20110811

August 4th, 2011, 15:00 UTC in #ubuntu-meeting.

Agenda

Action Items from last meeting

  • ogra to repair burn bot for ubuntu-armel -> arm name

  • NCommander and mahmoh to discuss server ARM image variants
  • NCommander to create/assign FTBFS bugs to ARM team members

Standing Items

Meeting Outcome

[15:08] <ogra_> #startmeeting
[15:08] <meetingology> Meeting started Thu Aug 11 15:08:36 2011 UTC.  The chair is ogra_. Information about MeetBot at http://wiki.ubuntu.com/AlanBell/mootbot.
[15:08] <meetingology> Useful Commands: #topic #action #link #idea #voters #vote #chair #action #agreed #help #info #endmeeting.
[15:08] <ogra_> hmm, i dont have the urls handy, gimme a sec
[15:10] <ogra_> grmbl
[15:10] <ogra_> https://wiki.ubuntu.com/MobileTeam/Meeting/2011/20110811 doesnt exist
[15:10] <ogra_> so i only have action items from a week ago
[15:11] <ogra_> i'll skip these for now ...
[15:11] <ogra_> [topic] standing items
=== meetingology changed the topic of #ubuntu-meeting to: standing items
[15:11] <meetingology> TOPIC: standing items
[15:11] <ogra_> http://people.canonical.com/~platform/workitems/oneiric/ubuntu-armel.html
[15:11] <ogra_> http://people.canonical.com/~platform/workitems/oneiric/ubuntu-armel-oneiric-beta-1.html
[15:12] <GrueMaster> try http://people.canonical.com/~platform/workitems/oneiric/canonical-arm-ubuntu-11.10-beta-1.html
[15:12] <infinity> The first graph seems more broken than I remember it.
[15:13] <ogra_> hmm, i still have to update the WI tracker code to pick up the new name
[15:13]  * NCommander coughs a lng up
[15:13] <NCommander> *lung
[15:13] <ogra_> i guess that causes the weird graphs ... bear with me, its on my todo for after mx5 and ac100 have landed
[15:14] <ogra_> NCommander, do you remember any action items from last week ? (there was no wikipage and i dont remember them)
[15:14] <NCommander> ogra_: there's a wikipage, I sent the link out in the reminder email
[15:14] <ogra_> anyway....
[15:15] <ogra_> [action] ogra to update WI tracker with new team name
[15:15] <meetingology> ACTION: ogra to update WI tracker with new team name
[15:15] <ogra_> [topic] ARM Server Status (NCommander, Daviey)
=== meetingology changed the topic of #ubuntu-meeting to: ARM Server Status (NCommander, Daviey)
[15:15] <meetingology> TOPIC: ARM Server Status (NCommander, Daviey)
[15:15] <NCommander> https://wiki.ubuntu.com/ARM/Meeting/2011/20110811
[15:15] <ogra_> oh whay is that arm now
[15:15] <ogra_> the others are all at /MobileTeam/Meeting/
[15:15] <NCommander> Because all the wiki pages were rearranged
[15:15] <ogra_> up to last week
[15:16] <ogra_> k
[15:16] <ogra_> anyway, how is the server looking ?
[15:17] <NCommander> Server images themselves work pretty well, though I found a bug  that occurs duringnetwork setup if you don't have internet access
[15:17] <infinity> That's unique to ARM?
[15:17] <ogra_> how was the LCX summit wrt arm stuff ?
[15:17] <ogra_> *LXC
[15:17] <NCommander> LXC is semi-foobared on omap4 due to kernel configuration issues, but a new kernel was uploaded yesterday to help to resolve it
[15:17] <NCommander> I'll be confirming LXC works out of the box when I can stop hacking up a lung
[15:17] <ogra_> funny that these werent taken over from .38
[15:18] <ogra_> iirc all these options were enabled in .38 already
[15:18] <infinity> ogra_: .38 had the same missing bits.
[15:18] <ogra_> ah, k
[15:18] <ogra_> just because the USER_NS bug was fix released already
[15:19] <ogra_> anything else for server ?
[15:19] <ogra_> .
[15:19] <ogra_> .
[15:20] <ogra_> .
[15:20] <ogra_> moving on
[15:20] <ogra_> [topic] Kernel Status (cooloney, ppisati)
=== meetingology changed the topic of #ubuntu-meeting to: Kernel Status (cooloney, ppisati)
[15:20] <meetingology> TOPIC: Kernel Status (cooloney, ppisati)
[15:20] <ogra_> neither is here
[15:20] <ogra_> anyone else got something for kernel (apart from the above wrt server)
[15:20] <janimo> should upload the linaro mx kernel today
[15:21] <janimo> sponsor it for linaro
[15:21] <ogra_> mahmoh is constantly cpmplaining that we have no omap4 srever flavour
[15:21] <GrueMaster> I ran into a snag with ipsec because of missing kernel bits.
[15:21] <GrueMaster> Should be fixed with latest kernel upload.
[15:21] <ogra_> yeah
[15:22] <infinity> I'm not convinced we need a -server kernel for a product that's essential a proof-of-concept, but...
[15:22] <ogra_> davidm, do we have any plans to make kernel teams life harder this release with an -omap4-server kernel ?
[15:22] <GrueMaster> If omap4 were a server product, I would see where having a server kernel would be nice.
[15:22] <ogra_> infinity, i agree
[15:22] <davidm> ogra_, I do not have any such plans
[15:22] <ogra_> good :)
[15:22] <ogra_> anything else for kernel ?
[15:22] <ogra_> .
[15:22] <ogra_> .
[15:23] <ogra_> .
[15:23] <ogra_> moving on
[15:23] <davidm> yep
[15:23] <ogra_> [topic] ARM Porting/FTBFS status (NCommander, janimo)
=== meetingology changed the topic of #ubuntu-meeting to: ARM Porting/FTBFS status (NCommander, janimo)
[15:23] <meetingology> TOPIC: ARM Porting/FTBFS status (NCommander, janimo)
[15:23] <ogra_> the ftbfs list looks awful
[15:23] <janimo> nothing to report aside from the occasional givebacks to LP
[15:23] <ogra_> i nearly fell off my chair today when i checked it
[15:23] <janimo> ogra_, hmm it seemed to me it was usual
[15:23]  * janimo checks again
[15:23] <infinity> Yeah, now that we're (well, sometime today) in FeatureFreeze, I'm jumping into FTBFS head-first.
[15:23] <ogra_> wasnt someone looking into ocaml three weeks ago ?
[15:24] <janimo> more than one person. NCommander and linaro gcc folk
[15:24] <infinity> NCommander was poking ocaml with a pointy stick.
[15:24] <infinity> And if/when that gets fixed, we can clean up a lot of our mess.
[15:24] <ogra_> might have left some holes ... but no fix :)
[15:24] <ogra_> right
[15:25] <janimo> emacs is not ARM specific but just happened to be a timing issue, likewise is server team. libvirt is new FTBFS
[15:25] <NCommander> I made some progress with ocaml but no fix
[15:25] <ogra_> yeah, and libvirt is weird
[15:25] <janimo> any progress with the panda cluster? That would also clean up the list quite a bit
[15:25] <ogra_> cjwatson came by in #ubuntu-arm asking about it ... i tried a local build and that fails in even earlier tests
[15:26] <ogra_> so another pair of eyes on libvirt would be appreciated
[15:26] <janimo> ogra_, I'll take a look at libvirt if noone else got it
[15:26] <ogra_> and probably a build with the last change rolled back to see if its introduced through that (which i think is unlikely)
[15:26] <ogra_> [action] janimo to look at libvirt after FF
[15:26] <meetingology> ACTION: janimo to look at libvirt after FF
[15:27] <ogra_> anything else ?
[15:27] <ogra_> .
[15:27] <ogra_> .
[15:27] <ogra_> .
[15:27] <ogra_> moving on
[15:27]  * janimo bets on gcc
[15:27] <ogra_> [topic] ARM Image Status (ogra, NCommander)
=== meetingology changed the topic of #ubuntu-meeting to: ARM Image Status (ogra, NCommander)
[15:27] <meetingology> TOPIC: ARM Image Status (ogra, NCommander)
[15:27] <ogra_> awful !
[15:28] <ogra_> the livefs builders committed suicide, x86 had livecd-rootfs removed and the archive is totally out of sync
[15:28] <davidm> Ouch
[15:29] <infinity> Yeah, the word of the day is "special".
[15:29] <ogra_> i was (trying to) working on the ac100 iamges ssince monday but instead of coding i'm actiually fighting the archive churn
[15:29] <infinity> But none of the special is ARM-specific.
[15:29] <NCommander> least A3 was last week
[15:29] <ogra_> infinity, the chrun is ports specific as usual
[15:29] <cjwatson> livecd-rootfs has been put back, FWIW, and measures taken to stop that happening again
[15:30] <ogra_> and it gets really annoying that i have to spend three days on the archive to actually be able to produce 1-2h of code work
[15:30] <GrueMaster> That might explain why my mirror decided to puke.
[15:30] <infinity> ogra_: Well, the unity churn was an issue (and I smacked that until it got fixed), but the ubuntu-one screwup was out of our hands, for instance.  And the livecd-rootfs thing.
[15:30] <cjwatson> I'm afraid to say that *somebody* has to spend that time on the archive at some point or stuff doesn't get fixed, in aggregate ...
[15:31] <ogra_> infinity, well, the ubuntuone breakage was fixed with a seed change and meta upload
[15:31] <ogra_> i was just to frustrated yesterday to actually do it, but it should be solved now
[15:31] <infinity> ogra_: Oh, they don't intend to reintroduce the gnome client?  The changelog seemed to indicate differently.
[15:31] <infinity> Grr.
[15:31] <ogra_> i dont care
[15:31] <infinity> cjwatson: *nod*
[15:31] <ogra_> if it comes back i'll happily seed it again
[15:32] <cjwatson> FWIW you didn't actually have to change the seeds - rebuilding ubuntu-meta would have sufficed
[15:32] <ogra_> but really, we need to get that archive skew thing fixed, thats why i'm moaning
[15:32] <cjwatson> and that would mean it would automatically revert if the package is reintroduced
[15:32] <infinity> cjwatson: Rebuilding meta would have dropped ubuntuone entirely, if it was being pulled in by the non-existant package.
[15:32] <ogra_> indeed
[15:32] <cjwatson> true
[15:33] <cjwatson> but ogra's seed change was *only* to comment out ubuntuone-client-gnome
[15:33] <infinity> Oh. :)
[15:33] <ogra_> anyway, i think we are burning lots of manhours fighting the archive issues every release ... just felt like mentioning it once again
[15:33] <infinity> Yeah, I didn't look.
[15:33] <cjwatson> for future reference rebuilding ubuntu-meta would've been enough
[15:34] <infinity> ogra_: Now that we're past FF, I'll be keeping a closer eye on archive stuff for us.
[15:34] <infinity> Well, almost past.  Whatever.
[15:34] <ogra_> infinity, and how would that help ?
[15:34]  * infinity looks at a clock.
[15:34] <ogra_> you cant remove the sync issues
[15:34] <infinity> ogra_: Because catching things as they're a problem is much better than letting them stack up for two days and become an unresolvable mess?
[15:35] <ogra_> sure, but thats just curing the symptoms (as we do since 3 years)
[15:35] <ogra_> anyway
[15:35] <ogra_> lets move
[15:35] <ogra_> unless anyone else has something for image
[15:35] <ogra_> .
[15:35] <ogra_> .
[15:35] <ogra_> .
[15:35] <ogra_> moving on
[15:35] <ogra_> [topic] QA Status (GrueMaster)
=== meetingology changed the topic of #ubuntu-meeting to: QA Status (GrueMaster)
[15:35] <meetingology> TOPIC: QA Status (GrueMaster)
[15:36] <GrueMaster> Landscape testing works well.  Running the qa entropy script now, expect results in two weeks.
[15:36] <ogra_> heh
[15:36] <infinity> Plug in a keyboard and hire a cat to walk on it.
[15:36] <GrueMaster> My mirror is slowly being rebuilt, so until it finishes I can't do much automation.
[15:37] <janimo> so much entropy from that rsync activity going to waste
[15:37] <GrueMaster> The entropy test is a server workitem.
[15:38] <GrueMaster> Will look into iSCSI or LAVA today.  Haven't decided which will be more exciting.
[15:38] <GrueMaster> Especially given that I can't reimage with netinstall atm.
[15:39] <GrueMaster> Nothing else to report.
[15:39] <ogra_> any questions ? else i'll move to AOB
[15:40] <ogra_> .
[15:40] <ogra_> .
[15:40] <ogra_> .
[15:40] <ogra_> moving on
[15:40] <ogra_> [topic] AOB
=== meetingology changed the topic of #ubuntu-meeting to: AOB
[15:40] <meetingology> TOPIC: AOB
[15:40] <ogra_> so seeing #linaro there are some changes ahead for us
[15:40] <ogra_> seems x-loader gets dropped
[15:40] <janimo> from where?
[15:40] <ogra_> which means we need to adjust the image builders
[15:41] <ogra_> janimo, by linaro
[15:41] <janimo> can omap boot without it?
[15:41] <GrueMaster> What are they replacing it with?
[15:41] <ogra_> they roll a u-boot that can run without x-loader
[15:41] <GrueMaster> ah
[15:41] <ogra_> see #linaro for further info ...
[15:41] <janimo> getting rid of software wothout losing features is good
[15:41] <ogra_> i just felt that fits into AOB :)
[15:41] <janimo> first time I hear about it
[15:41] <ogra_> me too
[15:42] <ogra_> just wanted to mention that debian-cd will need adjustment
[15:42] <ogra_> and jasper probably too
[15:42] <GrueMaster> It will be interesting to see if it works.
[15:42] <ogra_> thats all from my side
[15:42] <janimo> ah getting rid of the debian package for xloader
[15:42] <janimo> but still need MLO on the disk
[15:42] <GrueMaster> Considering the 2nd stage needs to fit in 64k and u-boot is 227k.
[15:42] <janimo> I was hoping they found a way to not use a bootloader to load a bootloader to load a bootloader as now
[15:43] <ogra_> there is a solution
[15:43] <ogra_> you can add a binary header to the u-boot binary
[15:43] <GrueMaster> Oh.  They're combining x-loader & u-boot into one package with 2 binaries?  Make sense.
[15:43] <ogra_> there are some docs on omappedia about that
[15:44] <ogra_> anyway, no need to discuss details in the meeting, we just need to make the builder changes at the same time the package change
[15:44] <ogra_> and it looks like that will happen after FF
[15:44] <janimo> are they planning it in the oneiric timeframe?
[15:44] <janimo> ok
[15:44] <ogra_> yeah
[15:44] <ogra_> there was talk about an FFe
[15:44] <ogra_> anything else ?
[15:44] <ogra_> .
[15:44] <ogra_> .
[15:44] <ogra_> .
[15:45] <janimo> interesting conqequences of having 6 months vs 1 month release cycles project working on similar things :)
[15:45] <ogra_> heh, yeah
[15:45] <ogra_> going once
[15:45] <ogra_> twice ....
[15:45] <ogra_> ... sold to the dogowning QA guy in portland
[15:45] <ogra_> #endmeeting

Action Items

  • None

ARM/Meeting/2011/20110811 (last edited 2011-08-17 23:29:23 by mcasadevall)