daviey: I remember a similar discussion for PHP3 ;-)
[14:44] <soren> Hm...
[14:44] <soren> I don't have anything to report, I think.
[14:44] <soren> Questions?
[14:44] <jiboumans> soren: we're interested in hwo the automated qa thing is going of course
[14:45] <soren> jiboumans: It's going well.
[14:45] <alexm> soren++, yesterday's class on server automation was great
[14:45] <ttx> soren: yep, sounds like funny stuff :)
[14:45] <soren> jiboumans: The PPA builds have been moved to a team ppa, so more people are being notified of build failures..
[14:45] <soren> I'm working on getting more of the qa-regression-testing tests run on a daily basis..
=== cjohnston is now known as FFEMTcJ
[14:46] <soren> and the ISO testing thing will hopefully be moved to the DC soon-ish.
=== FFEMTcJ is now known as cjohnston
[14:46] <mathiaz> soren: what's the ISO testing thing?
[14:46] <jiboumans> soren++ the testing phase was long overdue
[14:47] <ttx> soren: any progress on the server QA position ?
[14:47] <soren> mathiaz: My automated ISO testing thing. I had a UDW session on it last night, if you care to read the logs.
[14:47] <soren> mathiaz: I'll be demoing it next week, if not.
[14:47] <soren> ttx: Yes. "Some".
[14:47] <soren> :)
[14:47] <mathiaz> soren: with screenshots and the framebuffer?
[14:48] <soren> mathiaz: Yup.
[14:48] <alexm> mathiaz: https://wiki.ubuntu.com/MeetingLogs/devweek1001/AutoServerTests
[14:48] <soren> Hang on, let me find a link for you..
[14:48] <mathiaz> soren: alexm: thanks
[14:48] <ttx> alexm++
[14:48] <soren> http://people.canonical.com/~soren/lamplvminstall.avi
[14:48] <MootBot> LINK received:  http://people.canonical.com/~soren/lamplvminstall.avi
[14:49] <soren> That's an automatic LAMP install with LVM "partitioning".
[14:49] <soren> It fails in the end, but it sort of shows a bit of what's going on.
[14:49] <mathiaz> soren: great - thanks
[14:50] <jiboumans> any other questions for soren or the QA team?
[14:50] <jiboumans> ok, moving on
[14:50] <jiboumans> [TOPIC] Weekly Updates & Questions for the Kernel Team (jjohansen)
[14:50] <MootBot> New Topic:  Weekly Updates & Questions for the Kernel Team (jjohansen)
[14:50] <jiboumans> specifically: Summarize outcome of discussion about bug 494565 (smoser, jjohansen)
[14:50] <ubottu> Launchpad bug 494565 in linux "support ramdiskless boot for relevant kvm drive interfaces in -virtual" [High,Fix committed] https://launchpad.net/bugs/494565
[14:51] <smoser> jjohansen, ?
[14:51] <smoser> otherwise i'll just go.
[14:51] <jjohansen> we changed the suggested configs to builtin
[14:51] <jjohansen> ie. virtio, virtnet, and 1 scsi
[14:51] <jjohansen> it boots in limited testing
[14:52] <jjohansen> I have test images for any who want, or you can wait for the next kernel upload
[14:52] <ttx> jjohansen: eta for next kernel upload ?
[14:52] <jjohansen> there is some concern about these configs changing behavior, so we need wide testing
[14:52] <jjohansen> ttx: my guess is friday
[14:53] <jiboumans> jjohansen: can our QA setup help with this?
[14:53] <jjohansen> I don't know
[14:53] <jiboumans> soren: ^
[14:53] <ttx> jjohansen: I didn't see a real conclusion to that ML thread, I guess the kernel team decided it was reasonably risk-free ?
[14:53] <soren> Not really.
[14:53] <soren> Well..
[14:54] <jjohansen> well reasonably risk free
[14:54] <jjohansen> there is some concern
[14:54] <soren> Some, not not anywhere near the amount of testing we'd get by just putting it out there and have people report back when things blow up.
[14:54] <jjohansen> the virt configs always present themselves
=== akgraner` is now known as akgraner
[14:55] <ttx> jjohansen: My concern is if it would blow up some fringe server hardware that always worked well, and nobody tests before release
[14:55] <jiboumans> jjohansen: i guess our real question is how do we know it's safe to put this in an LTS, and when is the latest point we need to decide?
[14:56] <jjohansen> right, there is no guarentee
[14:56] <soren> I'm curious..
[14:56] <jjohansen> we don't know of anything breaking, and don't believe anything should
[14:56] <soren> What driver are people using if this one doesn't work for them?
[14:56] <soren> What is the canonical replacement for it?
[14:57] <jjohansen> soren: which driver are you referring to?
[14:57] <jjohansen> virtio, virtnet, scsi?
[14:57] <soren> the scsi one that seems to be the main concern.
[14:57] <jjohansen> actually I am more concerned on virt
[14:57] <soren> Why?
[14:58] <jjohansen> the scsi driver has been around long enough we are reasonably confident that its bugs have been worked out
[14:58] <jiboumans> i think this dense enough to warrant a face 2 face discussion
[14:58] <jiboumans> as you'll all be at next weeks sprint, i suggest we raise the issue there
[14:58] <jiboumans> is that ok?
[14:58] <jjohansen> good idea
[14:58] <soren> ok
[14:58] <jiboumans> thanks
[14:58] <jiboumans> jjohansen++ for fixing kernels
[14:58] <jiboumans> [TOPIC] Assigned and to-be-assigned bugs: http://qa.ubuntu.com/reports/team-assigned/canonical-server-assigned-bug-tasks.html (ttx)
[14:58] <MootBot> New Topic:  Assigned and to-be-assigned bugs: http://qa.ubuntu.com/reports/team-assigned/canonical-server-assigned-bug-tasks.html (ttx)
[14:59] <ttx> Nothing assigned to team
[14:59] <ttx> I noticed a ew "New" bugs there, those should probably be Confirmed or Triaged if they have an assignee
[15:00] <ttx> One question, about http://launchpad.net/bugs/499520
[15:00] <ubottu> Ubuntu bug 499520 in vm-builder "default uec-image requires at least 300 M of RAM to run - m1.small and c1.medium not needed by default" [High,New]
[15:00] <ttx> smoser: ^
[15:00] <ttx> Do we finally know what caused the surge in RAM usage ?
[15:00] <ttx> I know soren asked a question about it
[15:00] <smoser> i gave soren some feed back. i'm not sure.
[15:00] <soren> I talked to the kernel guys about it.
[15:01] <smoser> my suspicion is upstart, or rather loads of jobs running parrallel that previously ran serial
[15:01] <soren> They seemed to think we were reading the numbers wrong. They made a convincing argument.
[15:01] <smoser> at least in boot that would account for balloon
[15:01] <smoser> soren if we're reading the numbers wrong, then so is OOM
[15:01] <smoser> as that is what pointed out the problem, processes dieing
[15:02] <ttx> ok, I'd want to know where this one is coming from, and where it can be going. Should we just switch our minds to a mode where a minimal image needs 300Mb of RAM...
[15:02] <nijaba> another subject for the sprint?
[15:02] <soren> smoser: It doesn't say anything about OOM in the bug?
[15:02] <ttx> nijaba: yes
[15:02] <ttx> jiboumans: done, next
[15:02] <smoser> soren, well "fails to boot" i think . but maybe its not there.
[15:02] <jiboumans> [TOPIC] Open Discussion
[15:02] <MootBot> New Topic:  Open Discussion
[15:02] <jiboumans> eh, oops
[15:02] <jiboumans> [TOPIC] Weekly SRU review: https://wiki.ubuntu.com/ServerTeam/KnowledgeBase#SRU%20weekly%20review (mathiaz)
[15:02] <MootBot> New Topic:  Weekly SRU review: https://wiki.ubuntu.com/ServerTeam/KnowledgeBase#SRU%20weekly%20review (mathiaz)
[15:03] <jiboumans> that first ;)
[15:03] <kirkland> ttx: i'd like to think a minimal image should run in 256MB of mem
[15:03] <smoser> we no longer need 300Mb of RAM for boot, as swap is being enabled again. so that avoids OOM.
[15:03] <ttx> kirkland: yes, me too.
[15:03] <mathiaz> nothing on the nomination lists
[15:03] <mathiaz> http://qa.ubuntu.com/reports/ubuntu-server-team/fixedbugs.ubuntu-server.latest.html
[15:03] <MootBot> LINK received:  http://qa.ubuntu.com/reports/ubuntu-server-team/fixedbugs.ubuntu-server.latest.html
[15:03] <ttx> kirkland: that's why I talked about "mind adjustment"
[15:03] <mathiaz> anything worth SRU on the list above?
[15:03] <zul> checking
[15:04] <kirkland> ttx: yeah, i'm with you ... honestly, i never run my VM's with less than 512MB, but i think we're doing something wrong if we *require* 300MB just to boot
[15:04]  * soren agrees
[15:04] <soren> I'm running a lot of hardy VM's on 64 MB just fine.
[15:04] <soren> Hm.... with swap, though.
[15:04] <mathiaz> http://qa.ubuntu.com/reports/ubuntu-server-team/fixedbugs.ubuntu-server.latest.html
[15:04] <MootBot> LINK received:  http://qa.ubuntu.com/reports/ubuntu-server-team/fixedbugs.ubuntu-server.latest.html
[15:04] <mathiaz> anything worth SRU on the list above?
[15:04] <zul> 462172, 480152, 502878, 500457
[15:05] <zul> 504897 and that one as well
[15:06] <soren> bug 462172, bug 480152, bug 502878, bug 500457, bug 504897
[15:06] <kirkland> soren: yeah, come to think of it, I'm running hardy in my hosted VPS with 128MB (+swap)
[15:06] <ubottu> Launchpad bug 462172 in samba "samba "Too many files are currently in use."" [Medium,Fix released] https://launchpad.net/bugs/462172
[15:06] <ubottu> Launchpad bug 480152 in samba "Samba service doesn't start automatically" [Medium,Fix released] https://launchpad.net/bugs/480152
[15:06] <ubottu> Launchpad bug 502878 in samba "Samba 3.4.0 won't let win98 clients to connect" [Medium,Fix released] https://launchpad.net/bugs/502878
[15:06] <ubottu> Launchpad bug 500457 in samba "Please backport fix for point&print samba support" [Wishlist,Fix released] https://launchpad.net/bugs/500457
[15:06] <ubottu> Launchpad bug 504897 in nut "megatec_usb problem (did not claim interface)" [Undecided,Fix released] https://launchpad.net/bugs/504897
[15:07] <mathiaz> zul: seems like good candidates - nominate/accept them
[15:07] <zul> k
[15:07] <mathiaz> anything else?
[15:07] <mathiaz> that's all for the SRU review
[15:07] <mathiaz> then
[15:07] <jiboumans> [TOPIC] Open Discussion
[15:07] <MootBot> New Topic:  Open Discussion
[15:08] <jiboumans> going once...
[15:08] <jiboumans> going wtice...
[15:08] <jiboumans> alright
[15:08] <Daviey> BANG
[15:08] <jiboumans> [TOPIC] Agree on next meeting date and time
[15:08] <MootBot> New Topic:  Agree on next meeting date and time
[15:08] <jiboumans> next week we won't have a meeting, as the entire platform team will be at a sprnt
[15:08] <jiboumans> the next meeting will be Feb 11th, same time, same channel
[15:09] <jiboumans> thank you all for your time
[15:09] <jiboumans> #endmeeting
[15:09] <MootBot> Meeting finished at 09:09.

MeetingLogs/Server/20100127 (last edited 2010-02-16 16:36:42 by jib)