| ScottK | phillw: "If" | 00:53 |
|---|---|---|
| phillw | ScottK: ?? | 00:54 |
| ScottK | [18:39:09] <ScottK> ^^^ would make a nice opportunity target if we respin again, but it's perfectly SRU able. | 00:55 |
| ScottK | [18:44:23] <phillw> ScottK: again?!! :P | 00:55 |
| phillw | he he, I was thinking out loud at how the -release team expect all the ticks in the boxes for 'mandatory' and give the testers about 24 hours to do so... :D | 00:57 |
| ScottK | The release team doesn't respin stuff just for fun. | 00:58 |
| Len-nb | ScottK, how about if ubiquity crashes? Bug #1067566 | 01:14 |
| ubot2 | Launchpad bug 1067566 in ubiquity "ubiquity crashes when "something else" is selected" [Undecided,New] https://launchpad.net/bugs/1067566 | 01:14 |
| ScottK | Not sure. | 01:19 |
| * xnox is very confused how come cjwatson is not in the channel.... | 01:31 | |
| phillw | xnox: I think it is called exhaustion / I want some fresh air / sleep.... | 01:31 |
| phillw | ohh, and I forgot... food! | 01:32 |
| xnox | well he is usually ssh'ed in. | 01:32 |
| xnox | "cjwatson [22:31:57] goes to crash for a while." | 01:33 |
| phillw | xnox: yup, that borg drone has returned to his alcove for a regeneration cycle. | 01:34 |
| xnox | slangasek: stgraber: bug 1067566 maybe related to the latest ubiquity we landed. | 01:34 |
| ubot2 | Launchpad bug 1067566 in ubiquity "ubiquity crashes when "something else" is selected" [Undecided,New] https://launchpad.net/bugs/1067566 | 01:34 |
| xnox | phillw: don't find it funny. | 01:34 |
| phillw | xnox: everyone needs a break, | 01:35 |
| phillw | but,as that is 2 out of 2 who do not like my humour... I'll depart. | 01:36 |
| xnox | ScottK: who can actually respin images? (as in push the relevant buttons / pipes) | 01:42 |
| ScottK | Canonical employed members of ubuntu-release. | 01:43 |
| xnox | ScottK: thanks. | 01:43 |
| micahg | oh? it's not limited to the cdimage group? | 01:43 |
| ScottK | Non-employees don't get shell access into the data center. | 01:43 |
| ScottK | I don't think so. | 01:43 |
| ScottK | I think that is for controlling access to the code the builds the images. | 01:44 |
| xnox | ScottK: i'd thinks it's the gpg-keys... | 01:44 |
| xnox | ScottK: can you check if bug 1067566 affects Kubuntu image? It should be quick only need to see if you can enter manual partitioning.... | 02:06 |
| ubot2 | Launchpad bug 1067566 in ubiquity "ubiquity crashes when "something else" is selected" [Critical,New] https://launchpad.net/bugs/1067566 | 02:06 |
| ScottK | I could yesterday. | 02:06 |
| ScottK | I can check again. | 02:06 |
| xnox | ScottK: before the mass respin or after? | 02:06 |
| ScottK | Need to download the updated images. | 02:06 |
| ScottK | Before | 02:07 |
| xnox | ScottK: yeah I think the new ubiquity with fixes only handed in the Kubuntu 16.1 | 02:08 |
| ScottK | Need to finish one other test in progress on the box I can attempt an install on and then I'll check. | 02:12 |
| xnox | ScottK: thanks. | 02:18 |
| slangasek | xnox: are you staying up to babysit this bug? | 02:20 |
| xnox | slangasek: a little =) looks easy peasy, testing a fix locally. Might upload it into quantal-proposed and let the rest of you folks see what to do about it =) | 02:22 |
| slangasek | xnox: ok | 02:22 |
| * xnox has a fix, now testing combinations. | 02:26 | |
| ScottK | What package is the fix in? | 02:27 |
| xnox | ScottK: ubiquity, pure python, ubi-partman.py in the PageGtk get_grub_choice, and there was on added for the PageKDE as well... | 02:28 |
| xnox | looking at PageKDE now just to see if it's affected by inspection. | 02:28 |
| ScottK | If you can give me a diff, I'll cowboy it onto my existing image and test for you (if it's affected) | 02:28 |
| xnox | ScottK: well. first see if you can step into manual partitioning. if that doesn't break & you can even start an install from within manual partitioning it's alright ;-) | 02:29 |
| ScottK | OK. Still waiting for a backtrace to regenerate after I installed a ton and a half of debug packages for an unrelated problem. | 02:30 |
| xnox | yeah... we had jibel with dbgsym packages in live install catching apt earlier today | 02:31 |
| ScottK | This is a plamsa-netbook goes to 100% CPU and takes an hour or two to die kind of thing. | 02:32 |
| xnox | ScottK: right so I have done the fix in a safe way: added extra safety nets, such that even if qt is unaffected no additional / different harm should be caused. | 02:37 |
| xnox | instead of doing gtk specific way =) | 02:38 |
| ScottK | OK. Done with the other testing. One to this one. | 02:43 |
| ScottK | one/on | 02:44 |
| ScottK | xnox: Looks like Kubuntu is affected. When I click on the manual installation type the partitioning window doesn't open. | 02:51 |
| ScottK | slangasek: Since it looks like a respin is in the offing for sure now, would you please have a look at synaptics, waiting in queue for quantal-proposed as an opportunity target. | 02:53 |
| slangasek | ack | 02:53 |
| ScottK | Thanks. | 02:54 |
| ScottK | xnox: Want to pastebin me your patch? | 02:54 |
| slangasek | xnox: there's already a 52 in the pad | 02:56 |
| xnox | slangasek: in my ink ;-) | 02:57 |
| slangasek | xnox: no, the one down below | 02:57 |
| slangasek | you're 53 now :) | 02:57 |
| * xnox doesn't like this number scheme. best to adress all with lp bug #, as they are unique and do increment as well. | 02:58 | |
| xnox | slangasek: thanks =) | 02:58 |
| slangasek | wubi and preinstalled images also affected by ubiquity, fwiw (oem-config) | 03:00 |
| xnox | slangasek: ubi-partman not invoked in either. | 03:01 |
| === kubuntu is now known as Guest28127 | ||
| === Guest28127 is now known as ScottK-netbook | ||
| xnox | slangasek: so, I don't know how to mark it up. It's not "affected" by the bug, but should be respun? | 03:01 |
| ScottK-netbook | xnox: Waiting for a patch to try. | 03:01 |
| ScottK-netbook | xnox: Can't release out of date, so it's affected. | 03:01 |
| xnox | ScottK-netbook: does it crash in Qt? http://paste.ubuntu.com/1284298/ | 03:02 |
| slangasek | xnox: what ScottK-netbook said - it includes binary packages that we have to respin the image for | 03:02 |
| * xnox likes to know if the problem can be reproduced first. | 03:02 | |
| slangasek | xnox: 19:51 < ScottK> xnox: Looks like Kubuntu is affected. When I click on the manual installation type the partitioning window doesn't open. | 03:03 |
| ScottK-netbook | It may have crashed behind the scenes. The manual partitioning window never opened. | 03:03 |
| slangasek | does that count as "reproduced"? | 03:03 |
| xnox | Hmm.... thanks. /me dropped out network for a second there. | 03:03 |
| xnox | I'd say so =) | 03:03 |
| xnox | ScottK-netbook: there should be tracebacks in the /var/log/syslog | 03:04 |
| ScottK-netbook | OK. I didn't check there. Let me try again. | 03:04 |
| ScottK-netbook | This will take a bit. | 03:05 |
| ScottK-netbook | No traceback. | 03:09 |
| ScottK-netbook | Oct 17 03:06:48 kubuntu os-prober: debug: running /usr/lib/os-probes/mounted/90linux-distro on mounted /dev/sdb1 | 03:09 |
| ScottK-netbook | Oct 17 03:06:48 kubuntu os-prober: debug: running /usr/lib/os-probes/mounted/90solaris on mounted /dev/sdb1 | 03:09 |
| ScottK-netbook | Oct 17 03:06:48 kubuntu ubiquity[9633]: Device /dev/sda5 not found in os-prober output | 03:09 |
| ScottK-netbook | Oct 17 03:06:49 kubuntu ubiquity[9633]: switched to page Form | 03:09 |
| ScottK-netbook | xnox: ^^^ | 03:09 |
| slangasek | xnox: would it perhaps be more idiomatic to do try/except AttributeError pass? | 03:10 |
| slangasek | xnox: also, can this reasonably be unit tested as part of the fix? | 03:10 |
| xnox | the change in ok_handler is bogus, that will not work. | 03:10 |
| ScottK-netbook | Sigh. I just realized I was doing it wrong ... | 03:13 |
| * ScottK-netbook goes back and tries again. | 03:13 | |
| skaet | unseeded universe final freeze email has now been sent out. | 03:14 |
| ScottK | micahg: wins then. | 03:15 |
| ScottK | slangasek: Once I do it right, I definitely get the traceback without xnox | 03:15 |
| ScottK | 's fix | 03:15 |
| slangasek | ok | 03:16 |
| ScottK | xnox: Also I don't get the traceback with your fix. | 03:16 |
| ScottK | Trying to see if I can do an install now. | 03:16 |
| xnox | ScottK: nor a working grub installation =) | 03:16 |
| xnox | well, please try and wait till it actually fails..... | 03:16 |
| ScottK | Will do. | 03:17 |
| ScottK | Got through the manual paritioning and into installing now. | 03:21 |
| ScottK | The respin is phillw's fault. He jinxed us. | 03:27 |
| xnox | meh | 03:31 |
| ScottK | Install's still going. | 03:32 |
| xnox | running in debug mode here and actually I think I got over-protective =) | 03:33 |
| ScottK | Your patch definitely solved an actual problem that I could reproduce. | 03:34 |
| * xnox done auto install, manual, now need to do the same with multidisk | 03:35 | |
| skaet | Thanks for getting this tracked down xnox, ScottK, slangasek. I'm getting a bit too tired to be productive now, so will catch up with the state after some sleep. Good luck. | 03:35 |
| * skaet --> zzz | 03:35 | |
| xnox | skaet: night ;-) | 03:36 |
| ScottK-netbook | Install done. Time to reboot. | 03:40 |
| plars | hmm, the wubi binaries pointed to from isotracker don't seem to be the correct ones - they are still the old ones before the fixed certificate | 03:41 |
| plars | at http://people.canonical.com/~evand/wubi/quantal | 03:41 |
| ScottK | xnox: Kubuntu manual partition install succeed with your patch. | 03:44 |
| ScottK | 18 minutes until the next publisher run ... | 03:44 |
| xnox | ScottK: yeah, but now on gtk side it's preseeded too early. | 03:45 |
| slangasek | plars: yes; the updated ones are published to releases.ubuntu.com | 03:47 |
| micahg | can I retry previously aborted builds still? | 03:47 |
| slangasek | micahg: yes | 03:47 |
| ScottK | slangasek: I verified the synaptics fix from the built dep, so I think it's ready to be copied over at your convenience. | 03:52 |
| ScottK | (or tell me to do it, either way) | 03:53 |
| slangasek | ScottK: ok - I'll copy it at the same time that we get ubiquity | 03:53 |
| ScottK | ok. thanks | 03:55 |
| * ScottK snoozes. | 04:18 | |
| slangasek | xnox: still with us? | 04:33 |
| xnox | slangasek: yes. committing part of a fix. | 04:33 |
| xnox | slangasek: not uploading yet. | 04:34 |
| slangasek | ok | 04:34 |
| xnox | slangasek: opening a second bug as well. | 04:35 |
| slangasek | oh? | 04:36 |
| xnox | discovered while testing this one a bit more. just checking with pristine image without my modifications. | 04:37 |
| slangasek | ok | 04:39 |
| xnox | Ok. now that I understand significance of internal vs removable storage, there is no second bug. | 04:56 |
| xnox | uploading soon. | 04:57 |
| slangasek | great | 05:00 |
| xnox | slangasek: passing the baton over. | 05:11 |
| slangasek | xnox: thanks | 05:11 |
| slangasek | xnox: thanks for sticking with it :) | 05:11 |
| xnox | slangasek: meh =) i'm off to sleep for now ;-) | 05:11 |
| slangasek | well, that's not very good flood control | 05:52 |
| slangasek | ^^ there's the new ubiquity; builds will kick off automatically on the next publisher run | 06:10 |
| stgraber | doh, the only thing I forgot to add a check for in my patch was the one that blew up... thanks xnox for investigating and fixing | 07:00 |
| stgraber | hmm, isn't lubuntu alternate also affected by that bug? I see it wasn't disabled on the tracker | 07:03 |
| * stgraber starts downloading ubuntu server amd64 for another secureboot test while the rest of the world respins | 07:03 | |
| slangasek | stgraber: the ubiquity bug? | 07:18 |
| slangasek | the current respins are for a ubiquity bug, I wouldn't expect any alternates to be affected | 07:18 |
| slangasek | hrm, but oem-config is on the alternates, sigh | 07:20 |
| slangasek | ok, adding lubuntu + ubuntu-server to the respin | 07:23 |
| stgraber | slangasek: right, was thinking of oem-config | 07:37 |
| cjwatson | slangasek: we could have put ubiquity in -updates and avoided respinning alternates/preinstalled/blah ... oh well | 07:45 |
| slangasek | mm, sorry | 07:46 |
| * stgraber reboots for some ubuntu server secureboot fun. be back in 15min-ish | 07:46 | |
| stgraber | cjwatson: bad news in secureboot land. The latest server image still won't install linux-signed | 07:53 |
| stgraber | grabbing the log now | 07:53 |
| stgraber | cjwatson: http://paste.ubuntu.com/1284542/ | 07:53 |
| stgraber | that was with 20121016.1 | 07:54 |
| cjwatson | srsly | 07:54 |
| cjwatson | Oh, argh, base-installer/kernel/override-image | 07:55 |
| cjwatson | stgraber: Hack: boot with base-installer/kernel/override-image=linux-signed-generic | 07:56 |
| cjwatson | This is all rather ironic because due to kernel metapackage changes we are now overriding to the default | 07:57 |
| cjwatson | slangasek: Thoughts on a debian-cd-only respin for this? | 07:58 |
| slangasek | cjwatson: seems tolerable under the circumstances... | 07:58 |
| cjwatson | stgraber: Hmm, a boot parameter override won't work | 07:58 |
| slangasek | can we get a proper fix sorted out for .2? | 07:59 |
| cjwatson | Proper as opposed to ...? | 07:59 |
| slangasek | one that doesn't require a debian-cd hack? | 07:59 |
| cjwatson | This is removing a broken hack | 07:59 |
| slangasek | oh, ok | 07:59 |
| slangasek | then yeah, wfm | 07:59 |
| cjwatson | That said, the landscape is different in precise | 08:00 |
| cjwatson | Because the linux-server package is actually genuinely still different there | 08:00 |
| stgraber | I'm doing a quick secureboot desktop install test as I have never tried one and would feel a lot better knowing that it works here. I guess by the time I'm done with that we'll have another server image for me to test. | 08:00 |
| cjwatson | Oh, wait, linux-image-server/precise/amd64 depends on -generic | 08:00 |
| cjwatson | slangasek: http://paste.ubuntu.com/1284550/ FWIW | 08:02 |
| slangasek | looks good to me | 08:03 |
| stgraber | according to https://wiki.ubuntu.com/Kernel/Dev/Flavours the server flavour was removed as of 12.04, so backporting the seed change to 12.04.2 will be fine | 08:03 |
| * stgraber disappears for another test installer | 08:04 | |
| cjwatson | slangasek: fixed, deployed, padded as [55] | 08:05 |
| cjwatson | stgraber: Yeah, I checked out 'apt-cache show' in my precise-amd64 chroot first | 08:05 |
| slangasek | cjwatson: great | 08:05 |
| cjwatson | The pad is slightly more tolerable now that I've realised that repeated PgUp/PgDn work provided I click somewhere in between | 08:06 |
| cjwatson | (although *what*) | 08:06 |
| infinity | cjwatson: Yeah, PgDn/click/PgDn/click isn't exactly intuitive. :P | 08:12 |
| === yofel_ is now known as yofel | ||
| stgraber | argh, I was really hoping this wouldn't happen... | 08:22 |
| stgraber | I installed Edubuntu DVD amd64, got the shim and grub-signed but no signed kernel | 08:22 |
| stgraber | so can't boot | 08:22 |
| stgraber | /var/log/syslog clearly shows the installer removing linux-signed from the target... | 08:23 |
| cjwatson | Does that happen with Ubuntu? | 08:23 |
| cjwatson | (desktop) | 08:23 |
| stgraber | I'm 99% sure it'll but I'll try it next | 08:23 |
| === henrix_ is now known as henrix | ||
| infinity | Any objections to me copying gdm to release? | 08:24 |
| infinity | (unseeded, etc) | 08:25 |
| cjwatson | stgraber: not sure I can see an obvious reason why that's happening - unless some of the pieces are missing for it to boot in SB mode | 08:25 |
| cjwatson | infinity: fine by me | 08:27 |
| infinity | Done. | 08:27 |
| cjwatson | huh, whoever gave back pandoc/armel is optimistic | 08:28 |
| cjwatson | micahg: did you get your webkit stuff done? | 08:28 |
| infinity | pandoc isn't one of the ones that hangs buildds, is it? | 08:28 |
| Laney | usually times out normally | 08:29 |
| cjwatson | yeah | 08:29 |
| cjwatson | I'd care if we were short of buildd time at the moment | 08:30 |
| stgraber | cjwatson: relevant log: http://paste.ubuntu.com/1284595/ | 08:31 |
| stgraber | I'm grabbing a standard desktop image now | 08:31 |
| cjwatson | well, it thought it needed to install grub-efi-amd64-signed at least | 08:33 |
| cjwatson | and shim-signed | 08:33 |
| stgraber | yeah and then decided to remove linux-image-signed | 08:33 |
| cjwatson | probably need to set -x in /usr/share/ubiquity/check-kernels | 08:34 |
| stgraber | ok, will do with the ubuntu desktop install | 08:34 |
| stgraber | writing the usb stick now so should have the result in 15min or so | 08:35 |
| cjwatson | I suspect it's a ubiquity bug | 08:35 |
| cjwatson | May not be able to fix for 12.10 if it doesn't affect Ubuntu desktop | 08:35 |
| === doko_ is now known as doko | ||
| cjwatson | self.kernel_version is probably busted for signed kernels, but I don't know if that relates | 08:37 |
| cjwatson | http://paste.ubuntu.com/1284607/ possibly but I think that's too risky for release if it doesn't break Ubuntu desktop | 08:38 |
| stgraber | cjwatson: ubuntu desktop is also affected | 08:42 |
| cjwatson | argh | 08:43 |
| cjwatson | argh argh argh | 08:43 |
| cjwatson | need that set -x asap then | 08:44 |
| cjwatson | infinity: Chinese edition image failed to build? | 08:44 |
| stgraber | yep, installing pastebinit on that system, will have it in a sec | 08:44 |
| PeterMahnke | hi release team... this is Peter from the web team | 08:44 |
| PeterMahnke | can I ask an ARM question? | 08:44 |
| cjwatson | stgraber: also ubiquity debug output wouldn't hurt at some point | 08:44 |
| cjwatson | PeterMahnke: go | 08:45 |
| PeterMahnke | the site has an arm download page that says "Ubuntu Server and Ubuntu Desktop for ARM are a general purpose OS for ARM-based systems only. It supports, Calxeda ECX-1000 (Server only), Marvell Armada-XP (Server only) and TI Panda development boards." | 08:45 |
| PeterMahnke | are these correct boards? | 08:45 |
| infinity | cjwatson: *blink* | 08:45 |
| infinity | PeterMahnke: Yeahp. | 08:45 |
| PeterMahnke | ok... the other question is... we are sending people to http://cdimage.ubuntu.com/releases/12.04/release/ for 12.04 downloads | 08:46 |
| PeterMahnke | where do they go for 12.10? | 08:46 |
| PeterMahnke | will it be the same but 12.10? | 08:46 |
| stgraber | cjwatson: /var/lib/ubiquity/install-kernels: http://paste.ubuntu.com/1284614/ | 08:46 |
| stgraber | cjwatson: /var/log/syslog: http://paste.ubuntu.com/1284615/ | 08:47 |
| stgraber | cjwatson: check-kernel with -x: http://paste.ubuntu.com/1284619/ | 08:47 |
| infinity | cjwatson: Failed, or just didn't build at all. Weird. | 08:47 |
| infinity | cjwatson: Oh, wait, no, there it is in my log. The livefses built in my run. | 08:47 |
| cjwatson | PeterMahnke: The ARM OMAP4 images have been moved to releases.ubuntu.com, so http://releases.ubuntu.com/12.10/ | 08:47 |
| cjwatson | Not sure if we care about advertising the OMAP image | 08:47 |
| stgraber | cjwatson: I'll re-run ubiquity with --debug now | 08:47 |
| cjwatson | (3) | 08:47 |
| cjwatson | infinity: I have a "LiveFS (built by adconrad)" failure in my inbox | 08:48 |
| ogra_ | omap3 should stay on cdimage | 08:48 |
| cjwatson | infinity: want it bounced to you? | 08:48 |
| ogra_ | as well as ac100 | 08:48 |
| cjwatson | ogra_: Yeah, I just don't know whether it should be advertised on www.ubuntu.com | 08:48 |
| cjwatson | (or is advertised) | 08:48 |
| infinity | cjwatson: Err, sure. I don't see that in my screen history, which is odd. | 08:49 |
| ogra_ | ah, no, i think we only want to advertise the fully supported arches | 08:49 |
| infinity | ubuntu-i386 on cardamom.buildd finished at 2012-10-15 19:42:41 (success) | 08:49 |
| infinity | ubuntu-amd64 on kapok.buildd finished at 2012-10-15 20:06:34 (success) | 08:49 |
| infinity | ^-- Those were my Chinese builds last night. | 08:49 |
| cjwatson | Uh | 08:49 |
| cjwatson | Not unless you had a time machine | 08:49 |
| infinity | Uh, wrong old screen. | 08:49 |
| infinity | La la la. | 08:49 |
| infinity | ubuntu-amd64 on kapok.buildd finished at 2012-10-16 23:01:38 (failed) | 08:50 |
| infinity | That'd be the one. | 08:50 |
| cjwatson | stgraber: could use a critical-priority bug about this when you have a moment | 08:51 |
| PeterMahnke | so cjwatson where should I send the ARM 12.10 downloaders? | 08:52 |
| stgraber | cjwatson: ok, doing that now | 08:53 |
| stgraber | cjwatson: debug log: http://paste.ubuntu.com/1284628 | 08:53 |
| infinity | cjwatson: I assume someone's respun since then anyway. Or did that fail too? | 08:53 |
| cjwatson | PeterMahnke: http://releases.ubuntu.com/12.10/ I think | 08:54 |
| cjwatson | infinity: http://china-images.ubuntu.com/quantal/daily-live/current/ has out-of-date iso and manifest for amd64 | 08:54 |
| infinity | cjwatson: But Steve's in the middle of a respin right now, no? Or someone is... | 08:55 |
| infinity | Oh, FFS, why didn't we disable the precise dailies? | 08:55 |
| cjwatson | infinity: We'll have to go again for this SB thing anyway :-( | 08:56 |
| cjwatson | FED UP | 08:56 |
| stgraber | bug 1067659 | 08:56 |
| ubot2 | Launchpad bug 1067659 in ubiquity "ubiquity removes linux-signed during installation cleanup preventing (some) secureboot machines from booting" [Critical,Triaged] https://launchpad.net/bugs/1067659 | 08:56 |
| infinity | cjwatson: Right, I'm disabling precise dailies in cron. | 08:56 |
| cjwatson | k | 08:57 |
| infinity | cjwatson: If our builder goes idle, I can respin Chinese and see if it's actually broken, so we're confident before the next mass-respin. | 08:58 |
| apw | cjwatson, there is a contention this samsung efi brickage is happening as the kernel is booting, if that were the case, could we use hwmatch to detect the machine and prevent boot | 08:59 |
| cjwatson | stgraber: could I get /var/log/installer/debug too? | 08:59 |
| cjwatson | apw: er, in theory, but we're pretty far into untested code | 08:59 |
| cjwatson | apw: has it been demonstrated to hit 12.10? | 08:59 |
| infinity | cjwatson: I'd give good money to have a machine to test this theory on. :/ | 08:59 |
| apw | cjwatson, understood indeed, 12.10> unknown here, i think slangasek probabally knows | 09:00 |
| cjwatson | also, the signed image doesn't currently have hwmatch | 09:00 |
| cjwatson | slangasek doesn't | 09:00 |
| cjwatson | or didn't last I spoke to him about it | 09:00 |
| cjwatson | there's an unproven theory that the partition table rearrangement on our images might have got rid of it | 09:00 |
| apw | cjwatson, ok thats interesting | 09:01 |
| cjwatson | apw: but this could be utter BS | 09:01 |
| apw | cjwatson, yeah this is a bit of a lack of information issue | 09:02 |
| cjwatson | stgraber: I've found one part of the bug, but hopefully installer/debug will let me track down the other part | 09:02 |
| stgraber | cjwatson: sure, one sec | 09:04 |
| stgraber | cjwatson: http://paste.ubuntu.com/1284645 | 09:04 |
| * cjwatson accumulates browser tabs rapidly | 09:05 | |
| * stgraber back in 5min | 09:06 | |
| cjwatson | grah, this is inscrutable | 09:07 |
| cjwatson | stgraber: could I get a tarball of /var/lib/ubiquity/ ? | 09:08 |
| cjwatson | couple of other files there I want to check | 09:08 |
| stgraber | cjwatson: should be at http://www.stgraber.org/download/ubiquity-lib-sb.tar.gz | 09:14 |
| cjwatson | right, so not being explicitly removed by remove_unusable_kernels at least | 09:17 |
| cjwatson | Oh, I see, maybe. We do_install the kernel but we don't mark it to keep | 09:20 |
| stgraber | cjwatson: I'm not sure what bits you've looked at so far, but in plugininstall.py there's a check for linux-generic- which won't match a signed kernel | 09:21 |
| cjwatson | Yeah, paste above. But I'm not going to attempt to fix that, it controls something else, I think it probably isn't relevant here, and that's quite a risky change | 09:23 |
| cjwatson | stgraber: could you test http://paste.ubuntu.com/1284670/ ? | 09:23 |
| stgraber | sure | 09:23 |
| cjwatson | I suggest we punt this fix into -updates so that we don't have to rebuild everything again | 09:26 |
| stgraber | yay for installing from and to SSDs ;) I can now get a full desktop install in 2 minutes or so | 09:26 |
| stgraber | cjwatson: ok, with your change I now have the signed kernel installed | 09:28 |
| cjwatson | Can I double-check syslog and debug? | 09:28 |
| stgraber | cjwatson: hmm, default grub entry is unsigned... | 09:29 |
| stgraber | cjwatson: and I don't have any .signed file in /boot... | 09:30 |
| stgraber | posting the logs now | 09:30 |
| stgraber | cjwatson: http://paste.ubuntu.com/1284675 <- debug | 09:31 |
| stgraber | cjwatson: http://paste.ubuntu.com/1284676 <- syslog | 09:31 |
| stgraber | I don't think I cleaned those two before installing, so you may have to scroll down a bit | 09:31 |
| stgraber | if you want some perfectly clean ones I'm happy to flush /var/log /var/log/installer and /var/lib/ubiquity again and re-install | 09:32 |
| jibel | so, I get a grub prompt after installation to an external drive | 09:35 |
| cjwatson | stgraber: hmm, the logs indicate that the default grub entry is signed - I think a fresh boot would be helpful to clear the noise | 09:36 |
| cjwatson | jibel: hm, apw tested that yesterday | 09:36 |
| cjwatson | jibel: sure your BIOS is booting from the external drive, and not from a vestigial GRUB somewhere else? | 09:37 |
| jibel | cjwatson, ah maybe it switched to the wrong device when I removed the stick, checking | 09:37 |
| cjwatson | stgraber: and I do mean fresh boot rather than just trying to clear out the state - since we're short on time I'd like to be as sure as possible that there's no pollution | 09:38 |
| stgraber | cjwatson: yep. just running debsums in /target first, will do a reboot and clean run after that | 09:38 |
| stgraber | cjwatson: oh, the debsums result is interesting. let me paste that to you | 09:38 |
| stgraber | cjwatson: http://paste.ubuntu.com/1284692 | 09:39 |
| cjwatson | stgraber: curious, the logs indicate it was there at some point | 09:41 |
| cjwatson | oh, I'm misreading the log, that's it scanning a grub.cfg elsewhere | 09:41 |
| cjwatson | you have another installation on /dev/sda2 but are installing this somewhere else? | 09:42 |
| stgraber | right, /dev/sda2 is my main install on the internal ssd | 09:42 |
| stgraber | I'm booting from /dev/sdd and installing to /dev/sdb | 09:42 |
| cjwatson | still pretty confused - linux-signed-generic is in the live seed so it should have copied that and not removed it | 09:43 |
| stgraber | maybe /rofs is broken too, I forgot to check that | 09:43 |
| * cjwatson checks the special kernel copying code | 09:44 | |
| cjwatson | hm, there is at least *something* odd here, so don't restart just yet | 09:45 |
| stgraber | too late, I restarted minutes ago. Though test runs are pretty quick so I can replicate in 5min or so | 09:45 |
| cjwatson | ok | 09:45 |
| stgraber | /boot in the squashfs doesn't actually contain any kernel | 09:45 |
| cjwatson | yeah, it copies it off the iso9660 filesystem | 09:45 |
| jibel | no way. no external drive: rescue mode / internal drive first: grub prompt / external drive first: black screen with blinking cursor at the top left | 09:46 |
| cjwatson | but it probably isn't getting it right for signed kernels - that's just what I was thinking here | 09:46 |
| stgraber | ok, so that's the problem, because the iso9660 filesystem only contains a single signed kernel | 09:46 |
| cjwatson | apw: could you dig into jibel's problem? | 09:46 |
| cjwatson | stgraber: right, although the patch I gave you earlier is needed too | 09:46 |
| jibel | apw, my setup is a laptop with internal SSD, and external SATA drive on USB | 09:47 |
| stgraber | yep, we have two problems :) | 09:47 |
| stgraber | and that means that any system installed over the past few days has the wrong kernel binary as only the unsigned kernel is installed but the signed binary is in /boot | 09:47 |
| stgraber | I'd expect debsums to report a mismatch on /boot/vmlinuz-* on any 12.10 system installed since the last kernel update | 09:47 |
| cjwatson | Yeah, for the three people who've installed using SB | 09:48 |
| cjwatson | So I'm going to ignore that temporarily :-) | 09:49 |
| cjwatson | It'll get sorted out on upgrade | 09:49 |
| cjwatson | I'll focus on the installer bits | 09:49 |
| stgraber | hmm, you sure? My understanding is that vmlinuz is the signed one for everyone on the media | 09:49 |
| stgraber | so if ubiquity copies the kernel from the iso9660, then non-secureboot systems installed over the past few days have a signed vmlinuz in /boot | 09:50 |
| stgraber | (not that it'd be a big problem and as you said, it'll fix itself with the next kernel update) | 09:50 |
| cjwatson | Oh | 09:50 |
| cjwatson | Yeah, true, but ... that | 09:51 |
| stgraber | cjwatson: still want me to run another test install now? | 09:53 |
| * stgraber needs to find something better than his phone as an install media. Android gets confused and decides not to export the SD card requiring a reburn between every boot... | 09:54 | |
| cjwatson | stgraber: I think I'll want *a* test, but trying to figure out what | 09:57 |
| cjwatson | Unfortunately since it requires at least one CD modification it's very hard to test it in place | 10:02 |
| cjwatson | stgraber: OK, could I just get clean syslog and debug from the previous patch I gave you so I can check there isn't anything else, and then I'll run a possible approach by you? | 10:03 |
| === mmrazik is now known as mmrazik|afk | ||
| stgraber | cjwatson: ok, running an install now | 10:03 |
| stgraber | cjwatson: http://paste.ubuntu.com/1284742 | 10:08 |
| stgraber | cjwatson: http://paste.ubuntu.com/1284743 | 10:08 |
| cjwatson | stgraber: So, my thought is to put both unsigned and signed kernels in the iso9660 filesystem, call the signed kernel vmlinuz.efi.signed instead of vmlinuz, adjust grub.cfg on the image, and use http://paste.ubuntu.com/1284741/ | 10:08 |
| cjwatson | stgraber: Thanks, so all the other bits of this do look better | 10:10 |
| stgraber | cjwatson: plan sounds good. That'll make the cd 5MB larger but it's not like we have an alternative anyway. | 10:11 |
| stgraber | well, the alternative would be to only ship the signed kernel, drop unsigned completely, seed the signed kernel everywhere and change the path in ubiquity, but that's not the kind of thing we want to do a day before release :) | 10:12 |
| cjwatson | Quite. We have clearance under 800MB - *barely* | 10:14 |
| cjwatson | As in, 48KB | 10:14 |
| stgraber | cjwatson: I'll apply your patch and make the system think I have two kernels in /rofs/casper so I can do a test run with that code before we respin | 10:14 |
| cjwatson | But that's a self-imposed limit so given the circumstances we could simply waive it | 10:14 |
| infinity | Do we care that Chinese images are way over our arbitrary desktop size limit, or should I just give Chinese an extra 50MB or something for localisation breathing room? | 10:15 |
| stgraber | (or rather /cdrom) | 10:15 |
| cjwatson | stgraber: Thanks. FWIW I think we want http://paste.ubuntu.com/1284752/ in livecd-rootfs and http://paste.ubuntu.com/1284754/ in debian-cd | 10:15 |
| cjwatson | infinity: We probably ought to choose not to care | 10:15 |
| cjwatson | And give it an extra chunk | 10:15 |
| infinity | cjwatson: Yeah, I'll give it another 50MB in the publishing script. | 10:15 |
| stgraber | infinity: I gave them an extra 50MB for 12.04.1, so I guess we can stick to their limit being Ubuntu's + 50MB | 10:16 |
| stgraber | cjwatson: diffs look good | 10:17 |
| cjwatson | stgraber: ubuntu-defaults-image is wrong too | 10:18 |
| cjwatson | http://paste.ubuntu.com/1284763/ | 10:18 |
| cjwatson | Now, I have no idea whether ubuntu-defaults-builder will get SB right - it probably won't | 10:19 |
| cjwatson | But that's not RC, we have no time to fix it properly, and this at least ensures that it doesn't copy MD5-mismatching kernels | 10:20 |
| didrocks | so… I've noticed an issue during upgrade from precise -> quantal in compiz if you use the "gnome classic session" | 10:21 |
| didrocks | basically, you have no window mananger | 10:21 |
| didrocks | the issue is that if you start with this session, it will be a bad hack in the SRU to get window manager back (and nothing for those users in the 7 days until it's published into -updates) | 10:21 |
| cjwatson | stgraber: re size - for R hopefully we'll be able to just ship a detached signature and construct at run-time | 10:22 |
| didrocks | seeing the number of respin, is there anything plan? is there a change I can move the fix to finale? | 10:22 |
| cjwatson | any SRU would have to cope with people who'd upgraded to quantal during development anyway | 10:23 |
| cjwatson | so I don't see any gain from trying to cram that into images vs. a 0-day SRU | 10:23 |
| cjwatson | we can waive the waiting period | 10:23 |
| didrocks | cjwatson: well 0-day SRU isn't in -updates for the 0 day? it will be in 0 + 7 days, right? | 10:23 |
| didrocks | ah | 10:23 |
| didrocks | last time I asked, it wasn't possible :) | 10:23 |
| cjwatson | see "waive the waiting period", yes :) | 10:23 |
| stgraber | cjwatson: test install running | 10:24 |
| cjwatson | meh, it's *possible*, it's a decision whether we allow it | 10:24 |
| didrocks | ok, great, so let me do an upload with only that :) | 10:24 |
| didrocks | thanks cjwatson | 10:24 |
| cjwatson | and this sounds severe enough to make sense | 10:24 |
| didrocks | cjwatson: I agree, that's why I rejected the SRU0 for compiz snapshot above I did this morning ^ | 10:24 |
| didrocks | cjwatson: I'm doing an upload with just this fix, don't want to add the rest of the snapshot which needs a normal testing period of time | 10:24 |
| didrocks | thanks again, preparing that | 10:24 |
| cjwatson | Makes sense | 10:25 |
| infinity | didrocks: We've been trying to empower people to make intelligent and informed decisions about waiting periods, rather than just following the letter of the law. In some cases, fast-tracking things is definitely the right thing to do. | 10:25 |
| jibel | apw, ping ? | 10:26 |
| didrocks | infinity: perfectly fine with this one, it impacts upgrades and people installing gnome-session-fallback, so they do need an internet connexion anyway | 10:26 |
| stgraber | cjwatson: it looks like we're good. I have both kernels in /target/boot and grub.cfg points to .efi.signed | 10:27 |
| stgraber | cjwatson: and the unsigned kernel is my test file with "unsigned" written in it, so it copied the right files at the right place | 10:28 |
| stgraber | cjwatson: will paste the install logs now | 10:28 |
| stgraber | cjwatson: http://paste.ubuntu.com/1284771 | 10:28 |
| stgraber | cjwatson: http://paste.ubuntu.com/1284772 | 10:28 |
| stgraber | if that's all you need from the live environment, I'll reboot for a test boot just to be 100% sure that it worked | 10:29 |
| cjwatson | That's all I need, thanks, looks good | 10:30 |
| stgraber | and target boots fine! | 10:31 |
| cjwatson | stgraber: Brilliant - I'll start getting things uploaded | 10:32 |
| cjwatson | I think what makes sense is ubiquity -> proposed, livecd-rootfs -> release, ubuntu-defaults-builder -> release | 10:33 |
| cjwatson | no gain in putting the last two through proposed | 10:33 |
| stgraber | indeed | 10:33 |
| stgraber | and then have ubiquity copied to updates instead of release | 10:33 |
| cjwatson | Right | 10:34 |
| stgraber | btw, are the server seed changes done already? if so, I'd like to have it respun so I can confirm the fix | 10:34 |
| cjwatson | ? | 10:35 |
| cjwatson | oh, the preseed file | 10:35 |
| cjwatson | Yeah | 10:35 |
| stgraber | gah, not seed, preseed in debian-cd... | 10:35 |
| stgraber | I updated the pad for desktop images and server | 10:39 |
| cjwatson | thanks | 10:39 |
| cjwatson | respinning server now | 10:39 |
| stgraber | ok, marking the relevant images as rebuilding on the tracker | 10:40 |
| cjwatson | In fact, that can be iso9660-only, don't need to respin the livefs | 10:40 |
| cjwatson | right? | 10:40 |
| stgraber | indeed, in theory all we need is /preseed, so iso9660 should be all we need | 10:40 |
| stgraber | I went on the assumption that preinstalled and wubi images aren't affected by the ubiquity change. If I'm wrong, let me know and I'll update the tracker | 10:41 |
| cjwatson | shouldn't be, no, not when we're running them through -updates | 10:42 |
| === mmrazik|afk is now known as mmrazik | ||
| infinity | Hrm? Updated ubiquity affects nearly everything, from the POV of packages installed... | 10:43 |
| infinity | Oh, except for wubi. | 10:43 |
| infinity | But preinstalled, yes, since it uses oem-config. | 10:43 |
| stgraber | infinity: we're copying it directly to -updates so it won't cause the package lists to be out of date for alternate/preinstalled | 10:44 |
| cjwatson | exactly | 10:45 |
| cjwatson | specifically to avoid having to respin world *again* | 10:45 |
| cjwatson | we could have done that in the previous respin as well | 10:45 |
| infinity | Ahh. | 10:47 |
| cjwatson | ubuntu-defaults-builder on its way too, a few minutes behind | 10:51 |
| cjwatson | I guess I should only have respun server for amd64 | 10:51 |
| cjwatson | oops | 10:51 |
| * stgraber goes to review ubiquity and livecd-rootfs | 10:52 | |
| cjwatson | apw: how goes the investigation of jibel's bug? | 10:54 |
| apw | jibel, hey | 10:54 |
| jibel | apw, hey, I tried an installation to an external drive, no way to boot after installation | 10:56 |
| jibel | apw, I filed bug 1067691 | 10:56 |
| ubot2 | Launchpad bug 1067691 in ubiquity "Installation to external drive failed" [Undecided,New] https://launchpad.net/bugs/1067691 | 10:56 |
| apw | jibel, will try and repro shortly | 10:56 |
| jibel | apw, did I do something wrong or what info do you need to investigate this ? | 10:56 |
| apw | i want to do a quick install with a respun image to confirm if i hit the same | 10:57 |
| stgraber | ok, I'm bored, diffing manually... | 10:58 |
| stgraber | and that's the three of them accepted | 11:02 |
| stgraber | finishing lunch and I'll do a server test | 11:02 |
| stgraber | hopefully we'll finally be good with that SB mess | 11:03 |
| cjwatson | Hm, I need to fix cdimage to actually download .efi.signed too | 11:07 |
| cjwatson | OK, done, deployed | 11:15 |
| stgraber | server image written to usb stick, will be out for the next 15min for testing | 11:15 |
| stgraber | cjwatson: success! server install worked and booted fine. I checked grub.cfg, the checksums of the kernels and the package list post-install, all looks good AFAICS. | 11:22 |
| cjwatson | cautiously muted yay | 11:22 |
| stgraber | looks like something in d-i tries to look for linux-headers-signed-generic and fails to find it (obviously) but that's not causing anything more than a syslog entry | 11:23 |
| cjwatson | It might fail to install linux-headers-generic | 11:23 |
| cjwatson | But if so that's a release note at this point | 11:23 |
| stgraber | nope, linux-headers-generic is there as it's a dependency of linux-signed-generic | 11:25 |
| cjwatson | ok, great | 11:25 |
| cjwatson | waiting for ubiquity/arm* to finish - I suspect it'll just miss the next publisher run | 11:26 |
| cjwatson | apw: anything? | 11:26 |
| cjwatson | really want to know if this is another ubiquity upload, a grub2 upload, or what ... | 11:27 |
| apw | cjwatson, working ... to repro his issue i have had to zap the machine; its mid install right now | 11:27 |
| cjwatson | has anyone managed to do any in-place analysis? | 11:27 |
| apw | cjwatson, now mine is in progress i'll engage him | 11:28 |
| * stgraber starts smoke testing Edubuntu while waiting for the mass rebuild | 11:29 | |
| cjwatson | Looks like sda => SSD, sdb => installation medium, sdc => target | 11:29 |
| cjwatson | purely from disk sizes | 11:29 |
| cjwatson | grub-install was run on sda | 11:30 |
| cjwatson | Oct 17 09:24:10 ubiquity: No active iterator for grub device entry. | 11:31 |
| cjwatson | debconf (developer): <-- SET grub-installer/bootdev /dev/sda | 11:31 |
| cjwatson | stgraber: hmm, get_grub_choice is called here before selecting an autopartitioning option | 11:32 |
| stgraber | seriously? | 11:32 |
| cjwatson | presumably from maybe_update_grub | 11:33 |
| cjwatson | hum, no, that's just manual partitioning | 11:33 |
| cjwatson | maybe I'm misreading | 11:33 |
| cjwatson | stgraber: sorry, I know you've been very busy this morning, but this seems to touch on your recent changes | 11:34 |
| stgraber | yeah... the fact that it's showing "No active iterator" shows that it's at least entering the right part of the if statement, so now the question is why self.get_autopartition_choice doesn't return the right thing... | 11:36 |
| stgraber | jibel: how easily can you reproduce this? I can provide you with a patched ubi-partman.py that'll print a whole bunch of stuff into /var/log/installer/debug so we can try and figure out what's going on | 11:36 |
| cjwatson | the debconf debug log indicates that whole disk / sdc was selected | 11:38 |
| apw | cjwatson, so my reading of the logs we have from jibels bug ... it looks exactly like it did before any fixes | 11:39 |
| cjwatson | it's ubiquity 2.12.14 | 11:39 |
| jamespage | stgraber: hrm - really sorry but I've found another regression in open-iscsi - bug 1066945 | 11:40 |
| ubot2 | Launchpad bug 1066945 in open-iscsi "iSCSI root fails" [Critical,In progress] https://launchpad.net/bugs/1066945 | 11:40 |
| jamespage | on a more positive note I did just test a fix and it looks OK | 11:41 |
| cjwatson | stgraber: hmm | 11:41 |
| cjwatson | stgraber: I'm pretty sure disk_string is /var/lib/partman/devices/=dev=sdc, based on how partman-auto/select_disk is being preseeded | 11:42 |
| cjwatson | stgraber: what was the disk_string.split(None, 3) business based on? | 11:44 |
| stgraber | cjwatson: hmm, fun, makes me wonder why the format changes randomly then... | 11:44 |
| stgraber | cjwatson: it's supposed to be "SCSI3 (0,0,0) (sda) some fancy string" | 11:44 |
| cjwatson | *blink* no it's not | 11:44 |
| cjwatson | that's the translated one | 11:44 |
| cjwatson | get_autopartition_choices should be giving you back the untranslated text, whch is the partman devices directory | 11:45 |
| cjwatson | consistently | 11:45 |
| stgraber | well, I certainly was getting the translated one on my tests here as I made sure I was calling grub_default with the right parameter here... | 11:46 |
| stgraber | get_current_disk_partman_id in the Gtk frontend was giving me =dev=sda but there was no equivalent in the Kde frontend so that's why xnox suggested using get_autopartition_choices instead | 11:47 |
| * stgraber starts a VM to confirm he's not insane and indeed still gets the translated string out of get_autopartition_choices | 11:47 | |
| cjwatson | gtk or kde? | 11:47 |
| stgraber | I tested both | 11:47 |
| stgraber | if someone wants the debug version of the code, I'm usually running with: http://paste.ubuntu.com/1284860/ | 11:48 |
| cjwatson | in any case you can't legitimately do any machine processing on the translated form | 11:48 |
| cjwatson | it could be completely different in different languages, never mind anything else | 11:49 |
| cjwatson | hm, ok, looks like you're meant to process through self.extra_options['use_device'][1][val][0] | 11:49 |
| cjwatson | (obviously ...) | 11:50 |
| cjwatson | as in get_current_disk_partman_id | 11:50 |
| cjwatson | I think you ought to add an equivalent of get_current_disk_partman_id to ubiquity/frontend/kde_components/partAuto.py | 11:51 |
| apw | jibel, i assume you install in .fr ? | 11:52 |
| stgraber | cjwatson: is get_current_disk_partman_id().replace("=", "/") safe or do I need to copy/paste the extra_options logic and extract something better manually? | 11:52 |
| cjwatson | and then something like dev, _ = whatever.get_current_disk_partman_id(); dev = dev.replace("=", "/") | 11:52 |
| cjwatson | oh, ..._partman_id only gives you the dev anyway | 11:52 |
| cjwatson | = => / is safe | 11:53 |
| stgraber | ok, I'll do that then | 11:53 |
| cjwatson | apw: log says yes | 11:53 |
| apw | cjwatson, ok ... is your analysis clear, or do you want me to test both langs ? | 11:54 |
| apw | cjwatson, my .en install seems to have used sdc correctly | 11:54 |
| stgraber | apw, jibel: can you try http://paste.ubuntu.com/1284877/ while I figure out how to do it in the Kde frontend? | 11:58 |
| infinity | Accepting the two-line change in thunderbird, we'll see if it builds in time to get into an image and not be an SRU. | 11:58 |
| infinity | (Though I don't have high hopes) | 11:58 |
| stgraber | cjwatson: does ^ look sane to you? | 11:58 |
| cjwatson | apw: I'm not totally convinced by this analysis in that I can't see *exactly* how it would result in this failure, but it looks like a good bet. It'd be good if you could try in fr to make sure that we have >1 person who can reproduce this, though | 11:58 |
| cjwatson | stgraber: not entirely sure what the isinstance achieves now but I guess it doesn't hurt | 11:59 |
| cjwatson | stgraber: yes, that looks right | 11:59 |
| stgraber | cjwatson: yeah, it'd only be useful if the list entry contains garbage though that'd likely break the rest of the install anyway :) | 12:01 |
| apw | cjwatson, ack | 12:03 |
| cjwatson | jamespage: looks good - can you please add it to the pad (URL in topic)? | 12:07 |
| stgraber | cjwatson: http://paste.ubuntu.com/1284888/ is with the KDE code. Grabbing kubuntu-desktop now to check that it actually works | 12:08 |
| * jibel back from lunch | 12:08 | |
| jibel | stgraber, which paste should I test ? | 12:08 |
| cjwatson | stgraber: I'd probably have put the KDE function in partAuto.py, but meh, do what you have :) | 12:08 |
| stgraber | jibel: http://paste.ubuntu.com/1284877/ | 12:09 |
| jibel | stgraber, ack. I'll reinstall the machine first | 12:09 |
| jamespage | cjwatson, any particular section? | 12:10 |
| stgraber | cjwatson: yeah, I did that initially then I noticed how close the Kde and Gtk functions ended up being so I moved it back to the same file so they're less likely to get out of sync if we ever have to touch them again | 12:10 |
| cjwatson | jamespage: rebuild triggers | 12:10 |
| jamespage | cjwatson, ack | 12:11 |
| jamespage | cjwatson, done - I've linked it to the server ISO - not sure which other images include it. | 12:11 |
| Laney | kubuntu | 12:12 |
| Laney | (per seeded-in-ubuntu) | 12:12 |
| cjwatson | no | 12:12 |
| cjwatson | That's the Kubuntu DVD which isn't built any more | 12:12 |
| cjwatson | Haven't got round to figuring out seeded-in-ubuntu's data sources ... | 12:13 |
| Laney | wh | 12:13 |
| Laney | right ... | 12:13 |
| Laney | maybe deleting the old image would do it | 12:14 |
| nigelb | 20 | 12:14 |
| nigelb | (gah, sorry) | 12:14 |
| cjwatson | Laney: oh, yeah, that's still there - gone now | 12:14 |
| tumbleweed | cjwatson: re seeded-in-ubuntu sources: all .manifest and .lists on cdimage.u.c | 12:21 |
| cjwatson | OK, so deleting the image will sort it out | 12:23 |
| cjwatson | thanks | 12:23 |
| stgraber | cjwatson: code briefly tested on kubuntu, properly selected /dev/sdc here | 12:24 |
| stgraber | I'll give a quick go at manual partitioning to make sure we don't get into the same bug as was found overnight | 12:24 |
| xnox | =)))) | 12:25 |
| xnox | your paste looks better now. | 12:25 |
| stgraber | manual seems good on kde, I can enter the dialog create partitions and continue, so it looks good | 12:27 |
| stgraber | and according to the debug log, grub would have installed on the right device | 12:27 |
| * stgraber cleans up the ubiquity branch and prepares for upload | 12:28 | |
| stgraber | just waiting for confirmation from jibel and apw that the Gtk frontend works fine too | 12:28 |
| jibel | apw, installing in French or English doesn't make a difference | 12:28 |
| highvoltage | stgraber: good morning. do we have an idea when the edubuntu build will be ready for testing? | 12:29 |
| stgraber | highvoltage: I'm hoping to have ubiquity published in the next hour (for the 13:32 UTC publisher run), so Edubuntu should be built 3-4h after that, depending on the ordering | 12:30 |
| stgraber | highvoltage: I'd strongly suggest testing the current image for now | 12:30 |
| stgraber | highvoltage: and doing the upgrade testing | 12:30 |
| cjwatson | stgraber: You'll be lucky | 12:30 |
| cjwatson | Given we need two publisher runs | 12:30 |
| stgraber | oh, right, well, let's add 30min to that estimate then ;) | 12:31 |
| cjwatson | 14:03 UTC is more likely, yes | 12:31 |
| * skaet waves hello | 12:32 | |
| cjwatson | Hello. Still not done :-( | 12:35 |
| highvoltage | stgraber: ok | 12:36 |
| * stgraber updated the pad | 12:37 | |
| stgraber | cjwatson: I uploaded ubiquity to the queue. I'd recommend we let it build while we're waiting for test results from jibel and apw | 12:42 |
| skaet | cjwatson, so I see. Thanks for all the hard work release team's been putting in. Fingers crossed this is the last. | 12:45 |
| stgraber | if it's not I'm not sure how we'll get that thing out tomorrow, so yeah, better work fine after that one... | 12:47 |
| jbicha | good morning, what determines whether something is a 0-day SRU or if it needs to wait the traditional 7 days? (I'm asking about eds in particular) | 12:48 |
| skaet | jbicha, function of risk, scope of impact, level of testing done. | 12:50 |
| cjwatson | stgraber: reviewing | 12:56 |
| highvoltage | the upgrade notes when upgrading using update-manager still says that quantal is a beta release, when does that get removed? | 13:01 |
| jibel | stgraber, grub prompt on reboot, checking boot order now. | 13:02 |
| gema | jibel: where is / in the french keyboard? | 13:02 |
| jibel | gema, https://www.google.fr/search?client=ubuntu&channel=fs&q=french+keyboard+layout | 13:03 |
| jibel | stgraber, same result as before, now checking that I correctly patched ubiquity | 13:05 |
| stgraber | jibel: can you pastebin /var/log/installer/debug? the patch I sent you included debug statements so that should help quite a lot | 13:06 |
| apw | cjwatson, stgraber, well i did the install (without fix) in .fr and it still installed grub to /dev/sdc | 13:08 |
| apw | cjwatson, so i guess i am not a good test of the issue, i can of course test regression wise | 13:08 |
| jibel | stgraber, http://people.canonical.com/~j-lallement/debug_1067691-2.12.14-patched | 13:10 |
| stgraber | DEBUG: calling grub_default with disk_path=/dev/sdc | 13:10 |
| stgraber | DEBUG: == END get_grub_choice == | 13:10 |
| stgraber | debconf (developer): <-- SET grub-installer/bootdev /dev/sda | 13:10 |
| xnox | because both are internal disks. | 13:11 |
| xnox | ?! | 13:11 |
| stgraber | so my code did the right thing... | 13:11 |
| stgraber | yeah, that's my guess, /dev/sdc probably isn't marked as removable | 13:11 |
| jibel | is this machine special? | 13:11 |
| xnox | jibel: is /dev/sdc actually a USB device or interal disk? | 13:11 |
| stgraber | jibel: what's the content of /sys/class/block/sdc/removable ? | 13:11 |
| xnox | jibel: no, not machine special, but whether it's "internal" HDD (regular HDD in the VM) or USB attached "external" (in a VM or Real HW) | 13:12 |
| cjwatson | OK, so the case of non-removable devices is one that I think is in principle a bug but too scary to fix in 12.10 | 13:12 |
| jibel | xnox, as far as I can see, sdc is a little grey box, with a rotational drive inside, and a cable marked USB, connected to a USB port on the laptop | 13:13 |
| cjwatson | Because the other case that's more or less indistinguishable is a machine with two hard disks, somebody installs Ubuntu to the second, and expects it to automatically boot | 13:13 |
| xnox | Most computers regardless of amount of disks only boot from the first hard drive by default, unless person changes the boot order & disks in the bios. | 13:13 |
| cjwatson | We've had bugs about that in the past | 13:13 |
| xnox | jibel: ha =)))) | 13:13 |
| cjwatson | I'd like to change that because it'd be a massive simplification | 13:13 |
| cjwatson | But not for 12.10 | 13:13 |
| * skaet nods | 13:13 | |
| cjwatson | However, 14:11 <stgraber> jibel: what's the content of /sys/class/block/sdc/removable ? | 13:13 |
| cjwatson | should tell whether it's this bug | 13:13 |
| * xnox *suspense* | 13:14 | |
| stgraber | if the answer is 0, we'll be sure it's that bug, if it's not, there are some more udev-related things that could go wrong with some specific drives... | 13:14 |
| jibel | 1 minute, time to reboot from a usb stick | 13:14 |
| jibel | stgraber, cjwatson: 0 | 13:17 |
| stgraber | yay! | 13:17 |
| stgraber | jibel: for fun, can you pastebin: udevadm info -q property -n /dev/sdc? | 13:17 |
| plars | istr at one time there was a way in the installer to select which device should get the bootloader installed to it? I know it got removed on arm because of a funny "Help for GRUB device selection goes here." message in its place | 13:17 |
| stgraber | plars: you only have that option in manual partitioning | 13:18 |
| xnox | plars: there is no grub on arm..... and only one place to install boot loader there. | 13:18 |
| * xnox likes ARM now | 13:18 | |
| cjwatson | xnox: I think he knows that :) | 13:18 |
| jibel | pastebinit should be seeded | 13:18 |
| stgraber | jibel: it's seeded on xubuntu and edubuntu ;) | 13:19 |
| jibel | stgraber, http://paste.ubuntu.com/1284997/ | 13:19 |
| stgraber | jibel: and has been promoted to main, so once I release the python3 port and get rid of configobj, we probably could ship it by default in 13.04 as it'll just be an extra (uncompressed) 13K :) | 13:19 |
| plars | yeah, just that maybe that would be a good option to have (with a sensible default of the first HD if we have trouble telling) regardless of manual partitioning or automatic. And yes, I'm quite aware that we don't use grub on arm :) | 13:19 |
| stgraber | jibel: ID_BUS=ata | 13:19 |
| plars | but yes, not for 12.10 | 13:20 |
| cjwatson | plars: I would tend to agree that it should be offered in the autopartitioning case if there are multiple disks | 13:20 |
| stgraber | jibel: that's your problem. As far as the machine is concerned, it's directly attached to a SATA bus | 13:20 |
| cjwatson | I haven't always been able to persuade design of this kind of thing :) | 13:20 |
| xnox | plars: sorry. =) | 13:20 |
| cjwatson | So - I think we should still take .16 if it passes regression-testing, because I'm a lot more confident that this approach is robust across languages | 13:20 |
| cjwatson | FWIW, since we originally started trying to prefer the first disk, external disks have become a *lot* more common | 13:22 |
| cjwatson | So I don't think the old approach was wrong in context, but the balance seems to have shifted | 13:22 |
| plars | infinity: I haven't tried the latest -server builds from this morning, but the one I grabed last night should have had the fix for this keyboard bug I think and I still don't have a working keyboard for inputting the password to unlock partitions | 13:23 |
| stgraber | jibel: do you have a brand/model number for that USB disk? I'm wondering how it's that the kernel thinks it's a standard sata drive, maybe it has some kind of fancy controller embedded that does that | 13:23 |
| plars | is a rebuild running right now for -desktop isos? and if so, what all is included? 57-60? | 13:30 |
| jibel | stgraber, I'll have to open the enclosure, the bios sees it as USB:ST925031 5AS | 13:30 |
| jibel | I've no clue what it is, the drive is a western digital or hitachi something | 13:31 |
| stgraber | plars: no rebuild running at the moment, we're waiting for some more regression tests of the last ubiquity upload | 13:31 |
| skaet | plars, waiting for ubiquity to build, then those will be included in the rebuild is my understanding | 13:31 |
| === henrix is now known as henrix_ | ||
| plars | skaet, stgraber: so is that why the -desktop ones are already crossed out on iso tracker? I'm not certain what triggers that condition, or if they are manually marked obsolete | 13:32 |
| stgraber | plars: I marked them all as needing rebuilding as they currently don't install the right kernel anyway | 13:32 |
| jibel | stgraber, finally the disk is a seagate Momentus 5400.6 250GB | 13:32 |
| plars | stgraber: got it, thanks | 13:33 |
| stgraber | jibel: ok, any kind of brand/model written on the enclosure? | 13:33 |
| stgraber | jibel: I'm suspecting the USB->SATA chip that part of the enclosure to use some weird chip that the kernel doesn't identify as a usb device | 13:33 |
| jibel | stgraber, of course not, it's the cheapest crap on amazon | 13:34 |
| stgraber | hehe, ok ;) | 13:34 |
| jibel | note that I used the shopping lens to find it :) | 13:35 |
| stgraber | jibel: I actually have the same thing with one of my drives here. My phones properly report ID_BUS=usb but my fancy USB3/SATA3 enclosure shows up as ID_BUS=ata | 13:36 |
| stgraber | apw: any idea what the kernel/udev uses to determine whether something should show up as usb or ata when both are clearly connected on the usb bus (as DEVPATH indicates)? | 13:37 |
| cjwatson | OK, ubiquity has built everywhere, will publish starting at 14:03 UTC, at which point we can copy, and it'll publish to updates starting at 14:33 UTC, visible ~14:55 UTC | 13:37 |
| cjwatson | Then we can actually build | 13:38 |
| stgraber | apw: I initially thought it was a usb2 vs usb3 thing but jibel's showing up on usb2 so that's not it. One clear difference though is that my phone doesn't expose any of the ATA flags (features and so on) but my enclosure does, so maybe that's what's used to decide to show as ata instead of usb | 13:39 |
| cjwatson | So I think it's time for me to have a half-hour break | 13:40 |
| skaet | sounds good. | 13:41 |
| === henrix_ is now known as henrix | ||
| === Ursinha is now known as Ursinha-afk | ||
| stgraber | ^ what was that? | 13:55 |
| stgraber | not spotting any obvious difference on the tracker | 13:55 |
| skaet | stgraber, not sure either | 13:55 |
| * skaet not touching things there right now, balloons? | 13:55 | |
| * balloons smiles | 13:56 | |
| balloons | no, not messing with anything | 13:56 |
| infinity | T'wasn't I. | 13:57 |
| infinity | stgraber: Oh, hey, can you make Ubuntu Core/powerpc show up on the tracker? I thought someone had done that, but it looks not there. | 13:58 |
| didrocks | skaet: cjwatson: ^ here is the compiz version for the 0 day SRU I was talking about. Should I just ping you tomorrow so that it's fast-tracked to -updated? (it's been confirmed by seb128, jibel and I to work) | 13:58 |
| stgraber | infinity: checking | 13:59 |
| infinity | didrocks: For fast-tracking, you'll want SRUy people, you can certainly ping me about it. | 13:59 |
| stgraber | infinity: could be missing mapping in post-qa | 13:59 |
| skaet | didrocks, yes, pinging tomorrow is appropriate | 13:59 |
| didrocks | infinity: then ^ (you are still awaken or in London? ;)) | 13:59 |
| infinity | didrocks: I'll do the SRU review and accept now, and if people can do the usual SRU verification bits, that makes my job much easier later. | 14:00 |
| infinity | didrocks: Yeah, I'm in London this week and the next, and going straight to CPH from here. | 14:00 |
| didrocks | infinity: I started to fear for you TBH ;) | 14:00 |
| didrocks | infinity: thanks! | 14:00 |
| didrocks | infinity: sure, I'll get them to comment | 14:00 |
| infinity | didrocks: Well, I do only sleep 3 or 4 hours a night, so whatever. | 14:00 |
| stgraber | infinity: fixed | 14:00 |
| didrocks | infinity: take care still :/ | 14:01 |
| infinity | stgraber: Shiny, thanks. | 14:01 |
| skaet | stgraber, balloons - it was utlemming. all good. new set coming up. | 14:02 |
| * balloons standing by for zsync | 14:03 | |
| ScottK | https://bugs.launchpad.net/~ubuntu-release/+bugs shows no open bugs now. | 14:05 |
| plars | skaet: https://bugs.launchpad.net/ubuntu/+source/initramfs-tools/+bug/1066376 | 14:11 |
| ubot2 | Launchpad bug 1066376 in initramfs-tools "keyboard doesn't work to enter password with panda and encrypted partitions" [Undecided,In progress] | 14:11 |
| * infinity looks. | 14:13 | |
| infinity | plars: It won't be broken for logitech keyboards in the way it was before, so I'm a bit curious. | 14:14 |
| plars | infinity: but remember that I said it was also broken for usbhid generic keyboards also, and still is broken for both | 14:14 |
| infinity | plars: Can I get a copy of the initrd from /boot (don't really want to tear apart a uInitrd). | 14:14 |
| infinity | plars: See, it shouldn't be broken for either now. | 14:15 |
| infinity | plars: So, yeah, dmesg, lsusb, and your current initrd... | 14:15 |
| plars | infinity: yes, I was about to go looking at that, and was also going to pull apart the uInitrd in the boot partition (I suspect I will not find those modules there just as before) | 14:15 |
| plars | infinity: I'm inserting the bits to boot with serial console enabled as we speak | 14:15 |
| infinity | plars: Hrm, when I generate an initrd here, I definitely get the right modules. But I don't have a machine I can try a fresh install on. :/ | 14:16 |
| plars | infinity: I believe you, but I suspect that is never getting installed to the boot partition | 14:17 |
| infinity | plars: Well, that's where the initrd in the boot partition comes from... | 14:17 |
| skaet | infinity, we missed you on the call. Any updates from London? | 14:19 |
| infinity | skaet: Ahh, totally missed the call, was mangling chroots. Nothing really from us here, just waiting on installer bits to trickle through, and trying to hunt information on Samsung bricking. | 14:19 |
| infinity | slangasek: You awake yet? Andy wants you. | 14:20 |
| skaet | yup slangasek's awake | 14:20 |
| slangasek | mmmm rrrrr | 14:20 |
| slangasek | yes | 14:20 |
| cjwatson | Sorry, I missed the call too, was having a break | 14:24 |
| cjwatson | Copied ubiquity to -updates now | 14:24 |
| cjwatson | stgraber: FWIW post-qa is written such that we only need to add mappings for products/image-types, not individual architectures | 14:25 |
| stgraber | cjwatson: right. In this case, it was just a missing product on the tracker. | 14:27 |
| * cjwatson nods | 14:27 | |
| cjwatson | Oh, I forgot, publishing quantal-updates is going to be fast, isn't it | 14:36 |
| cjwatson | Yay, that's helpful | 14:36 |
| infinity | didrocks: compiz accepted to proposed. | 14:36 |
| didrocks | infinity: excellent! will get people to comment :) | 14:36 |
| infinity | cjwatson: Yeah, when no release pockets change, life is good. | 14:37 |
| didrocks | infinity: I need to ping you again tomorrow for moving to -updates? | 14:37 |
| infinity | didrocks: Once it builds, make sure we do the usual sru-verification on the actual built binaries. | 14:37 |
| infinity | didrocks: And, for the sake of me peace of mind for fast-tracking, get more than one person to verify. | 14:37 |
| infinity | didrocks: Then I'll be happy to do a quick turnaround on the release. | 14:37 |
| didrocks | infinity: ok, will get seb and jibel as a victim again :) | 14:37 |
| infinity | s/me peace/my peace/ | 14:37 |
| infinity | didrocks: Perfect. | 14:38 |
| didrocks | thanks ;) | 14:38 |
| cjwatson | So, let's see. I'm just respinning x86 desktops, yes? | 14:38 |
| jdstrand | is the publisher offline? | 14:38 |
| cjwatson | I mean, non-arm | 14:38 |
| cjwatson | jdstrand: very much online | 14:38 |
| jdstrand | hmm | 14:38 |
| cjwatson | jdstrand: what are you waiting for? | 14:38 |
| jdstrand | infinity: do you still have superpowers on seeing why something isn't mirroring to security.u.c | 14:39 |
| jdstrand | cjwatson: python2.5 to mirror | 14:39 |
| cjwatson | jdstrand: suite and version? | 14:39 |
| jdstrand | actually, I guess that doesn't have anything to do with the publisher | 14:39 |
| infinity | jdstrand: Nope, you want IS, if it's actually failing to mirror. We can certainly see if it's on ftpmaster, though. | 14:39 |
| cjwatson | 2.5.2-2ubuntu6.2 published an hour or so ago, if that's what you mean | 14:40 |
| cjwatson | (from publisher logs, not from LP web UI) | 14:40 |
| infinity | cjwatson: Hrm? Why only x86 desktops? | 14:40 |
| jdstrand | cjwatson: yes, that is it | 14:40 |
| jdstrand | LP claims it made it to -updates too | 14:40 |
| cjwatson | In fact, that was the copy to -updates, not the initial publish to -security half an hour before that | 14:40 |
| * jdstrand is tapping his fingers on LP atm | 14:40 | |
| jdstrand | cjwatson, infinity: thanks | 14:41 |
| cjwatson | infinity: None of the bugs we're respinning for affect ARM, do they? | 14:41 |
| === henrix is now known as henrix_ | ||
| cjwatson | (But this is why I'm asking rather than just starting the builds) | 14:41 |
| cjwatson | Oh, we have ARM desktops not just preinstalled | 14:41 |
| cjwatson | OK, then yeah, we should update those | 14:41 |
| * infinity is a bit confused about the "triggers" section. | 14:43 | |
| infinity | Is iscsi a trigger, or pending investigation? If it's a trigger, we kinda failed to accept and build it. :P | 14:43 |
| cjwatson | Er, I accepted it? | 14:43 |
| infinity | Oh, maybe it failed to close the bug. | 14:43 |
| cjwatson | Although I didn't copy it to release or -updates | 14:43 |
| infinity | Ahh, which is why the bug isn't closed. | 14:44 |
| cjwatson | It's only on one image so I think release pocket would be fine | 14:44 |
| cjwatson | ? | 14:44 |
| infinity | Yeah, I'm fine with release. | 14:44 |
| * cjwatson asks #is about mirroring | 14:45 | |
| infinity | cjwatson: If you're copying open-iscsi, I wouldn't mind picking up libunity-webapps too, but that affects ubuntu/edubuntu. | 14:46 |
| infinity | Not that it's a critical fix or anything, just that it's tiny, and obviously not broken. But that makes it a simple 0-day too. | 14:46 |
| infinity | So, whichever. | 14:46 |
| cjwatson | I was going to start desktop before open-iscsi had time to copy | 14:49 |
| infinity | Alright, then screw it. Had I thought about it before now, I would have copied it, so clearly I didn't think it was important. :P | 14:53 |
| === henrix_ is now known as henrix | ||
| cjwatson | Building: Ubuntu desktop, Kubuntu desktop, Xubuntu desktop, Kubuntu Active, Lubuntu desktop, Edubuntu DVD, Ubuntu Studio DVD, and Chinese edition | 14:55 |
| highvoltage | hope chinese edition is in chinese this time :) | 14:58 |
| skaet | Thanks cjwatson | 15:00 |
| * skaet was talking to is | 15:01 | |
| infinity | didrocks: Are these version numbers based on month or something? | 15:02 |
| cjwatson | highvoltage: It wasn't before? | 15:02 |
| infinity | didrocks: Cause I can't sort out how a 7-line diff goes from 3.8.0 to 3.10.0 instead of, say, 3.8.1. :P | 15:02 |
| didrocks | infinity: no, they do use the gnome versionning way of major_version.+2 | 15:03 |
| didrocks | infinity: I agree that it ends up with not using the minor | 15:03 |
| infinity | didrocks: Sure, but shouldn't these post-release fixes use microversions? | 15:03 |
| didrocks | and of course, some lenses don't follow that schema | 15:03 |
| infinity | didrocks: (Not that this is your issue, nor that the version number actually matters, it's just a bit odd) | 15:03 |
| didrocks | infinity: well, my plan is to discuss versionning at UDS with them :) | 15:03 |
| highvoltage | cjwatson: I remember stgraber mentioning problems with that in a previous release | 15:03 |
| didrocks | because yeah, it's troublesome, I agree :) | 15:03 |
| didrocks | they do use odd/even | 15:04 |
| didrocks | for development version (between two releases) | 15:04 |
| didrocks | but the last digit is finally never used | 15:04 |
| cjwatson | highvoltage: OK, I couldn't tell whether you meant something over the last few days | 15:04 |
| highvoltage | cjwatson: ah, sorry for unintentionally raising a flag | 15:05 |
| didrocks | infinity: there is no hurry in the unity stack itself (apart the compiz thingy you already reviewed) if you are busy with other release duty :) | 15:06 |
| phillw | skaet: can you IM me when you have a spare minute, thanks | 15:10 |
| cjwatson | I've copied open-iscsi to release now, slightly belatedly | 15:15 |
| plars | uh, network problems? I just noticed my netboot install failed, but I don't think it's me, I can't get to ports.ubuntu.com or launchpad.net | 15:18 |
| plars | ok, they seem to be back now | 15:18 |
| plars | odd, I could get to other things fine during that time though | 15:18 |
| stgraber | ssh -t user@example.com screen -UDr | 15:20 |
| stgraber | oops :) | 15:20 |
| stgraber | (new phone with fancy default shortcuts) | 15:20 |
| ogra_ | clever shortcut | 15:21 |
| ogra_ | will log you in to all unconfigured debian and ubuntu boxes :) | 15:21 |
| cjwatson | plars: yeah, #is chatter suggests something | 15:22 |
| ScottK | kernel.ubuntu.com vanished for a bit about the same time. | 15:24 |
| barry | yes, there have been some intermittent network problems, but it seems to have cleared up | 15:24 |
| ev | if someone could shepherd that gnat-gps upload through, I'd greatly appreciate it. Of course, if the universe is on fire with release, no worries :) | 15:24 |
| cjwatson | ev: I was waiting to be a little bit more sure that we weren't going to be contended on buildds | 15:24 |
| ev | cjwatson: noted; thanks | 15:25 |
| cjwatson | ev: SRU verification will be a pain without an LP bug in the changelog | 15:25 |
| cjwatson | Any chance of one? | 15:25 |
| ev | cjwatson: sure, reject and I'll re-upload | 15:25 |
| cjwatson | ev: done, thanks - I'm fine with the upload otherwise | 15:26 |
| cjwatson | The buildds actually are fairly busy at the moment with security updates and the like :-/ | 15:27 |
| cjwatson | I'm trying to keep at least one of each architecture free | 15:27 |
| === Ursinha-afk is now known as Ursinha | ||
| roaksoax | n/win 13 | 16:00 |
| xnox | ogra_: cjwatson: i have a fix for bug 1053030 . This also means that you can install on-to remaining non-pre-partitioned free space of the sd-card on panda boards. | 16:09 |
| ubot2 | Launchpad bug 1053030 in ubiquity "highly confusing UI on desktop when installation media is big enough and no external storage is attached" [High,Confirmed] https://launchpad.net/bugs/1053030 | 16:09 |
| cjwatson | I understood that wasn't release-critical? | 16:10 |
| xnox | cjwatson: it does involve pre-seeding 'ubiquity/partman-skip-unmount' to true only when no auto-partitioning options are available. | 16:10 |
| === bjf is now known as bjf[afk] | ||
| xnox | cjwatson: true, it's not critical, but it is a regression from precise, where we did support preinstalled panda on sd-card only without usb storage. | 16:11 |
| cjwatson | I think we're out of time for installer changes that aren't critical | 16:11 |
| cjwatson | We're pushing our luck as it is :-) | 16:11 |
| xnox | ok. | 16:11 |
| xnox | http://paste.ubuntu.com/1285315/ | 16:11 |
| xnox | I'm queuing this up in my r-proposed then ;-) | 16:12 |
| cjwatson | infinity said last UDS that we really needed to encourage Panda users to attach external storage anyway. | 16:12 |
| ogra_ | ++ | 16:12 |
| cjwatson | So I'd regard any fix as cleanup rather than an essential way to address some legitimate use case. | 16:12 |
| * ogra_ will bear the moaning of SD card users in #ubuntu-arm :) | 16:12 | |
| ogra_ | cjwatson, if i have an image where ssh is preinstalled, there was a trick to force fresh key generation on first boot that i cant remember anymore, could you refresh my memory ? | 16:13 |
| infinity | Yeah, I really don't want to encourage people to install to SD, though having the option would be cute and fun, it's totally not critical. | 16:13 |
| cjwatson | ogra_: Not sure I remember such an option at all | 16:14 |
| cjwatson | I remember people talking about it ... | 16:14 |
| xnox | well, this patch also solves the "dd iso to /dev/sda1 & install onto free space on /dev/sda2" - fixes a crash, and unbreaks the previous use-case cjwatson mentioned when ubiquity/skip-partman-umount was introduced. | 16:14 |
| ogra_ | cjwatson, hmm, k | 16:14 |
| ogra_ | i thought we had such a thing in the past somewhere | 16:14 |
| cjwatson | xnox: So that is somewhat more common and definitely reasonable, and a couple of days ago I'd have said we should take it | 16:14 |
| cjwatson | But I don't think I can justify pushing the testing schedule back even further for it now | 16:15 |
| xnox | ack. | 16:15 |
| cjwatson | xnox: Make sure it's on the release notes, because we'll get asked about that | 16:15 |
| xnox | ack. Will do when I'll be back later. | 16:15 |
| cjwatson | And might as well make it clear it's something we realised late, rather than deliberately withdrew support for | 16:16 |
| * cjwatson attempts to prepare for upcoming auto-syncs: I'm arranging for interactive copies to jump the queue so that they don't get starved | 16:16 | |
| cjwatson | This might cause auto-syncs to be somewhat slower, but anyone who cares about auto-sync performance that much is weird :-) | 16:17 |
| skaet | :) | 16:17 |
| plars | infinity: given the awful performance of SD, I'd really like to actively discourage people from installing to SD rather than just not encourage it | 16:18 |
| infinity | plars: Yeah, I don't want to TECHNICALLY discourage (as in break) the option, but I do want to warn people that it's silly. | 16:19 |
| infinity | plars: Because the user experience of running from SD is awful, as you say. | 16:19 |
| cjwatson | Any reason I shouldn't build server images, now that open-iscsi is in? | 16:20 |
| infinity | cjwatson: Nope, that's all we were waiting on. | 16:20 |
| cjwatson | On its way, then, modulo livefs buildd time. | 16:21 |
| plars | cjwatson: did you and others notice that amd64 iso came back "oversized" on this latest build? http://cdimage.ubuntu.com/daily-live/20121017.2/ The size bump was from 758M to 763M | 16:24 |
| cjwatson | plars: Yeah, we'll tweak the size | 16:24 |
| cjwatson | It's very very marginal and was needed to fix secure boot | 16:24 |
| cjwatson | Sorry, I mean we'll tweak the size limit | 16:25 |
| plars | iirc, those were fairly arbitrary now that we're past cd size, just wanted to point it out | 16:25 |
| cjwatson | 800MB was a self-imposed limit anyway; under the circumstances we can waive it | 16:25 |
| plars | right | 16:25 |
| cjwatson | The alternative was to remove a language pack, which I wasn't comfortable with trying to ram in at the same time | 16:25 |
| cjwatson | So yeah, I'll fix the site, but just ignore it for now | 16:26 |
| cjwatson | It's 63488 bytes over | 16:26 |
| infinity | cjwatson: Was Chinese in the set you're building right now too? | 16:28 |
| cjwatson | Yeah | 16:28 |
| infinity | cjwatson: Things never quieted down enough to test-build that, so I'm hoping it doesn't fail. | 16:28 |
| cjwatson | We'll find out. It's near or at the end of the queue, though. | 16:29 |
| cjwatson | plars: fixed | 16:29 |
| cjwatson | I think I should pre-publish desktop images nowish. Any objections? | 16:30 |
| cjwatson | I'd like to give mirrors a chance. | 16:30 |
| infinity | cjwatson: Please do. | 16:31 |
| infinity | cjwatson: The earlier, the gooder. | 16:31 |
| cjwatson | Can do Wubi fs images as wewll. | 16:32 |
| cjwatson | *well | 16:32 |
| * stgraber grabs ubuntu desktop amd64 and re-test on secureboot just to be 100% sure that this all works now | 16:38 | |
| cjwatson | Please! | 16:38 |
| slangasek | I can test here as well, but stgraber's test seems to be more, ah, rigorous in light of his firmware differences | 16:39 |
| slangasek | xnox: bug #1009973> you closed a release notes task saying "made it into 12.04.1"; I don't see any mention of this package in the release notes? | 16:41 |
| ubot2 | Launchpad bug 1009973 in mdadm "SRU upstream bugfix micro point release 3.2.5" [Wishlist,Fix released] https://launchpad.net/bugs/1009973 | 16:41 |
| stgraber | slangasek: hehe, yeah, I guess yours let you boot an unsigned kernel... if anyone wants to figure out why grub won't let me boot unsigned stuff, I'll have the machine with me at UDS :) | 16:41 |
| === henrix is now known as henrix_ | ||
| cjwatson | stgraber: I suspect we may spend some time swearing atit | 16:43 |
| cjwatson | *at it | 16:43 |
| slangasek | stgraber: that is definitely something we should figure out, given that unsigned kernels are supposed to be bootable still | 16:43 |
| cjwatson | stgraber: Are you at least able to tell it to use alternative keys? | 16:44 |
| stgraber | cjwatson: nope, the firmware doesn't have that option... so at this point we need to get it to the archive every time... | 16:45 |
| === henrix_ is now known as henrix | ||
| stgraber | that's unless there's some signed efi binary somewhere that lets me add keys, then I could just use that | 16:45 |
| cjwatson | Not until Matthew's shim mok work lands, I think | 16:46 |
| jbicha | Riddell: http://www.kubuntu.org/news/kubuntu-12.10 ? | 16:46 |
| cjwatson | WTF. Would be nice if Kubuntu stopped releasing early without telling us | 16:47 |
| cjwatson | Seeing as we're still rebuilding Kubuntu images | 16:47 |
| * cjwatson hopes that's a mistake | 16:47 | |
| jbicha | well it's October 18 somewhere in the world ;) | 16:48 |
| cjwatson | Desktop and wubi pre-published | 16:48 |
| cjwatson | Server will wait a bit longer | 16:48 |
| * stgraber reboots to test desktop on sb | 16:49 | |
| stgraber | cjwatson: got grub but kernel wouldn't boot. Before we start to panic I'm going to have a quick look at what's actually in the .iso | 16:52 |
| stgraber | cjwatson: I only see a single vmlinuz in /casper | 16:52 |
| * stgraber checks he has the latest build | 16:52 | |
| cjwatson | aargh | 16:52 |
| stgraber | yep, md5 matches latest from cdimage (20121017.2) | 16:53 |
| slangasek | isn't there only supposed to be a single vmlinuz in /casper, the signed one? | 16:53 |
| cjwatson | No, had to change that | 16:54 |
| slangasek | ah poo | 16:54 |
| slangasek | ok | 16:54 |
| cjwatson | Because the result of that was a checksum validation failure after install | 16:54 |
| * skaet figures Kubuntu just wants to beat OMG!Ubuntu for a change ;) | 16:54 | |
| cjwatson | And a missing signed kernel too | 16:54 |
| ScottK | Unpublished | 16:54 |
| cjwatson | thans | 16:54 |
| cjwatson | +k | 16:54 |
| skaet | thanks | 16:54 |
| Riddell | jbicha: erk | 16:54 |
| Riddell | cjwatson: yes sorry | 16:55 |
| cjwatson | stgraber: syncing and debugging as fast as I can | 16:55 |
| ScottK | Riddell: Fixed already. | 16:55 |
| stgraber | the good news is, we won't have to respin the livefs (I tested the ubiquity side of it faking /cdrom/casper so I'm 99.9% sure the ubiquity change is right) | 16:55 |
| cjwatson | Right, excellent chance I screwed up livecd-rootfs or cdimage | 16:56 |
| cjwatson | Sigh | 16:56 |
| slangasek | cjwatson: let me ask the necessary question: is this worth a respin? | 16:57 |
| slangasek | we do have SB support | 16:57 |
| slangasek | it just doesn't work on stgraber's hardware | 16:57 |
| cjwatson | livecd-rootfs has done the right thing | 16:58 |
| cjwatson | It's a necessary question, but I think we should - not using a signed kernel puts us into less-tested territory, and when you say "we do have SB support", we don't actually know if that's true with current images given recent changes | 16:59 |
| cjwatson | I would rather finish the job if at all possible, since the last respin was so recent and this can be iso9660-only | 16:59 |
| infinity | cjwatson: Given that it's cdimage-only, I'm with you, if you have a handle on what went wrong. | 17:00 |
| === bjf[afk] is now known as bjf | ||
| cjwatson | I expect to in ~5mins | 17:00 |
| infinity | cjwatson: Shiny. | 17:00 |
| stgraber | with the current image assuming you are on a firmware that boots (so not a lenovo firmware), you'd get linux-image and linux-signed-image installed but would be missing the .efi.signed file in /boot, so yeah, fixing cdimage seems easier than reverting ubiquity + upload + copying + publishing + respinning everything to get a consistent /boot again | 17:00 |
| slangasek | cjwatson: ok then | 17:01 |
| * skaet waiting to hear results | 17:01 | |
| slangasek | I can certainly test SB here, but not in < ~5min :) | 17:01 |
| infinity | cjwatson: I see no Chinese image in that last set, did you quit before you got there? | 17:02 |
| cjwatson | infinity: The last set is still in progress | 17:02 |
| cjwatson | It has ARM images in it so ... | 17:02 |
| infinity | cjwatson: Righto. | 17:02 |
| cjwatson | Right, so this was in the bit of cdimage that I haven't rewritten in Python with unit tests yet | 17:05 |
| cjwatson | And a missing bit of code in debian-cd | 17:05 |
| cjwatson | Fixing | 17:05 |
| highvoltage | stgraber: edubuntu upgrade testing for i386/amd64 worked without a hitch | 17:07 |
| cjwatson | Fixes deployed - respinning Ubuntu desktop cdimage-only, and will eyeball afterwards | 17:08 |
| cjwatson | I'll need to do Kubuntu desktop as well after that since I was just too late for it | 17:09 |
| cjwatson | But the rest will pick it up automatically | 17:09 |
| stgraber | highvoltage: yay, one of the few good news of the day! keep them coming ;) | 17:09 |
| cjwatson | Riddell,ScottK: FWIW it had already hit plant | 17:10 |
| cjwatson | *planet | 17:10 |
| cjwatson | Oh, Riddell followed up, never mind | 17:10 |
| highvoltage | stgraber: any problems that might need a rebuilt again? | 17:11 |
| Riddell | yeah, playing for pity with my brain damage to avert the anger | 17:11 |
| highvoltage | *rebuild | 17:11 |
| cjwatson | Riddell: heh | 17:11 |
| cjwatson | I C-ced a build in order to just do it for amd64/amd64+mac instead - sorry for mailspam | 17:13 |
| stgraber | highvoltage: in theory all is well now that cjwatson fixed the cdimage/debian-cd scripts, but well, we've said that so many times over the past two days | 17:14 |
| cjwatson | Should have a better idea in a few minutes | 17:14 |
| * highvoltage holds thumbs | 17:14 | |
| cjwatson | "No desktop image for amd64!" wait what | 17:16 |
| * cjwatson digs | 17:16 | |
| cjwatson | Oh, bah, stupid EEXIST | 17:17 |
| * cjwatson refixes | 17:19 | |
| cjwatson | Kubuntu doesn't need a respin because it doesn't have the signed kernel in its seeds | 17:21 |
| cjwatson | And too late to fix that | 17:21 |
| cjwatson | So should be OK to go ahead and test those Kubuntu images | 17:21 |
| Riddell | thanks | 17:22 |
| stgraber | right, only ubuntu desktop and edubuntu dvd have linux-signed installed | 17:22 |
| cjwatson | OK, .4 has /casper/vmlinuz.efi.signed | 17:22 |
| stgraber | yay! | 17:22 |
| cjwatson | And a grub.cfg pointing to it | 17:23 |
| cjwatson | So please test :) | 17:23 |
| stgraber | downloading | 17:24 |
| highvoltage | stgraber: didn't there used to be a https://wiki.ubuntu.com/QuantalQuetzel/ReleaseNotes/Edubuntu page? it seems to have vanished | 17:25 |
| stgraber | highvoltage: fix the typo in the URL and it'll work :) | 17:25 |
| highvoltage | (d'oh) | 17:26 |
| seb128 | ^ rejecting, I will do another upload update to 3.6.1 as well which fixes a regression from the current version | 17:32 |
| cjwatson | Hm, that server build only has an unsigned kernel | 17:36 |
| cjwatson | Oh, because it works differently there | 17:36 |
| cjwatson | It's actually signed but called vmlinuz | 17:37 |
| cjwatson | Should still be OK | 17:37 |
| cjwatson | Unless it has a similar bug with kernel checksums, but if so, it may be tough :-/ | 17:38 |
| cjwatson | It'd take a d-i upload and a live-installer upload and possibly other stuff to change that, and I don't plan to do that in the absence of actual install-breaks bugs like we had on desktop | 17:39 |
| * skaet nods | 17:40 | |
| stgraber | success!!! | 17:45 |
| stgraber | it booted, installed and booted again! | 17:46 |
| infinity | stgraber: Fancy. Desktop? | 17:46 |
| stgraber | yep, desktop on secureboot | 17:46 |
| stgraber | I tested server on secureboot already today and it worked, so I'm not expecting any difference with the respin for the open-iscsi stuff | 17:46 |
| infinity | Right, back to pre-publishing and a Thursday of calm and happy? | 17:49 |
| * skaet likes idea of calm and happy | 17:54 | |
| cjwatson | stgraber: excellent. If you have any time and energy left then a server test would be good to make sure I didn't break it with the desktop fixes in cdimage ... | 18:04 |
| cjwatson | infinity: are you re-prepublishing? Otherwise I'll do so in a bit. | 18:04 |
| cjwatson | Now I just hope that non-SB still works :-) | 18:05 |
| stgraber | cjwatson: ok, will do | 18:07 |
| infinity | cjwatson: I figured you'd be doing it since you know where you were in the building and such. | 18:07 |
| cjwatson | Righto. Will do after dinner. | 18:07 |
| popey | is -proposed on hold or some magic preventing compiz 1:0.9.8.4-0ubuntu3 from hitting the archive? | 18:13 |
| popey | rmadison can see it, but when I apt-get update, I don't get the package in apt-cache policy | 18:14 |
| stgraber | popey: archive mirroring was broken at least a few hours ago, might still be the case | 18:15 |
| popey | stgraber, broken intentionally or otherwise? | 18:16 |
| sbeattie | popey: otherwise. somehow it looks like the archive mirror get reset back to august | 18:17 |
| sbeattie | s/get/got | 18:17 |
| sbeattie | http://archive.ubuntu.com/ubuntu/dists/quantal-proposed/main/binary-amd64/ lists august 17 as the last time it was updated. | 18:18 |
| sbeattie | #is has been notified | 18:18 |
| === henrix is now known as henrix_ | ||
| popey | ta | 18:19 |
| psivaa | i get /casper/vvmlinuz.efi.signed file not found for amd64 oem installation | 18:24 |
| stgraber | psivaa: is that a typo you made on IRC or is that typo part of the error? | 18:25 |
| psivaa | stgraber: sorry that is a typo | 18:25 |
| psivaa | vmlinuz in fact | 18:25 |
| stgraber | ok, at what stage are you getting this error and what's giving you the error? | 18:26 |
| stgraber | is that in the initial install or in the OEM install (user first boot)? | 18:26 |
| plars | stgraber: also if I do check disk for defects, I get that same error | 18:27 |
| psivaa | stgraber: yes | 18:27 |
| infinity | This is on the images just being spun right now? | 18:28 |
| psivaa | stgraber: that is in the initial install menu where you press f4 and select oem installation and then hit enter on 'Install ubuntu'' | 18:28 |
| stgraber | psivaa: can you give me the md5 of that .iso? (just want to confirm that we're indeed in big trouble) | 18:29 |
| psivaa | infinity: yes md5sum d42faaa9aa65f681065c724aaa7d92fc | 18:29 |
| stgraber | argh | 18:29 |
| psivaa | i checked it and its the current | 18:29 |
| stgraber | right, confirmed here in kvm | 18:30 |
| * stgraber investigates | 18:30 | |
| psivaa | stgraber: thanks | 18:30 |
| stgraber | ok, so basically any of the entry fails with file not found | 18:30 |
| stgraber | now, inspecting the .iso, the files are clearly there... | 18:30 |
| * stgraber digs into gfxboot/isolinux | 18:30 | |
| cjwatson | oh meh | 18:32 |
| stgraber | ../casper/vmlinuz.efi.signed works | 18:32 |
| cjwatson | isolinux cares about 8.3 | 18:32 |
| cjwatson | because it's 2012 | 18:32 |
| cjwatson | but why does anything other than ... oh | 18:32 |
| cjwatson | ok, <- idiot | 18:32 |
| cjwatson | I used $CASPER_KERNEL in rather too many places | 18:33 |
| cjwatson | should've only used it for grub.cfg | 18:33 |
| stgraber | ignore my earlier comment | 18:33 |
| cjwatson | cdimage-only respin again | 18:34 |
| slangasek | that affects just ubuntu and edubuntu, right? | 18:36 |
| cjwatson | Yes | 18:36 |
| cjwatson | Maybe not even Edubuntu - it's building right now so it depends if I caught it in time | 18:37 |
| cjwatson | I can check easily | 18:37 |
| * slangasek nods | 18:37 | |
| infinity | cjwatson: I'm still not seeing new Chinese images, which I assume also need to get in on this rebuild. | 18:39 |
| * infinity hasn't seen new ones during any of this. | 18:40 | |
| infinity | Anyhow. I'm going to go grab some food. | 18:40 |
| cjwatson | infinity: It's still in progress. Edubuntu iso9660 and Ubuntu Studio livefs are building now; Chinese will follow. | 18:41 |
| cjwatson | I gave the ordering here at 14:55 UTC. | 18:42 |
| stgraber | also worth noting that chinese won't show up on IRC as they're not posting to the tracker (they are tracked on a separate tracker without auto-posting) | 18:42 |
| cjwatson | I was just in time; Edubuntu is OK | 18:45 |
| stgraber | highvoltage: can you test? ^ | 18:45 |
| cjwatson | psivaa: Please try Ubuntu desktop 20121017.5 | 18:45 |
| cjwatson | (amd64) | 18:45 |
| stgraber | I'm going to do a server test install now on secureboot | 18:45 |
| stgraber | cjwatson: is it worth having me re-test the new ubuntu desktop amd64 on secureboot or are we good here? | 18:45 |
| highvoltage | stgraber: syncing... | 18:46 |
| cjwatson | I didn't change anything which could plausibly cause a problem | 18:46 |
| cjwatson | But I'm really paranoid right now | 18:46 |
| * cjwatson re-prepublishes desktop amd64, and prepublishes server | 18:47 | |
| stgraber | ok ;) I'll re-test once I'm done with server, it doesn't take long to do, it's just a pain because I can't do anything else during that time... | 18:47 |
| cjwatson | Yeah, sorry :-( | 18:47 |
| psivaa | cjwatson: ok | 18:52 |
| slangasek | stgraber: you can kick the amd64 SB test over to me | 18:52 |
| slangasek | for desktop | 18:52 |
| slangasek | I was going to re-test it here anyway since I haven't tried any of the latest images on my hardware | 18:53 |
| cjwatson | FWIW it's possible that the general archive.u.c problems will invalidate install testing happening at the moment | 18:53 |
| cjwatson | Hopefully that'll clear soon ... | 18:53 |
| slangasek | and I can sanity-check that the vmlinuz is signed at all points | 18:53 |
| cjwatson | Well, not invalidate, I suppose, but create many spurious failures | 18:53 |
| stgraber | cjwatson: server amd64 SB is good on 20121017.2 | 18:54 |
| cjwatson | ace | 18:54 |
| * skaet keeping toes now crossed as well as fingers | 18:58 | |
| plars | xnox: I think this is probably related to https://bugs.launchpad.net/ubuntu/+source/ubiquity/+bug/1042647 but I can't seem to manually repartition on my arm board at least, if I've previously installed with lvm | 18:59 |
| ubot2 | Launchpad bug 1042647 in ubiquity "[FFe] [UIFe] Manual Partitioning LVM" [High,New] | 18:59 |
| slangasek | blah, system shutdown on a base install takes too long | 19:01 |
| * stgraber reboots for two efi desktop installs (sb and non-sb to check a report from jibel), back in 20min | 19:03 | |
| skaet | utlemming, Ubuntu Server EC2 HVM (Europe) amd64, just noticed no results on it. tests still running or some other glitch? | 19:04 |
| utlemming | skaet: sorry, tested fine | 19:04 |
| utlemming | skaet: I just didn't mark 'em | 19:04 |
| * utlemming does that now | 19:04 | |
| skaet | thanks | 19:04 |
| utlemming | skaet: done | 19:06 |
| slangasek | jdstrand, stgraber: could one of you two please provide an SRU test case for bug #1067473? | 19:07 |
| ubot2 | Launchpad bug 1067473 in isc-dhcp "[quantal] isc-dhcp-client dropped network-interface-security symlink and therefore may run unconfined" [High,In progress] https://launchpad.net/bugs/1067473 | 19:07 |
| slangasek | (preferably something non-racy) | 19:09 |
| stgraber | desktop amd64 sb is good | 19:10 |
| stgraber | jdstrand: ^ I'm thinking of stopping apparmor, manually triggering the job and running aa-status, should that work? | 19:11 |
| seb128 | ^ whoever is still reviewing universe uploads, any chance to get that geary one in? it's basically adding a build-dep on intltool so it gets a menu entry generated and a one liner bug fix | 19:12 |
| seb128 | geary is made by the yorba guys who do shotwell | 19:12 |
| seb128 | they would probably appreciate to have those fix in, it's their first release ;-) | 19:13 |
| cjwatson | infinity: Chinese edition built now | 19:13 |
| * stgraber reboots again | 19:13 | |
| slangasek | seb128: universe final freeze was yesterday; so no https://lists.ubuntu.com/archives/ubuntu-devel-announce/2012-October/000987.html | 19:16 |
| slangasek | seb128: I can reject and you can re-upload to -proposed | 19:16 |
| seb128 | slangasek, crap, I hate dch :p | 19:17 |
| seb128 | slangasek, I meant to upload to proposed ;-) | 19:17 |
| skaet | thanks seb128, slangasek | 19:17 |
| seb128 | slangasek, but you mean I need to do the SRU paperwork anyway, the freeze was an harder one that the main one? ;-) | 19:17 |
| skaet | :P | 19:17 |
| slangasek | seb128: main is equally frozen at this point | 19:18 |
| skaet | seb128, SRU process following please. | 19:18 |
| slangasek | this is the "you don't get into -release unless you're causing the kittenpocalypse" freeze. | 19:18 |
| * skaet laughs | 19:19 | |
| seb128 | slangasek, it's an universe kittenpocalypse if the yorba guys don't get their launcher, trust me :p | 19:19 |
| seb128 | </worth a try> | 19:19 |
| seb128 | ;-) | 19:19 |
| * slangasek makes a mental note: don't trust the desktop team, they always exaggerate :) | 19:20 | |
| * skaet thinks kittenpocalypse is a better description than the kittenkiller. | 19:21 | |
| jbicha | I'd really like to see the evolution-data-server fix get in as a 0day SRU, it's 0 risk and fixes a high-impact bug | 19:23 |
| stgraber | jibel: efi without sb works fine, so that bug report is a bit weird but certainly doesn't reflect all the non-sb efi installs | 19:23 |
| rtg | skaet, I just uploaded Quantal LBM (bug #1066123). Originally we thought we wouldn't have to, but it turns out there are drivers in compat-wireless that are not yet mainline. Loss of those drivers would cause regressions for some folks. I'll eventually need someone to promote it to main since its a new package for Quantal. (LBM is always new for each release) | 19:24 |
| ubot2 | Launchpad bug 1066123 in linux-meta "Add comapt-wireless v3.6 stack to Precise LBM" [Medium,Fix committed] https://launchpad.net/bugs/1066123 | 19:24 |
| seb128 | slangasek, btw since you mentioned SRU I will drop a note that precise SRU reviews would be welcome ;-) I appreciate it's quantal release busy time but I just hope that busy time doesn't go over until after UDS :p e.g next week would be a good time to review some ;-) | 19:24 |
| slangasek | seb128: well, my review day is Friday and there's a good chance I won't get overrun by 12.10 stuff this week | 19:25 |
| slangasek | so I hope to help there | 19:25 |
| cjwatson | stgraber: what bug is this? | 19:25 |
| seb128 | \o/ | 19:25 |
| skaet | rtg, ack. Release note appropriate then? | 19:25 |
| rtg | skaet, not really | 19:25 |
| rtg | it should have no impact on folks without the HW | 19:26 |
| skaet | k | 19:26 |
| slangasek | rtg: but it has a rather major impact on folks with the hw, doesn't it? | 19:26 |
| rtg | skaet, and LBM has never been on the isntall media anyways | 19:26 |
| jibel | cjwatson, bug 1067822 | 19:26 |
| ubot2 | Launchpad bug 1067822 in ubiquity "Can't boot to newly installed system "grub rescue> "" [Undecided,New] https://launchpad.net/bugs/1067822 | 19:26 |
| slangasek | rtg: won't this result in a potentially unusable system on upgrade? | 19:26 |
| stgraber | cjwatson: bug 1067822 | 19:26 |
| rtg | slangasek, there is no upgrade path for LBM, so they'd have to install it by hand anyways. | 19:27 |
| cjwatson | Yeah, I saw that via answers and followd up | 19:27 |
| cjwatson | +e | 19:27 |
| cjwatson | I expect that some kind of user error is at least implicated even if it isn't the whole story | 19:27 |
| cjwatson | That's certainly the case for the UEFI failure | 19:27 |
| cjwatson | (No EFI System Partition) | 19:27 |
| slangasek | rtg: ok; agreed that it doesn't need a release note then | 19:28 |
| infinity | rtg: I'll check your LBM when I get home from food. | 19:28 |
| slangasek | (not because it's hardware-specific, but because this is not a new requirement for users of LBM) | 19:28 |
| rtg | slangasek, that is my take on it. I'll upload the meta package but it can stall in unapproved until afterrelease | 19:28 |
| * skaet nods | 19:28 | |
| cjwatson | So, yeah, the replace option should have created the extra partition; but a fresh install would have done it, and what this user is trying to do is effectively upgrade from BIOS to UEFI installation in-place which is always going to be fairly advanced | 19:28 |
| rtg | ogasawara, ^^ | 19:29 |
| cjwatson | What's happening with BIOS there is a different story, but firmware is often unclear about what's going on in places where it needs to be clear, and it's quite possible that a partial UEFI installation is interfering somehow | 19:29 |
| cjwatson | I don't see an immediate reason to consider that bug RC | 19:29 |
| cjwatson | At least not if BIOS install tests are otherwise going OK | 19:31 |
| rtg | slangasek, would you pretty please reject the LBM I just uploaded? I neglected to notice the ABI number was incorrect. | 19:31 |
| rtg | skaet, ^^ oops | 19:32 |
| cjwatson | rtg: done | 19:32 |
| rtg | cjwatson, thanks. I'll redo in a minute. | 19:33 |
| * skaet can't get there as fast at cjwatson it appears | 19:33 | |
| skaet | :) | 19:33 |
| slangasek | not sure why y'all are trying, he asked me :-P | 19:33 |
| cjwatson | slangasek: overachieving | 19:33 |
| skaet | :) | 19:33 |
| * slangasek redirects all channel highlights to cjwatson and takes a nap | 19:33 | |
| * skaet finds it satisfying to hit reject these days without complaints ;) | 19:34 | |
| ScottK | slangasek: cjwatson thrives on rejection, I guess. | 19:34 |
| cjwatson | That explains the pain in all the diodes down my left side. | 19:39 |
| rtg | cjwatson, skaet, slangasek: re-uploaded Quantal LBM with the ABI if the kernel that is currently in the archive. We can upload meta now, or wait until after release. Your call. | 19:40 |
| * xnox suddenly thinks about Marvin from hitch-hiker's guide to the galaxy. | 19:41 | |
| slangasek | rtg: no need for you to wait before uploading | 19:41 |
| rtg | slangasek, k, gimme a few minutes. its got the same issue as LBM (re: ABI number) | 19:41 |
| cjwatson | xnox: That was the intent ... | 19:41 |
| xnox | =) | 19:42 |
| cafetiere | rtg how is meta behind? | 19:44 |
| rtg | cafetiere, ogasawara's pull request had the wrong ABI | 19:44 |
| cafetiere | bad Ogasawara | 19:44 |
| ogasawara | cafetiere: it's not behind in the archive, just in my pull request | 19:45 |
| cafetiere | :) ok | 19:45 |
| seb128 | slangasek, ^ geary is back :p with SRU compliant bug and everything ;-) | 19:50 |
| slangasek | seb128: thanks :) | 19:50 |
| slangasek | stgraber, cjwatson: did a SB desktop install test here with 20121017.5, worked fine | 19:54 |
| skaet | cjwatson, have the sources images been built? | 19:54 |
| skaet | (cron.sources) | 19:54 |
| skaet | cron.source rather | 19:54 |
| SpamapS | So, should I be worried about buildds right now? I'm doing normal precise-proposed SRU processing and I see gcc-4.6 in there... | 19:55 |
| ScottK | SpamapS: Last I heard, cjwatson was trying to keep one open buildd per arch, so I'd wait for powerpc to clear up a bit. | 19:56 |
| ScottK | SpamapS: If you want, you can accept my synaptics SRU, it's arch all, so there's plenty of buildds. | 19:57 |
| ScottK | :-) | 19:57 |
| cjwatson | Yeah, please don't start long-running builds like GCC before release now | 19:57 |
| cjwatson | slangasek: great | 19:57 |
| cjwatson | skaet: yes | 19:57 |
| skaet | thanks | 19:58 |
| SpamapS | Will skip any big ones, thanks | 19:58 |
| cjwatson | Pruned back .manifest now too | 20:00 |
| cjwatson | So all of my tasks from "Release minus 1 day" are done | 20:01 |
| cjwatson | OK, if there are no other fires I might try to get an early night and get up early tomorrow | 20:05 |
| cjwatson | But if the installer explodes or something I'd appreciate an SMS | 20:05 |
| xnox | night =) | 20:05 |
| slangasek | cjwatson: ack, g'night | 20:14 |
| skaet | Thanks cjwatson, good night. | 20:22 |
| stgraber | I just added two new indexes to the tracker DB which should fix the biggest performance issue that people noticed. I'm not expecting any side effect and those will be upstream in the next release. | 20:34 |
| stgraber | (querying bug data was horribly slow leading to load times of over 30s for the result pages, it's now down to less than a second) | 20:34 |
| ScottK | stgraber: Much better. Thanks. | 20:38 |
| skaet | thanks stgraber. | 20:50 |
| xnox | stgraber: very snappy =) | 21:32 |
| ogra_ | sigh, is the datacenter already overloaded, zsyncing the images takes ages | 21:37 |
| === Ursinha is now known as Ursinha-afk | ||
| * skaet sees that stgraber and highvoltage have been a testing... :D | 22:16 | |
| stgraber | hehe yeah :) finishing a wubi run on windows8 here and then will mark i386 as good to go | 22:18 |
| Riddell | skaet: what time is the call tomorrow? | 22:20 |
| skaet | Riddell, you've got a calendar invite | 22:21 |
| Riddell | skaet: I know, I can't find it! | 22:21 |
| skaet | I'll resend. | 22:22 |
| skaet | :) excellent stgraber. | 22:23 |
| stgraber | edubuntu testing done and not seeing anything scary on the tracker so I'll go sleep for a bit, see you all in the morning | 23:08 |
| slangasek | 'night, stgraber | 23:09 |
| * ogra_ twiddles thimbs watching two netinst instally | 23:12 | |
| ogra_ | *installs | 23:12 |
| skaet | http://www.markshuttleworth.com/archives/1195 - Raring Ringtail | 23:52 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!