[00:07] <TheMuso> evand: Re dmraid and what I was talking about with cjwatson in -devel. I saw partmon-dmraid mentioned over the weekend in my logs, but see nothing in the archive. Is this being worked on/a bzr branch somewhere I can pull to have a look? And whats its current status?
[00:08] <soren> There is dmraid magic in the installer..
[00:09] <soren> You need to pass disk-detect/dmraid/enable=true on the kernel command line to activate it.
[00:09] <TheMuso> Right. So dmraid is part of an existing partmon package?
[00:10] <soren> The interesting bits are in a) dmraid-udeb and b) the disk-detect script in the hw-detect package (udeb).
[00:10] <TheMuso> soren: Right.
[00:11] <TheMuso> I happen to have hardware to test this with, so will get things sorted for doing so later, and I'd also like to integrate it with the initramfs error handling spec stuff, which I can at least make a start on. Thank you both.
[00:12] <cjwatson> dmraid is still in universe, so none of the dmraid stuff will work yet
[00:12] <TheMuso> I'm als assuming one has to modify an alternate image to include dmraid. :)
[00:12] <cjwatson> and partman-dmraid
[00:12] <TheMuso> cjwatson: Yeah I just saw that
[00:13] <TheMuso> but I don't see a partman-dmraid...
[00:13] <TheMuso> oh of course, its in the installer section.
[00:13] <cjwatson> right, we need to get it properly integrated, I asked evand about that on IRC but I don't recall whether he replied (I've lost my logs due to my client going barmy)
[00:13] <TheMuso> Right.
[00:13] <cjwatson> the UI it exposes was pretty mad last I looked, but functional
[00:13] <TheMuso> Ok.
[00:14] <soren> The new iscsi magic isn't exactly pretty either, but it seems to work.
[00:14] <cjwatson> ooh! launchpad supports inter-branch seed inclusion now
[00:15]  * soren tries to turn that into something that makes sense in his head
[00:15]  * soren fails
[00:25] <cjwatson> RN  installer.OTHER => installer.OTHER.OTHER
[00:25] <cjwatson> thanks, bzr
[00:25] <cjwatson> soren: I'll explain via ubuntu-devel-announce shortly-ish
[00:26] <cjwatson> but basically the thing I talked with you the other day about server seeds
[00:27] <soren> I think I get that bit.. I just don't quite get how launchpad fits into that. No worries, I'll just read it on u-d-a. :)
[00:28] <cjwatson> oh, launchpad runs germinate and needed a few tweaks to cope
[00:28] <cjwatson> in order to generate task headers
[00:29] <soren> launchpad runs germinate? Erm... Ok. I don't suppose you just mean the buildd's when building d-i?
[00:29] <soren> Oh, for tasks.
[00:29] <soren> Right, beginning to make sense.
[00:30] <soren> Done.
[00:38] <TheMuso> Woo. After reading http://wiki.debian.org/DebianInstaller/SataRaid, things really do seem to be rather experimental. Do we really want to integrate this yet? I'm of the view that we don't.
[00:38] <TheMuso> Dispite the fact that just about every mother board sold, at least in Australia, supports fakeraid.
[00:55] <cjwatson> I'd like to offer something, even if it's a bit homebrew
[00:55] <cjwatson> various vendors are starting to make desperate noises
[01:05] <TheMuso> Yeah having said that, there are a lot of users who wish to, and do use it for windows installs, as its the only sane way of getting Windows onto a form of RAID without doing something weird like dynamic disks, or expensive controllers.
[03:40] <CIA-24> ubiquity: evand * r2452 ubiquity/ (3 files in 3 dirs):
[03:40] <CIA-24> ubiquity: * Always show the advanced partitioner buttons, greying them out
[03:40] <CIA-24> ubiquity:  conditionally instead of hiding them.
[11:23] <CIA-22> oem-config: cjwatson * r412 oem-config/ (d-i/update-control debian/changelog debian/control): * Build-depend on dctrl-tools rather than grep-dctrl.
[12:18] <CIA-22> ubiquity: cjwatson * r2453 ubiquity/ (d-i/update-control debian/changelog debian/control): * Build-depend on dctrl-tools rather than grep-dctrl.
[14:23] <xivulon> Have done the winfoss replacement (umenu)
[14:23] <xivulon> http://img166.imageshack.us/img166/5553/umenuys5.jpg
[14:23] <xivulon> As discussed with evand and heno
[14:24] <xivulon> Will upload the code tonight to umenu project
[14:24] <xivulon> evand can you pls fwd to Henrik?
[14:24] <evand> xivulon: Fantastic work, thanks a lot!  I'll forward it on to Henrik.
[14:31] <xivulon> thx
[14:36] <evand> cjwatson: provided that heno approves of the above implementation, do you have any objections to me adding this to find-live-filesystem?
[14:37] <cjwatson> not at all, please go ahead
[14:37] <evand> great
[14:49] <soren> cjwatson: Got any good hints about how to debug grub-installer?
[14:50] <soren> cjwatson: Calling it from the commandline in ("grub-installer /target" like it says in the postinst) does not give me any output, but has a 0 exit code, but from the installer menu, it fails.
[14:50] <soren> syslog just says it returned 1.
[14:51] <cjwatson> have you tried the usual 'set -x' and DEBCONF_DEBUG=developer tricks?
[14:51] <cjwatson> i.e. edit 'set -x' into /usr/bin/grub-installer and then run it from the installer main menu
[14:52]  * soren does so
[14:52] <soren> cjwatson: Ah.
[14:52] <soren> cjwatson: Right, that explains..
[14:52] <soren> Thanks!
[14:52] <soren> :)
[14:58] <saispo> hi soren :)
[14:58] <saispo> hi cjwatson too :)
[14:58] <soren> saispo: hi
[15:00] <saispo> soren: you're the virtualization man for Ubuntu ? :)
[15:00] <soren> I am.
[15:00] <saispo> :)
[15:00] <saispo> i open a bug about syncing qemu 0.91 to debian unstable, you will or not ?
[15:02] <cjwatson> soren: what was the problem?
[15:03] <soren> cjwatson: Well, the first part of it was that the virtio block devices are called /dev/vd[a-z], which grub-installer didn't recognize.
[15:03] <soren> Right now I'm trying to figure out what's wrong with the device.map generation bit..
[15:05] <cjwatson> ah
[15:05] <cjwatson> note grub-installer duplicates it from grub :-/
[15:08] <soren> saispo: I'm doing it today.
[15:09] <saispo> ok great ! thanks, i haven't seen it on list :)
[15:22] <soren> Which fd to send stuff to from something like grub-installer for it to show up in syslog?
[15:25] <cjwatson> 2
[15:25] <cjwatson> grub-installer already has log, error, info functions though
[15:26] <cjwatson> you should normally use them
[15:37] <soren> Ah. Good call.
[15:38] <soren> What actually generates the device.map? grub itself?
[15:38] <evand> indeed
[15:39]  * soren sobs
[15:39] <saispo> any git expert ? :)
[15:39]  * evand pats soren on the back
[15:40]  * evand proposes a grub support group
[15:45] <evand> oh, actually
[15:45] <evand> maybe not
[15:45] <xivulon> soren what is the issue?
[15:45] <xivulon> device.map is missing, or is it wrong?
[15:45] <soren> grub not "getting" virtio devices.
[15:46] <xivulon> can't you append them to device.map maybe?
[15:46] <xivulon> device map is generated via grub --device-map
[15:47] <soren> Yes, I've gathered by now.
[15:47] <xivulon> which in turns is called by grub-install which is called by grub-installer
[15:47] <soren> And the device.map generation doesn't know how to handle virtio devices.
[23:10] <soren> cjwatson: Still around? Can you in a few words explain why grub-installer replicates so much of grub-install's logic? Why not just call grub-install?
[23:41] <CIA-22> grub-installer: soren * r726 grub-installer/ (debian/changelog grub-installer):
[23:41] <CIA-22> grub-installer: * Teach grub-installer about virtio devices (/dev/vd*).
[23:41] <CIA-22> grub-installer: * Release 1.27ubuntu4.