[00:44] partman-base: cjwatson * r96 ubuntu/debian/changelog: releasing version 121ubuntu1 === cjwatson_ is now known as cjwatson [15:02] bdmurray: I added a test case for 241420 [15:03] xivulon: I had updated wubi.hardy to include that change as r504 yesterday. Are you sure you're synced up with LP? [15:44] evand strange I checked yesterday and was r503 [15:45] in fact it didn't flag any conflict [15:45] https://code.edge.launchpad.net/~ubuntu-installer/wubi/hardy [15:45] argh, I thought that branch was bound. [15:45] Apparently not. [15:45] I'll revert my changes and merge. [15:46] the bug evaded my as well, debugging something that happens at the end of the installation late at night is quite annoying... [15:47] do you have any insight there? [15:47] it's definitely the bind mounting, I'm just not sure why it would cause such a problem. [15:47] I'll keep poking at it, though I welcome any help or suggestions. [15:49] is it possible to narrow down what the actual "subprocess" running at the time it fails? [15:52] I'm fairly certain it starts in python-apt, curiously enough. [15:53] It's probably a red herring as mentioned before but printing self.percent makes the bug disappear (which makes absolutely no sense to me). [15:54] yes had the same thing happening to e [15:54] me [16:20] evand: that's it? [16:25] oh, actually no [16:25] I'll fix momentarily [16:25] and I thought I'd just finished it. :( [17:11] xivulon: wubi seem to install okay with the patched selfextract version that evand did yesterday [17:18] xivulon: until ofcourse the installation failed [17:19] evand: xivulon: want a copy of the .log files? [17:19] is or it likely to be the same issue as before? [17:20] same issue, thanks though [17:25] evand: what does Finnish look like in the language choice list? [17:26] evand other than that the other think to look at is that panic messages are not displayed in non-recovery mode [17:26] evand: np's self extractor works till then anyway :) [17:26] bdmurray: Suomi, I believe [17:26] davmor2, can you pls comment in bug (if you haven't done so already) [17:26] evand: okay, that's what I was going to guess [17:27] I am not sure if it is intentional or not, I have to test what is the normal panic behaviour [17:27] but I think that if you are thrown into busybox any panic message should be visible whether you are in recovery more or not [17:57] evand: was the self-extract fix based on 504 or 503? [17:57] 504 [17:57] evand: ta adding it to the bug report [17:59] xivulon: comment added [19:08] xivulon: I won't be around now till Wednesday next week on hols so if you need tester just ping the testing channel :) [19:33] oh thank heavens [19:33] I am trying to install ubuntu hardy from a usb stick using a netinstall base [19:33] do I need to specifiy something special to get the modules to load? [19:34] now it loads up to 50% and then fails through [19:36] holst: did you follow section 4.3 of the installation guide to the letter? [19:37] https://bugs.launchpad.net/ubuntu/+source/net-retriever/+bug/234486 [19:37] it looks like that one [19:40] i dont know which document you are referring to [19:40] there are about 1001 guides for this sutff [19:41] ah, give the updated images as mentioned in that bug report a shot then [19:41] there's really only one official guide [19:41] sudo apt-get install installation-guide-i386 [19:42] oh [19:42] thakns [19:42] (damn keyboard :P) [19:42] then point your browser at /usr/share/doc/installation-guide-i386/en/index.html [20:03] yes, very cool [20:03] now it works, with that update [20:05] it seems that the boot record on the memory stick is overwritten by the installer [20:06] because i cannot boot on the stick after the installer has completed [20:06] or, not overwritten, but overrided [20:08] if you install the bootloader to hd0, then you're installing to the USB stick [20:09] thats probably what happens [20:21] there is no real need to have /boot on a separate partition, unless you are encrypting, right? [20:21] (or using software raid or some other exotic stuff) [20:22] and also, is there some more portable way to specifiy "first hard disk"? [20:22] or if you have an older BIOS that can only read the first 1024 cylinders of the disk. [20:22] using /dev/sdf does not seem to be that secure [20:23] I mean, on another PC that might be /dev/sdb, or something else [20:23] unfortunately, no. GRUB and Linux have differing opinions on what ends up being the first device. One of the kernel team members is adding UUID support to GRUB for 8.10 though. [20:23] Which will fix this. [20:23] cool [20:24] then there will be no ambiguity what is "the hard disk"? [20:24] or (first) harddisk even [20:25] well, all our pc:s here are of the same model [20:25] so if there are a higher power on our side, and Murphy stays on vacation, then it might work until 8.10 :) [20:31] heh [20:34] hmm, there is a (last) snag in the installation [20:34] we have something called /NOBACKUP on all boxes [20:34] which of course must be saved during reinstalls [20:35] otherwise the users will kill us :) [20:35] so, I want to format it if it does not exist [20:35] but if it DOES exist [20:35] then I just want to mount it [20:35] http://rafb.net/p/PdwPIi29.html [20:35] I tried that [20:35] but it does not get formated, or mounted [20:40] maybe I missunderstood how flexible that stuff was [21:21] holst: it doesn't sound from your description as if /NOBACKUP is a separate partition [21:21] holst: if it's just an ordinary directory on /, then you can do it by just telling the partitioner to mount / but not to format it (as of 8.04) [21:22] holst: however, unfortunately, it's really hard to do this automatically at the moment, because the automatic partitioner doesn't have any support for selecting an existing partition [21:22] this is something I'm hoping to fix - but at the moment all it can do is create new partitions [21:23] if /NOBACKUP is a separate partition, then I've seen ways to do this by deleting all the other partitions you *don't* want to keep before the partitioner runs, then telling the automatic partitioner to operate on the free space, then mounting it at the end - but they're all a bit hackish [21:35] hi [21:35] just do manual partitioning, right? [21:36] mount it, select not to format it, and cross your fingers that you didn't do it wrong. [21:36] right, but that's not always viable ... [21:36] if you're doing 200 boxes it's not really sane :) [21:36] ah.. I should read back [21:37] (I don't know he's doing 200 boxes, but infer that there is a reasonable number from the fact that preseeding was involved at all) [21:37] also.. where do you work? [21:37] I want to work somewhere with 200 ubuntu boxes.. [21:38] I think the short and sweet here is that unless all the boxes are identical in partitioning, or specifically have a "tagged" fiesystem (label, uuid database, something) that you will have a somewhat destructive install method. [21:39] I'm guessing here, but if you can script a preseed file to check out the existing fstabs then you should be able to use existing partitions to map out what's not NOTBACKUP and simply format those then add the notbackup fstab line back later.