/srv/irclogs.ubuntu.com/2009/10/01/#ubuntu-installer.txt

* evand blinks at https://bugs.edge.launchpad.net/ubuntu/+source/ubiquity/+bug/28911502:22
ubottuError: This bug is private02:22
evandI'm trying to do a pass of all the configure_bootloader bugs.  Some of them are quite interesting.02:28
evandif there was any doubt, installing to an SD card seems to work fine02:28
evandthough I can't find a machine capable of booting from one :)02:28
=== cjwatson_ is now known as cjwatson
=== cjwatson_ is now known as cjwatson
=== dpm__ is now known as dpm
=== cjwatson_ is now known as cjwatson
cjwatsonevand: bug 289115> there are several squashfs errors before the python backtrace in the log - I think it's a case for a "sorry, mate, your CD's bust" stock reply if you have one lying around11:55
ubottuBug 289115 on http://launchpad.net/bugs/289115 is private11:55
evandah, I missed that in amongst all the NetworkManager noise when I looked last night11:59
evandgood deal11:59
evandthanks cjwatson11:59
davmor2cjwatson, evand: is this something similar https://bugs.launchpad.net/ubuntu/+source/casper/+bug/436418 guys gets stdin errors but when I checked yesterday the install was fine.12:04
ubottuLaunchpad bug 436418 in casper "Installation from USB media fails on Acer Aspire One" [Undecided,New]12:04
cjwatsonnot directly the same thing, certainly12:04
cjwatsonwe get all sorts of bugs that are hardware problems of some kind, though it's not always easy to sort them out12:05
cjwatsonwe nearly missed a genuine serious squashfs bug once because I was automatically dismissing all squashfs errors as hardware problems :-/12:05
cjwatson"works with 9.04 but not with alpha 6" doesn't sound easily dismissable as a hardware bug ...12:06
cjwatson(especially on USB where you don't have random burn errors)12:06
cjwatsonit'd be nice if we knew which process was saying "stdin: error 0"12:06
evandif memory serves, that appears at least once for me on a perfectly fine iteration of casper.12:07
evandbut obviously not necessarily the same thing, as you say12:07
* evand should read logs first more often :-/12:08
CIA-33usb-creator: evand * r225 trunk/debian/changelog: releasing version 0.2.814:27
CIA-33usb-creator: evand * r226 trunk/ (debian/changelog usbcreator/install.py):14:30
CIA-33usb-creator: Properly catch exceptions around modifying the syslinux14:30
CIA-33usb-creator: configuration (LP: #439977).14:30
=== robbiew-afk is now known as robbiew
rgreening_yay14:32
xivulondavmor2, evand, have noticed 43927914:33
davmor2xivulon: Yes I wrote14:33
davmor2it14:33
evand0.2.8 was uploaded ages ago, I just forgot to debcommit -r.  It's waiting to be accepted, which will occur after beta.14:34
xivulonhmm haven't seen that myself, quite annoying, does it happen often?14:34
CIA-33grub-installer: cjwatson * r813 ubuntu/ (debian/changelog grub-installer): Fix typo in fallback case when computing $disc_offered_devfs.14:36
davmor2xivulon: nearly every time on kubuntu it fails.  On ubuntu it work 80% of the time but the other 20% needs rebooting at least14:37
cjwatsonxivulon: I've observed it quite a bit, but it may be because syncio apparently got lost14:37
cjwatsonalso, wubi wasn't passing syncio correctly14:37
cjwatsonI've fixed both of those in changes queued for post-beta14:37
cjwatsonevand: I have http://paste.ubuntu.com/282991/ in ~/bin/ubuntu-release14:38
xivulonthanks for that, sorry for the trouble14:38
cjwatsonI don't know if that's all of it. I'm not sure if syncio explains the random hangs I see in the middle of the installer14:38
cjwatsonI was speculating that we might need to wait for sreadahead to finish, but that could be a total red herring14:39
xivulonany such report on the ntfs-3g mailing list?14:39
evandcjwatson: neat, thanks!14:39
cjwatsonI think I'll see how it goes with the new ntfs-3g/wubi and work from there14:39
=== rgreening_ is now known as rgreening
cjwatsonxivulon: syncio was always an Ubuntu patch so I don't think it's something upstream should be bothered about14:39
xivulontrue14:39
CIA-33grub-installer: cjwatson * r814 ubuntu/ (debian/changelog grub-installer):14:41
CIA-33grub-installer: If /boot is on an MD device and we're using GRUB 2, install GRUB there14:41
CIA-33grub-installer: rather than (hd0); GRUB 2 will interpret that as meaning that it needs14:41
CIA-33grub-installer: to install to each of the RAID members (LP: #427048).14:41
xivulonbut if it was a sreadahead / syncio interaction, wouldn't the issue also be present post-install?14:41
xivulonwell I assume once you install and reboot the test is usually finished...14:43
xivulonand why do reboots help? Does it always work when you reboot?14:44
cjwatsonI mean rebooting to try the installer again14:47
cjwatsonI haven't made any significant use of the installed system14:48
cjwatsonin davmor2's failing tests, the installed system was just unbootable because grub was screwed in some way; for example, missing grub.cfg with no indication in the logs that it failed to write grub.cfg (which there *really* should have been)14:48
cjwatsonthat's what made me think syncio might be involved14:48
xivulonthat seems plausible, I would be surprised though if you have a high failure rate on Kubuntu on first reboot and high (full?) success rate on second reboot14:51
xivulonwith same syncio settings14:51
cjwatsonI don't think anyone's managed to measure success rate on second reboot14:52
cjwatsonit's definitely very intermittent for me14:52
cjwatsonand davmor2's results seem all but random14:52
cjwatsonsome successes, some failures, not much in the way of rhyme or reason14:52
cjwatsonwe'll see what happens once we know syncio is in place, I guess14:52
xivulonok then, I misinterpreted the bug report14:53
cjwatsonthe fact that partitioning seems to hang sometimes suggests to me that maybe something is hanging on to the disk14:55
cjwatsonbut the fact that I seem entirely unable to switch away from X in wubi impedes investigation14:56
xivulonThe syncio is only a boot arg so davmor2 should be able to test on kubuntu15:03
cjwatsonno15:04
cjwatsonsupport for it was accidentally removed from ntfs-3g in a merge15:04
cjwatsonactually during the jaunty cycle, so of course this may not be the whole issue15:04
cjwatsonI put it back in and there's an upload in the queue for post-beta15:04
xivulonhmm could it be that it was moved upstream so that -o sync is enough?15:05
xivulonI also removed it in jaunty cycle, and I do not recall if it was an omission during python translation or deliberate because of -o sync15:06
cjwatsonno, I grepped for the underlying O_SYNC etc. options and there were no matches15:07
cjwatsonI spoke to the person who merged it, and he wasn't aware of why he'd removed it, and thought it was an accident15:07
cjwatsonhmm15:08
cjwatsonthat said it may well be that -o sync would be good enough15:08
cjwatsonbut wubi doesn't pass that15:08
=== nxvl_ is now known as nxvl
cjwatsonnot unless it's on FAT anyway15:09
cjwatsonso I'm not convinced - at this point we should probably get syncio working again and stick with that15:09
cjwatsonthe fact that other instability is new since jaunty does sort of suggest to me that it might be something to do with upstart doing things in parallel, and sreadahead is a likely candidate for trouble there, but I've had no opportunity to prove this yet15:10
xivulonI'll have a quick chat with cking15:11
cjwatsonI'd like to minimise change *for karmic* even if cking says it's ok; we can make bigger changes for lucid15:16
cjwatsonwe're having so much trouble getting this to work as it is that I think it is important to keep things under control15:17
xivuloncjwatson, cking does not remember any relevant change in that respect, so it is likely that it was lost in traslation on my side, and similarly on the ntfs-3g side15:19
xivulonso yes adding it back seems the most reasonable path15:20
CIA-33apt-setup: cjwatson * r172 ubuntu/ (apt-setup-verify debian/changelog):16:40
CIA-33apt-setup: Use debconf-apt-progress for our Ubuntu-specific all-in-one-go 'apt-get16:40
CIA-33apt-setup: update' run.16:40
CIA-33apt-setup: cjwatson * r173 ubuntu/ (6 files in 3 dirs):16:50
CIA-33apt-setup: Re-enable progress bar cancellation. Thanks to debian-installer-utils16:50
CIA-33apt-setup: 1.70 and fixes in debconf, it all seems to work fine now (LP: #172879).16:50
cody-somervilleWhat does the noprompt boot option do?17:06
cjwatsonsee casper 1.109 changelog17:08
CIA-33console-setup: cjwatson * r114 ubuntu/ (Keyboard/ckbcomp debian/changelog):20:17
CIA-33console-setup: Forbid Unicode keysyms in the range 0xf000-0xffff, as kbd 1.15-1 rejects20:17
CIA-33console-setup: the entire keymap if it contains any keysyms in that range (LP:20:17
CIA-33console-setup: #416949).20:17
cjwatsonAnAnt: ^- told you I'd get round to it eventually ;-)20:19
CIA-33console-setup: cjwatson * r115 ubuntu/debian/ (changelog config.proto): Add default codeset/layout/variant for Asturian.20:21
CIA-33console-setup: cjwatson * r116 ubuntu/debian/ (changelog config.proto):20:22
CIA-33console-setup: * Backport from trunk (Jordi Mallach):20:22
CIA-33console-setup:  - Set XKBVARIANT to "cat" for Catalan.20:22
CIA-33console-setup: cjwatson * r117 ubuntu/debian/changelog: releasing version 1.34ubuntu320:24
CIA-33console-setup: cjwatson * r118 ubuntu/debian/ (changelog config.proto): Fix backports of Asturian and Catalan changes in 1.34ubuntu3.20:27
CIA-33console-setup: cjwatson * r119 ubuntu/debian/changelog: releasing version 1.34ubuntu420:28
AnAntcjwatson: thanks !20:43
AnAntto be honest, I lost hope20:44
AnAntthat you would ever have time for it20:44
cjwatsonthat'll teach you a lesson then ;-)20:45
CarlFKcjwatson:    bug 439456  "missing nic firmware" I mentioned yesterday.  is this something you can decide on and make happen?21:14
ubottuLaunchpad bug 439456 in linux "netboot e100/d101m_ucode.bin not found" [Undecided,New] https://launchpad.net/bugs/43945621:14
cjwatsonCarlFK: I'm not a kernel developer - #ubuntu-kernel would probably be a better place to poke21:23
davmor2cjwatson: if there  was a fix for the keyboard detection in oem  on the cd's it didn't get past to the dvds21:24
cjwatsonI don't even know how that was fixed, I think it was luck :)21:24
cjwatsonI suspect it was in fact not really fixed and you happened not to run across it for some reason, TBH21:24
CarlFKcjwatson: will do21:30
=== nxvl_ is now known as nxvl
TheMusocjwatson: I noticed in the installer for UbuntuStudio, there are still the prompts about the grub command-line for linux/defaults etc. I have also noticed that this is fixed for Ubuntu proper. Do I have to do anything to get rid of these prompts for the studio installer?21:49
cjwatsonTheMuso: uh, first I've heard of it, can you get me an installation syslog made with DEBCONF_DEBUG=developer on the installer boot command line please?22:57
TheMusocjwatson: Will do as soon as I can.22:59
CIA-33apt-setup: cjwatson * r174 ubuntu/debian/changelog: releasing version 1:0.41ubuntu223:16
CIA-33casper: cjwatson * r703 trunk/debian/changelog: releasing version 1.19723:29
=== robbiew is now known as robbiew-afk

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!