[00:02] Yeah. [04:53] Remind me, what's the magic thingy/hostname=bar:3142 thingy/directory=/ubuntu to use a local host instead of security.ubuntu.com for the security repo? [04:53] (Expert more still doesn't ask about that, and our network here is annoying.) [04:58] Hmm, looks like security_host. [05:02] ...except that doesn't work on Hardy. [05:09] Oh fuck me, tcp window scaling is still broken. [05:09] I hate out network :-/ [08:27] debian-installer: cjwatson * r964 ubuntu/ (3 files in 2 dirs): Move mainline architectures to 2.6.27-4 kernels. [11:25] davmor2: did you play with ntfs syncio? [11:33] No but I can today now that all the testing is out of the way :0 [11:33] :) even [11:35] Right just run through exactly what you need you need me to do then xivulon [11:43] davmor2 you have to mount an ntfs partition with -o syncio (assuming that ntfs in intrepid has the cking patch built-in) [11:43] hmm [11:44] 2 ticks then I'll just reinstall xp [11:44] I actually built the driver myself. Then try to copy an ISO to the new mountpoint. In my case the operation was terribly slow [11:44] There is no need to have XP, only an ntfs partition. [11:46] it only takes a couple of minutes to transfer it back over :) [11:47] xivulon: so is this from live or from within wubi? [11:48] from a standard intrepid installation, you should compare the performance of ntfs when mounted with and without syncio [11:49] np's [11:49] make sure that syncio is actually "on" [11:50] IIRC it will print a message confirming that syncio is active once you mount the partition [11:51] is the full command for mount on the bug report? [11:53] is it just a case of mount -t ntfs-3g -o syncio /dev/sda1 /media/disk ? [11:53] mount -t ntfs -o syncio /dev/XXX /mnt [11:54] I am looking at https://edge.launchpad.net/ubuntu/+source/ntfs-3g not sure if ntfs-3g is in release or update as both ubuntu2 and ubuntu3 version contain the same changelog [11:54] the table at the top of that page tells you what suites contain what versions of ntfs-3g [11:55] s/not sure if ntfs-3g is in release or update/not sure if ntfs-3g is in release or proposed/ [11:55] you're misreading the version numbers; look at the bit before "ubuntu" as well :) [11:55] that UI lists the packages in upload order, not necessarily in version number order [11:55] ah yep! [11:56] so davmor2 the version in intrepid (1:1.2506-1ubuntu2) should be fine [11:56] cjwatson: so does it only need to list type as ntfs or does it still need to be ntfs-3g or are the 2 the same now? [11:57] they've been the same for *ages* [11:57] since ntfs-3g was introduced properly in, what was it, gutsy? [11:57] cool [11:57] lrwxrwxrwx 1 root root 12 2008-07-11 16:54 /sbin/mount.ntfs -> /bin/ntfs-3g [11:57] if you want the old one you use -t ntfs-fuse [11:58] people should generally just use -t ntfs [11:59] okay cool thanks I just thought I'd double check that, that wasn't the cause of the issue before I start :) [12:03] hmm, cjwatson was looking at https://bugs.edge.launchpad.net/wubi/+bug/236975/comments/3 was curious why is step 1 required to begin with [12:04] I remember that init parses cmdline and extracts rootflags, which then are used by mountroot [12:04] it isn't, as far as I can see. init already exports ROOTFLAGS [12:07] where does this local-premount/root_locale script come from? [12:09] it is just a recipe of some users trying to change the mount attributes of /host [12:10] https://bugs.edge.launchpad.net/wubi/+bug/236975 [12:12] ah probably the chap is referring to hardy (0.85) and export ROOTFLAGS was introudced in intrepid (0.92a) [12:15] ah, ok [12:42] xivulon: definitely slower. It's going at 2.6 MB/sec I transferred it from my server to my test machine faster than that :( [12:43] I'll check it against standard in a second for you :) [12:47] 21 mb/sec + for standard [12:48] xivulon: ^ [12:49] davmor2 please post your findings in 204133 [12:49] same as me 10X slower :( [12:50] I have notified cking [12:53] Np's let me know if it needs testing again I've subscribed to the bug to keep an eye on it's progress anyway [12:54] davmor2 thanks === davmor2 is now known as davmor2-lunch === davmor2-lunch is now known as davmor2 === TheMuso_ is now known as TheMuso === superm1 is now known as superm1|away === superm1|away is now known as superm1 [16:38] debian-installer: cjwatson * r965 ubuntu/debian/changelog: releasing version 20080522ubuntu15 [19:11] cjwatson, et al: is there somewhere in /etc, /proc, or /sys where the keyboard layout is stored? [19:12] cjwatson: i'd like to implement a programmatic workaround for the kvm keymapping bug that's still requiring "kvm -k en-us" (or the like) by dynamically determining the value of that "en-us" bit from the host [21:47] fwiw, ~ubuntu-installer/ubiquity/trunk has been upgraded. [22:38] kirkland: /etc/default/console-setup [22:39] evand: cool, thanks [22:39] kirkland: note that -k en-us still isn't the whole story - when you set the keyboard layout in the guest, e.g. by running the installer, it still produces a broken layout then [22:40] kirkland: you can try this by selecting the British layout and pressing what IIRC is your \ key (our # key) [22:41] cjwatson: interesting, is there a bug on that issue? [22:41] I don't know if there's a specific one, I assumed it was part of the general fuckedness [22:41] haven't tried it with the VNC frontend yet [22:42] is the SDL frontend really so hard to fix? [22:42] :-D [22:42] i really have no idea how hard the SDL is to fix; soren has indicated that it's non-trivial [22:44] cjwatson: is there a problem if you send it "-k en-uk" and you select the British layout in the installer? [22:44] en-gb, but let me try [22:44] ISTR yes [22:45] oh, umm, at the very least things will be confused early on since the live CD defaults to a US keymap [22:47] cjwatson: hrm, i see === superm1 is now known as superm1|away [23:48] odd partition table "Warning: extended partition does not start at a cylinder boundary." but XP is currently booting/useing it OK. [23:48] alt-installer doesn't show any partitions, but does show the raw drive as a place to install to [23:48] i don't see a way to re-partition [23:49] http://dpaste.com/79878/ sfdisk --list [23:52] yes, parted is rather pedantic about some of those kinds of things [23:53] the most useful piece of information is usually 'sudo od -tx1 -Ax -N512 /dev/sda' to dump out the primary partition table [23:53] (or whatever the drive name is) [23:53] then it can be dumped into a file and used to test parted [23:54] it ought not to cause d-i to fail. We'd need /var/log/syslog and /var/log/partman (as usual ...) to see why. I assume that libparted is raising an exception where sfdisk just prints a warning [23:54] http://dev.personnelware.com/carl/temp/Sep22/a/dhcp243/ logs [23:55] but I may have poluted them trying to figure out where the partitions were... I can restart the install easy enough [23:55] Sep 22 17:17:25 kernel: [ 79.381737] attempt to access beyond end of device [23:55] Sep 22 17:17:25 kernel: [ 79.381772] sda: rw=0, want=58635120, limit=58633344 [23:55] Sep 22 17:17:25 partman: [23:55] looks kinda suspicious ... [23:55] Sep 22 17:17:25 partman: /dev/sda5: read failed after 0 of 512 at 2933850112: Input/output error [23:55] ~ # od -tx1 -Ax -N512 /dev/sda [23:55] od: invalid option -- 't' [23:56] im in installer's BusyBox [23:56] proper od [23:56] (e.g. from live CD) [23:57] the one with -t is called od_bloaty.c in busybox so I'd rather not turn it on :) [23:57] heh [23:57] I thnk I have an 8.4 cd around here somewhere... [23:58] Sep 22 17:17:16 kernel: [ 70.857180] sda: p5 exceeds device capacity [23:58] earlier on way before parted gets a look-in [23:59] note that sfdisk says your drive has 3649 cylinders and sda2 apparently ends at cylinder 3877 ... [23:59] so not even just an off-by-$small error