[00:43] cjwatson: you around? [00:46] shtylman: sort of (don't ask to ask, though, just ask) [00:47] cjwatson: :) [00:47] so we are trying to install hardy (yea... its old) on a new IBM x3550 [00:47] and during the server install [00:48] several things go wrong... the first of which is it can't find the cdrom module [00:48] like the install boots [00:48] and we go through the language page [00:48] but then it says: "No common CD-ROM drive was detected" [00:48] even though it booted from a cd?! [00:49] does any of that sound familiar? [00:49] and is there a way around it (which still installing from a cd) ? [00:50] "even though it booted from a cd"> a lot of people have this confusion - booting from a CD means that the BIOS could read the CD, whereas the error message means that Linux can't read the CD [00:50] but how can linux not read the cd if the installer started? [00:50] usually this means that either the kernel doesn't support the CD for some reason, or it requires a new module which isn't in the cdrom initrd [00:50] the installer started by means of the BIOS reading data from the CD [00:50] I see [00:50] starting the installer does not require Linux to read from the CD [00:51] remember, the first bit of the installer lives entirely in the kernel and initrd, and those are both read from the CD by the bootloader (using the BIOS) [00:51] I see [00:52] dmesg does say it detected sr0 tho [00:52] for a cdrom device... or is that unrelated? [00:52] I think it's only part of it [00:52] basically saying it can see it ... but can't really use it [00:53] you're not using the netboot installer burned to a CD or anything funny like that? [00:54] nope [00:55] right. it's probably a kernel bug (I think). they'll need to know the PCI id of your CD-ROM controller [00:55] (lspci) [00:55] k [00:56] well... its fixed in lucid ... so yea... [00:56] ok..that issue asside... as there is a more pressing issue I wanted to ask about [00:56] we can do a net install to bypass this cdrom issue [00:56] so that is ok.. [00:56] but then... it doesn't detect any harddrives... [00:57] now.. the is interesting for a few reasons [00:57] sort of sounds like the CD is fucked. what URL did you get it from? [00:57] or I suppose it could be that it's completely failing to deal with the PCI controller or something [00:57] the raid controller the drives are connected to are the same controllers as in another box (which is running hardy) ... yet in this new box they are not detected [00:58] the failing to talk to pci would most certainly line up with the hdd/raid detection issue [00:58] as lspci does list both [00:58] it would, but it's not the only possible cause [00:58] I see [00:58] it might not even be the most likely one [00:58] thoughts? [00:58] 00:57 sort of sounds like the CD is fucked. what URL did you get it from? [00:58] checking [01:00] ubuntu's website [01:00] sorry, not specific enough [01:01] :) [01:01] it was a while ago [01:01] I need a URL on releases.ubuntu.com or cdimage.ubuntu.com, or one of their mirrors [01:01] should we try a new image? [01:01] the reason I ask is that sometimes the kernel version on the CD and the kernel udebs shipped on the CD get out of sync, and then nothing works [01:01] this can particularly be the case if you picked a random daily build rather than a proper blessed release [01:02] this is 8.04.1 [01:02] but I do understand the concern [01:02] we can try a new iso image... but the hdd issue comes up over netboot as well [01:04] are there any other steps I could take wrt the hdd detection? [01:05] something I should look for? [01:06] netboot is actually much more vulnerable to kernel version skew [01:06] great :) [01:06] but for the rest, it would really be better to ask the kernel team, I'm not great at debugging this kind of thing [01:07] noted [01:07] as a parting note: http://ubuntuforums.org/showthread.php?t=1219514 [01:07] that mount cdrom is the same issue [01:07] same machine as well [01:08] but the "fixes" seem very shady to me [01:10] judging from that thread, frankly, it sounds like you have a new SATA controller (which both hard disks and the CD-ROM drive are hanging off) that simply isn't supported by 8.04 [01:11] :) [01:11] 8.04.4 is worth a try but I wouldn't hold out much hope [01:11] k [01:11] this is more plausible than my PCI controller hypothesis above (PCI doesn't change that much ...) and explains the symptoms in the same way [01:11] k [01:14] cjwatson: well...I appreciate the help [01:14] I will poke around [01:30] ubiquity: cjwatson * r3914 ubiquity/ (4 files in 3 dirs): [01:30] ubiquity: Keep track of how partitioning progress is displayed without requiring [01:30] ubiquity: an argument to be passed to progress_stop (LP: #535630). [01:34] shtylman: is bug 526534 still open? [01:34] Launchpad bug 526534 in ubiquity "kubuntu ubiquity crashes after first page" [Undecided,New] https://launchpad.net/bugs/526534 [01:35] oh, I guess it might be, the bug summary is not really an accurate summary of what's left of the bug after Riddell's workaround [01:36] guess somebody needs to experimentally unhide the progress bar on a live image and see if it still crashes [01:36] cjwatson: oh.. that was a weird crash [01:36] I have unhid it [01:36] and have never run into that crash [01:36] I think riddell did the install today with the new progress bar and it was ok [01:37] cjwatson: im not sure if that problem is reproducable [01:40] ubiquity: cjwatson * r3915 ubiquity/ (d-i/manifest debian/changelog): Automatic update of included source packages: user-setup 1.28ubuntu5. [01:42] ok, well, somebody with relevant clue should update the bug 'cos it's on the beta-1 milestone list :) [01:43] cjwatson: haha... ok.. I will mark it as closed cause it is as far as we are concerned... [01:44] cjwatson: although the failed to lock is kinda upsetting [01:44] and Riddell did mention having "some sort" of permission problem [01:44] but it wasn't related to the installer crashing [01:44] so that bug report has probly gotten out of hand and can be closed anyway [01:48] you might also find those permissions problems have gone away (or at least got less serious) due to my KApplication initialisation change in 2.1.32 [01:48] cjwatson: yea..i saw comments for that... [01:48] hopefully that have [01:48] *they [01:49] ubiquity: cjwatson * r3916 ubiquity/debian/changelog: releasing version 2.1.34 [02:44] ubiquity: cjwatson * r3917 ubiquity/debian/ (changelog ubiquity-frontend-kde.install): [02:44] ubiquity: * KDE frontend: [02:44] ubiquity: - There are no more files matching gui/qt/images/*.jpg. Don't try to [02:44] ubiquity: install them. [02:48] ubiquity: cjwatson * r3918 ubiquity/debian/changelog: releasing version 2.1.35 [12:55] cjwatson: I'm still getting the issue with the software updates popup when I enter live desktop. [13:15] partman-base: cjwatson * r188 ubuntu/ (debian/changelog parted_server.c): [13:15] partman-base: Apply optimal alignment constraints to new partitions, or when [13:15] partman-base: maximising an extended partition (LP: #530071). [13:36] Desktop images are broken for Xubuntu today: The following packages have unmet dependencies: libjpeg-progs: Depends: libjpeg7 but it is not installable E: Broken packages [13:37] not really an installer issue :) [13:37] Thanks [13:38] http://people.canonical.com/~ubuntu-archive/testing/lucid_probs.html lists it - it's just uninstallable, period [13:39] Does it have to removed from the image, then? [13:39] I suspect libjpeg-progs is in fact required [13:39] some numpty removed libjpeg7 from the archive before we were actually ready to get rid of it, I think [13:40] sigh, MARTIN [13:41] great... [13:41] needs some proper repair work [14:03] lol [14:11] Oh, you know I reported a bug against that new installer screen ? [14:11] bug 536838 [14:11] Launchpad bug 536838 in ubiquity "[Lucid] Desktop cd starts at install screen with no quit/cancel option " [Undecided,New] https://launchpad.net/bugs/536838 [14:11] Takes 8 minutes to get to it [14:14] charlie-tca: What would you expect "Quit" or "Cancel" to do? How about "Restart" and "Shutdown" instead? (the 8 minutes being a different issue). [14:15] They don't work [14:15] That was a different bug [14:16] I would expect at least to be able to do something. the choices allowed are really limited [14:16] I could accept restart or shutdown, though [14:16] I'm just not sure I understand what either "Quit" or "Shutdown" would be supposed to do at that point. [14:16] Right now, when it fails, you can't even debug it [14:16] Err, "Quit" or "Cancel". [14:17] It won't even allow you to switch to another tty [14:18] Something to let me say "oops, I didn't want this afterall" ? [14:18] Quit would quit the installer and drop to a live session. [14:18] Right now, it is called a hard power off, turn the system back on, then I can remove the cd [14:19] although as it happens "Try 10.04" would have the same effect [14:19] I used to be able to remove the cd at the menu and use Ctrl+Alt+del to restart, now I can't [14:19] won't even allow you to switch to another tty> that is *not* within the installer's control [14:20] it's a bug somewhere else [14:20] we couldn't prevent that in installer code even if we wanted to [14:20] I see [14:20] anyway to get a debug for that? [14:21] I would wait until Keybuk uploads his plymouth work before bothering [14:21] okay [14:21] that's the sort of thing that plymouth bugs could break [14:22] is that on gfxboot ? [14:22] i mean the stage thats described in the bug [14:22] no. [14:23] kickseed: cjwatson * r279 ubuntu/ (debian/changelog kickseed.sh): [14:23] kickseed: Drop nouveau preseeding, no longer needed with the drm backport in [14:23] kickseed: 2.6.32-16 kernels. [14:23] for Ubuntu images, the choice now normally appears at the start of the installer [14:24] unless you press a key within five seconds of the bootloader starting up [14:24] ah [14:24] i havent seen any x86 since felt 10 years :) [14:24] at least image/installer wise [14:25] kickseed: cjwatson * r280 ubuntu/debian/changelog: releasing version 0.53ubuntu4 === mcasadevall is now known as NCommander [16:56] forgot about this - http://chipx86.github.com/gtkparasite/ - should come in handy [17:11] is there a way to preseed ubiquity to prevent it from attempting to install additional packages from the CD, re bug #534605? [17:11] Launchpad bug 534605 in ubiquity "During netinstall, ubiquity prompts when failing to install packages from the CD." [Undecided,New] https://launchpad.net/bugs/534605 [17:12] this happens when attempting to netinstall on a system which happens to have a cdrom drive, no problem on netbooks which typically don't have a cdrom drive [17:13] cr3: Is that netboot only, or does the same happen with alternate install from USB? [17:14] persia: haven't tried usb installs in a while [17:16] Just wondering if it comes from overly broad interpretation of what it means for an optical drive to be available, or if it's specific to netboot. [17:17] But someone else might have deeper insight into what you could check. [17:17] 0xb8b1a8 [17:18] ? [18:08] cjwatson: fyi 8.04.4 worked... it was able to see the drives when using netboot (bypassing the cd issue altogether) [18:08] thanks for pointing me to that [18:16] oh cool, that was a bit of a long shot [19:22] partman-base: cjwatson * r189 ubuntu/ (debian/changelog parted_server.c partman-command): [19:22] partman-base: Add an ALIGNMENT_OFFSET command which can be used to detect whether a [19:22] partman-base: partition is misaligned. [19:27] debian-installer: cjwatson * r1256 ubuntu/ (build/config/armel/dove.cfg debian/changelog): Move Dove images to 2.6.32-202 kernels. [20:51] debian-installer: cjwatson * r1257 ubuntu/ (38 files in 2 dirs): Update help translation files. [20:57] debian-installer: cjwatson * r1258 ubuntu/debian/changelog: releasing version 20081029ubuntu90 [21:01] partman-base: cjwatson * r190 ubuntu/debian/changelog: releasing version 138ubuntu3 === robbiew is now known as robbiew_ [23:11] wubi: Agostino Russo * r174 trunk/ (4 files in 2 dirs): Updated references to 10.04/lucid