=== gnomelogger is now known as supremeDalek === supremeDalek is now known as apachelogger === apachelogger is now known as supremeapachelog === jjohansen is now known as jj-afk [01:57] slangasek, or pitti can someone post http://uec-images.ubuntu.com/server/maverick/20100803.2/ to tracker ? [01:58] slangasek, wrote lp:~ubuntu-archive/ubuntu-archive-tools/trunk/post-amis-to-iso-tracker.py to aid in doing that. it takes input from http://uec-images.ubuntu.com/server/maverick/20100803.2/published-ec2-daily.txt [01:58] smoser: done [01:58] gracias [03:34] can someone (slangasek?) populate the UEC images for iso testing ? use 20100803.2 [06:58] Good morning [07:05] * pitti promotes liblauncher to main, so that we can build the new -efl [07:05] ogra: ^ FYI [07:06] darn, 2 minutes after publisher [07:30] pitti: do you have the keys to nominate UEC images into the ISO testing tracker ? They are still missing... [07:31] "Ubuntu Server UEC amd64" and "Ubuntu Server UEC i386" [07:32] we want them to point to 20100803.2 [07:32] oh, I thought slangasek did that last night [07:32] he did the EC2 ones [07:33] not the "UEC" flavour [07:33] hm, so I never did that, but let's have a look at Steve's new script [07:33] there is no publication to do, just an entry to the ISo tracker [07:34] (but I have no idea how to trigger that) [07:36] ttx: so how do I get the AMI numbers? [07:36] there is no AMI number for the UEC cloud images [07:36] ah, os it's literally just 20100803.2 ? [07:37] I think so -- it's the tarballs that are at the bottom of http://uec-images.ubuntu.com/server/maverick/20100803.2/ [07:37] I just need the entries on the tracker so that we can register test results -- they just need to say "20100803.2" [07:38] test cases describe where to find them [07:38] ttx: ok, sorry for the confusion; how is that? [07:38] so it should really just be two SQL entries [07:38] \o/ [07:39] perfect ! [07:39] I might bug you in a few to bump the build score for eucalyptus (when it is uploaded) so that we respin at the earliest possible [07:39] Daviey is still running a few tests [07:40] sure [07:40] ttx: buildds are idle, though [07:40] within an hour, it should be uploaded [07:41] pitti: Hmm.. i tried a PPA build and that seems to be doing badly.. been uploaded +25mins, wait time let est. 24mins [07:42] right, PPAs are clogged [07:42] let me know if you need a PPA build score bump [07:42] it would help :) [07:42] I just looked at the ubuntu builders [07:42] Daviey: url? [07:42] https://edge.launchpad.net/~davewalker/+archive/uec-devel/+build/1903331 [07:42] done [07:42] next in line now [07:42] \o/ [07:43] thanks [07:56] pitti: got a question for you. the new eucalyptus adds a recommends:tgt, a mir was filed and accepted for tgt. If we rush build->publish->respin would tgt be blocked in component-mismatches and not reach the cd ? [07:56] ttx: no, recommends just get ignored [07:56] they would appear in c-m, but wouldn't render it uninstallable [07:56] ok [07:56] arh. [07:57] apparently there is another one for which MIR wasn't processed yet [07:57] https://bugs.edge.launchpad.net/ubuntu/+source/libcrypt-openssl-x509-perl/+bug/609992 [07:57] Launchpad bug 609992 in libcrypt-openssl-x509-perl (Ubuntu Maverick) (and 1 other project) "[MIR] libcrypt-openssl-x509-perl (affects: 1) (heat: 432)" [Wishlist,New] [07:58] pitti: so it would not reach the CD ? [07:58] ttx: no [07:58] oh, is that a dependency? [07:58] no it's a recommend as well [07:58] Daviey: right ? [07:58] ttx: I can pre-promote it, looks relatively harmless [07:59] pitti: that would be great. [07:59] * pitti does and updates the bug [07:59] ttx: yes [08:00] pitti: I just pinged lool, asking for a MIR review [08:00] I assume he's be working shortly.. hopefully he'll be able to look at it. [08:01] pitti: since the tgt mir was accepted, is there any way to pre-approve it so that it's accepted on the CD as well ? [08:01] "accepted" == "approved"? [08:01] ttx: can do, yes [08:01] https://bugs.launchpad.net/ubuntu/+source/tgt/+bug/594372 [08:01] Launchpad bug 594372 in tgt (Ubuntu Maverick) (and 7 other projects) "MIR: tgt (affects: 1) (heat: 85)" [Medium,Fix released] [08:02] ttx: ah, sure [08:02] "released"? oh [08:02] so, it's released now, promoted [08:02] \o/ [08:02] right in time for this publisher [08:03] hm, looks like I shouldn't have fixreleased it [08:03] we should have waited for the dependency to show up [08:04] pitti: ok, so now if we upload the new eucalyptus that has those new recommends, build/.publish/respin, the CD should magically get all of those without further human interaction ? [08:07] right [08:07] cool, thx [08:21] pitti: ok, uploaded, "Start in 4 seconds" for the last minute or so [08:22] buildds are very slow again unfortunately [08:22] https://launchpad.net/ubuntu/maverick/+source/eucalyptus/2.0~bzr1218-0ubuntu1 [08:22] I mean, the queue builder to actually assign a build to a buildd [08:22] ah, ok [08:22] ttx: how long does it build? [08:22] ~20 min [08:22] ah, should be fine [08:22] yep [08:22] we have 40 mins until next publisher === jj-afk is now known as jjohansen [08:31] pitti: That slowness should be pretty much fixed in a week, FWIW. [08:31] Although some of it will remain until the next LP release. [08:31] wgrant: ooh! [08:31] wgrant: do you know what causes it? [08:32] some days ago they were fast again, and now they are back to slow; I didn't quite see a pattern there [08:32] It's possible it won't get merged in time, but it should. [08:32] there is no large build queue right now [08:32] pitti: It's mostly to do with the number of builds that have finished since the last scanning cycle. [08:32] well, PPAs have, but they are always full anyway [08:32] A few days ago most of the PPA builders were gone, so there weren't many builds finishing. [08:32] So it scanned quickly. [08:32] aah [08:32] right, that was it [08:32] But now there are lots of builds finishing, and the upload processor takes ~15s per upload, and runs synchronously. [08:33] So then it doesn't scan for ages... leaving time for lots more builds to finish. [08:47] pitti, Daviey: packages built, willbe picked up by next publisher run [08:47] \o/ [08:47] nice [08:48] time for coffee, then ! [08:48] ttx: so, once they are published, I'll rebuild server ISOs and re-post [08:48] ttx: do we also need to update the UEC/EC2 images? [08:48] pitti: no [08:49] only the ISO [08:49] the ISOs [08:51] ok [09:00] pitti: If you notice it's published before I do, would you be kind enough to ping me please? :) [09:00] Daviey: I set up a trigger to respin server ISOs as soon as it's published [09:01] Daviey: but if you need to do anything in between, I can stop it and ping you instead [09:02] pitti: oh, interesting.. it's not a curl + grep + if statement + ./spin_cd.sh .. is it? :) [09:02] Daviey: something like that, yes [09:02] nice :) [09:03] it checks antimony's local mirror [09:03] $ wait-for-package eucalyptus-cloud_2.0~bzr1218-0ubuntu1 && for-project ubuntu-server cron.daily [09:06] ahh.. much better :) [10:12] pitti: 20100804.1 is up on cdimage, could you bump the ISo testing tracker reference ? [10:12] ttx: already at it [10:13] http://cdimage.ubuntu.com/ubuntu-server/daily/20100804.1/ [10:13] ttx: UEC, too? [10:13] no [10:13] Daviey: ^ [10:13] just ISO/amd6' and ISO/i386 [10:13] ttx: done [10:14] pitti: thanks ! [10:20] ah, super - thanks === supremeapachelog is now known as apachelogger [11:01] hm, the UEC install fails. Missing tgt deps on CD [11:02] libibverbs1 librdmacm1 libconfig-general-perl [11:02] grumble [11:03] meh [11:03] pitti: looks like those also need to be cleared, they are in https://bugs.launchpad.net/ubuntu/+source/tgt/+bug/594372 [11:03] Launchpad bug 594372 in tgt (Ubuntu Maverick) (and 7 other projects) "MIR: tgt (affects: 1) (heat: 85)" [Medium,Fix released] [11:04] ttx: Hmm [11:04] i got the same results [11:04] Daviey: they have not been promoted yet, only tgt was [11:05] ttx: right after publisher start :-/ [11:05] arh :) [11:05] * pitti promotes [11:05] :(... pitti, did those depends show in CM? [11:05] yes [11:06] pitti: it needs to be picked up by a publisher run ? arh! [11:06] I now verified that those three don't have any further universe depends [11:06] sorry, but http://people.canonical.com/~ubuntu-archive/component-mismatches.txt is a real mess right now, not easy to see what's relevant [11:06] a lot needs to be fixed in packages themselves [11:06] wow, just looked at CM.. it's huuuuge [11:07] pitti: so ETA is like... 90min ? [11:07] ttx: more like 2 hours [11:07] ok [11:08] pitti / ttx: I trust the ant CM won't be an issue? [11:08] no, most there aren't [11:08] pitti: could we mark the two ISO entries as "rebuilding" so that nobody wastes any ISo tseting effort ? [11:09] report.html on the cdimage directories is the interesting part [11:09] ttx: sure [11:09] done [11:09] great [11:09] http://cdimage.ubuntu.com/ubuntu-server/daily/20100804.1/report.html [11:09] right [11:09] sorry, we should have checked that right away [11:10] no problem, I didn't intend to do that much testing over lunch hour anyway :) [11:10] * Daviey has some admin he should be doing anyway.. :) [11:10] pitti: Thanks for sorting that out. [11:11] let's see who wins first - netbook armel or server :) [11:11] looks like they could both make it at the same time now [11:11] ttx: Is that likely to put us oversize?! [11:11] there is some risk [11:11] on amd64, could be [11:11] for some reason hplip is still in there [11:12] * Daviey files a remove from archive bug report.. that'll teach it a lesson! ;) [11:12] I wonder why though [11:12] http://people.canonical.com/~ubuntu-archive/germinate-output/ubuntu.maverick/print-server shows it's no longer in the seed [11:13] ttx: is hplib on both i386 and amd64, or you've only seen it on amd64? [11:13] bah [11:13] Ubuntu.Maverick server-ship seed [11:14] it was in duplicate places [11:14] ttx: If you can modify the seed before this is spun, then we should be ok on size? [11:14] I can modify the seed, not sure if that will be taken into account in time for the spin though [11:14] pitti: ^ ? [11:14] pitti: Does the spin process re-germinate? [11:15] the publisher runs germinate [11:15] ok, lets try this then [11:15] but ship doesn't have a Task: header [11:15] so I think this should get effective immediately on CD image build [11:16] ttx: do the server-ship seed change now, and I'll rebuild a test image to verify that hplip is gone [11:16] Task: ubuntu-desktop, kubuntu-desktop, kubuntu-mobile, kubuntu-netbook, edubuntu-desktop, xubuntu-desktop, ubuntu-netbook [11:16] it seems gone from the print-server task, anyway [11:16] that's the bit which needs 2 publishers [11:17] changed, committed and pushed [11:17] * ttx greps for hplip to make sure it's really dead now [11:20] pitti: in unrelated news I'll soon polish and propose an alternative chart for the work-items-tracker [11:21] it skips weekend days and concentrates on work done and velocity [11:21] ah, nice [11:21] see before: http://people.canonical.com/~ttx/before.svg [11:21] and after: http://people.canonical.com/~ttx/after.svg [11:21] it looks better in cases where work items keep being added during the subcycle [11:22] but isn't that what you should avoid usually? [11:22] in agile, I'd say yes [11:22] it would stop showing you when you have to postpone stuff or involve more people because you are running behind [11:22] since feature creep is bad [11:22] running over trend line has been very useful [11:22] showing it, I mean [11:22] well, my graph still shows that [11:23] if you don't postpone enough, the dotted line will look unrealistic [11:23] this one avoids you having to reset the trends line after you postpone stuff [11:23] ttx: burn up chart :) [11:24] it concentrates on your velocity, which is hard to derive from the before.svg [11:24] pitti: anyway, it should be opt-in [11:25] the default should still be the "regular" one [11:25] I guess I first need to convince Clint Byrum to reduce the per-user charts [11:25] they take awfully long [11:25] heh, I'm not surpised by that [11:25] not even surprised [11:26] * ttx goes to lunch while the publisher sleeps [11:27] pitti: how did your test image rebuild go ? No more hplip ? [11:27] building now; sorry, was looking at n-l-efl [11:29] ok, back in ~1h [11:49] ttx: http://cdimage.ubuntu.com/ubuntu-server/daily/20100804.2/ [11:49] looks good [11:49] 682M amd64 [12:01] publisher done \o/ [12:02] so, one more publisher to go to for the tgt dependencies to propagate to main [12:03] pitti: splendid! [12:03] pitti, ogra: are these two builds supposed to be in the tracker? [12:03] Netboot arm dove (20100803) 0/1 None [12:03] download info Netboot arm imx51 (20100803) [12:03] ara: no [12:04] ara: still fighting to get netbook-laucher-efl built [12:04] we still build them, but they are not tracker worthy [12:04] ara: ETA 2.5 h [12:04] ara: the current images are uninstallable [12:04] pitti, add another 2-2.5h for the image builds [12:04] oh, oops; I accounted .5 h [12:04] so, 4.5 h then [12:05] pitti, ok, thanks for the update, can you mark them as "rebuilding" so people don't get confused? [12:05] well, the armel buildds have USB disks [12:05] ogra: that's quite long -- I thought image builds would by and large need IO, and not much CPU [12:05] dont expect high speed :) [12:05] ara: I didn't think I added them in the first place? [12:05] ara: ah, sorry; "netboot" [12:05] I misread as "netbook" [12:05] :D [12:05] netboot should vanish from the tracker completely [12:05] ara: done [12:05] ogra: no netboot for arm? [12:05] ok, I can disable them [12:06] pitti, thanks [12:06] ogra: done [12:06] pitti, as long as they dont break d-i buiulds i'm happy to keep them building, but they dont need to be on the tracker [12:07] FTR, I went to all the red bugs and triaged/commented them; wubi and offline OEM are broken, the rest looks good so far [12:08] ev: if you have some minutes, could you have a look at bug 613288 and bug 600578? do we need any further information? [12:08] Launchpad bug 613288 in ubiquity (Ubuntu) "wubi installation failed - boot configuration store could not be opened (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/613288 [12:08] Launchpad bug 600578 in wubi "installer drops into grub shell after rebooting from windows. (affects: 1) (heat: 8)" [Undecided,New] https://launchpad.net/bugs/600578 [12:08] (they might be duplicates, but hard to say for me) [12:10] ev: especially the latter confuses me -- I thought wubi would always start right out of windows, and not touch grub etc.? [12:15] pitti, it does add an entry in the grub menu, so it looks like another partition, when it is not [12:15] http://blog.cyphermox.net/2010/06/wubi-installing-ubuntu-inside-windows.html [12:15] ara: ah, thanks [12:15] * pitti never saw it [12:15] then again, I just ordered a new laptop, which will come with the unavoidable win 7 [12:15] so for the first time in years I will actually have an otherwise useless windows to test those bits :) [12:16] :) [12:20] I'm out for a quick lunch; can't do anything for the next 40 mins anyway [12:22] enjoy your meal! [13:05] re [13:05] ara: I did, thanks [13:06] ttx: ok, tgt seems happy again, triggering images [13:06] pitti: heh, holding my breath :) [13:14] * ara -> lunch [13:16] pitti: published, looks ok at first glance [13:16] pitti: if they look ok to you too, please promote to ISO tracker [13:17] ttx: http://cdimage.ubuntu.com/ubuntu-server/daily/20100804.3/report.html looks spotless [13:17] so it should be fine [13:17] ttx: congrats, you won the race against netbook/armel :) [13:17] yay [13:17] server wins again [13:17] added to tracker [13:17] thx [13:27] ttx: anything new in server-land which should be mentioned on https://wiki.ubuntu.com/MaverickMeerkat/TechnicalOverview ? [13:28] pitti: I'll have to think about that. smoser will probably have a blurb abour kernel upgrades in cloud images [13:30] pitti: Can we have a note that Xfce4 was upgraded to the current 4.6.2 release? [13:31] Also, the systray bug is gone in Xubuntu. Should I update the notes? [13:31] charlie-tca: oh, that bug is still open [13:31] charlie-tca: I'm currently editing some GNOME parts, but please do afterwards; I'll ping you [13:31] Thanks [13:37] charlie-tca: please go ahead [13:37] Thank you [13:37] ttx: Trying the new iso now [13:38] Daviey: I'm on topology1 as well [13:38] ttx: I just noticed it was published, so you have a head start [13:38] * Daviey contemplates a cron job to check for new ISO's and email him. [13:39] Daviey: you mean you are not subscribed to the new candidates yet ? [13:41] pitti: was the log I added to the wubi bug any use to you? [13:41] davmor2: not sure; I pinged ev about it [13:41] davmor2: it didn't show any error, anyway [13:42] ttx: The subscribe button on iso.qa.ubuntu.com doesn't seem to work [13:42] ara: ^ ? [13:42] Daviey: are you logged in? [13:42] It's showing that i am. [13:42] i can try cycling [13:43] Daviey: hang on a second [13:43] Daviey: it's working here [13:43] Daviey: so the all-in-one installs now [13:44] http://iso.qa.ubuntu.com/qatracker/test/4389 -> hit Subscribe -> nothing shows in "My Subscriptions" [13:44] Daviey: rebooting to see if it starts [13:44] Daviey: have you selected the test to subscribe to the checkbox on the right? [13:44] ttx: groovy.. is it worth me replicating that... ISO has just finished burning [13:45] davmor2: err, "no comment"... /me sulks away thinking about moaning to mpt about usability :) [13:46] Daviey: registration fails, might be the bug you ran into [13:46] Daviey: you need to select the tests with the checkboxes and then click on subscribe. The person icon will change to colour when you click on subscribe [13:46] ttx: In that case, i'll give it a spin [13:46] see if i get the same result [13:46] * ttx reboots to see if that's transient [13:46] I'm pretty sure it is [13:46] Daviey: At the time it was the only it could be setup, there are possibly better ways now [13:47] davmor2: Oh aye.. i just thought i could go to the page and hit subscribe.. I've mastered it now :) [13:47] Thanks davmor2 [13:50] Daviey: np's [13:55] Daviey: beh, it looks very brittle [13:56] ttx: Did it work that time? [13:56] no [13:56] crptic "can't register, log in through admin interface and check cloud status" error message [13:57] Hmm.. i saw that before [13:59] that's a useless catchall error. [13:59] * pitti tries a btrfs install from desktop [14:01] netbook-launcher-efl : Depends: libevas-svn-05-engines-x but it is not installable [14:01] gar, what now?? [14:01] ogra: ^ not our day [14:02] it's libevas-svn-06-engines-x now [14:02] * pitti pings asac [14:02] Daviey: we should move that discussion to #ubuntu-server [14:03] agreed [14:18] pitti, hmm [14:19] ogra: I uploaded a fix === jjohansen is now known as jj-afk [14:19] just a dependency issue i guess ? [14:19] yes [14:19] hardcoded -05- abi in depends: [14:20] yeah [14:33] pitti: replied to both Wubi bugs. I'm going to be a bit difficult to reach today as Debconf is going on a field trip and the hotel wifi is spotty at best. Give me a ring on my cell if any further issues arise. [14:34] ev: ah, thanks; I just documented it for now, I don't think we shold block the release on it [14:34] ev: enjoy debconf! [14:35] ev: ah, thanks for the replies [14:36] sure thing, and thanks! [14:44] ah, btrfs desktop install works flawlessly, great job cjwatson/ev [14:46] whohoo! [14:49] ogra: do we need bug 600478 for alpha-3 really? [14:49] Launchpad bug 600478 in livecd-rootfs (Ubuntu Maverick) (and 2 other projects) "livecd.sh should remove foreign subarch headers during livefs build (affects: 1) (heat: 127)" [Medium,Confirmed] https://launchpad.net/bugs/600478 [14:49] ogra: the description makes it sound like an optimization thing, so we could also defer it? [14:49] yeah [14:50] fine for post A3 [14:50] ogra: also, bug 605972 [14:50] Launchpad bug 605972 in jasper-initramfs (Ubuntu Maverick) (and 1 other project) "Need to set hostname to ubuntu during first boot. (affects: 1) (heat: 246)" [Medium,New] https://launchpad.net/bugs/605972 [14:50] yeah [14:50] ogra: this sounds a bit strange -- isn't that something that should be done server-side on image cration? [14:51] that could be done at creation time too, but given that jasper functions similar to casper the logical solution would be to set it the same way [14:52] its cosmetic anyway and you only see the buildd hostname if you swithc to a console while oem-config runs [14:52] ogra: ah, oem-config will fix it, too [14:52] i was actually planning to fix both before A3 but that was at a time when i thought i'd have images on monday :P [14:52] ok, moving to beta then, thanks for the heads-up! [14:53] thanks for notifying :) [14:54] ogra: bug 600359 looks more serious -- sounds like it could seriously break armel boots? [14:54] Launchpad bug 600359 in ureadahead (Ubuntu Maverick) (and 1 other project) "ureadahead generating oom messages during boot. (affects: 2) (heat: 16)" [Medium,Confirmed] https://launchpad.net/bugs/600359 [14:54] perhaps we should disable ureadahead on armel as a workaround, WDYT? [14:54] or isn't it that common? [14:54] well, it doesnt break but kills the splash [14:55] its common on all beagle C series boards (256M) [14:55] ah, it could kill just about anything, no? [14:55] on the XM (512M) i havent seen it yet since we have more serious errors there [14:55] (cant boot at all atm due to MMC issues) [14:56] i havent seen it on the omap4 boards and i suspect it also wont happen on the XM once we can verify [14:57] (i had actually hoped tim gardners fix would make it go away) [14:59] ogra: that's what I initially thought, but I asked on the bug, and it still seems to happen [14:59] right [15:03] ogra: I'll move bug 605831 to beta as well then? [15:05] yeah [15:05] unlikely that we get it fixed today [15:07] ogra: sorry for all those pings, but I'm cleaning up teh alpha-3 checklist to ensure that we didn't forget anything really serious [15:08] yeah, dont worry, i usually do that on friday before the release team meeting [15:08] just ping away as needed :) === bjf[afk] is now known as bjf [15:19] pitti: I've added a reply to ev, with set debug=all inplace it's still droping straight into grubshell with no text other than the standard Gnu Grub Shell info. [15:21] davmor2: thanks [15:42] * pitti looks at https://bugs.edge.launchpad.net/ubuntu/maverick/+bugs?field.searchtext=&orderby=status&field.milestone%3Alist=27561 and is much more satisfied now [15:43] * pitti release-notes the last one [15:47] pitti, btw, seems n-l-efl is in the archive now [15:47] so we should be able to build images [15:47] ogra: yay! [15:47] http://people.canonical.com/~ubuntu-archive/testing/maverick_probs.html [15:47] it indeed just dropped from there [15:47] ogra: it still needs a bit to propagate to syncproxy etc. [15:47] i see it on ports already [15:47] oh, nice [15:48] * pitti checks on antimony [15:53] ogra: ARCHES='armel+omap armel+omap4' daily-preinstalled running [15:53] thanks ! [15:53] so, this will take 2.5 h? [15:53] guessed, i just changed the build scripts to speed them up [15:53] ok; I need to leave at 19:30 [15:53] so it might just about make it on time [15:53] so that I can add it to the tracker [15:53] it used to be 3.5h should be 2 or max 2.5 [15:54] but I guess more people can [15:54] right [15:54] i think GrueMAster can take care for that [15:54] he will also do most of the testing [15:54] so dont worry, arm team will take care [15:59] * ogra goes afk and tries to find something better than waiting for images === jj-afk is now known as jjohansen [17:11] Riddell: According to rmadison, when you promoted libdigest-sha-perl yesterday, you promoted the binaries, but not the source ... [17:34] ScottK: rmadison is quite right, fixed === nigelbabu is now known as nigelb [17:44] ttx, Daviey: TBH I don't quite understand bug 613033; should this be release-noted? [17:44] Launchpad bug 613033 in eucalyptus (Ubuntu Maverick) (and 2 other projects) "eucalyptus-cloud: cloud fails to start on separate install (affects: 1) (heat: 6)" [Critical,Triaged] https://launchpad.net/bugs/613033 [17:45] pitti: I would say it does.. basically if you have a topology of having all componets (except nodes) on one server, and then other servers just being nodes - we are ok [17:46] If you split the components (which should be supported), we are running into problems [17:46] ttx, Has gone home for the delay, and i'm reluctant to speak on his behalf on this. What is the deadline that you need the text by? [17:46]