[00:19] I just did an "guided LVM setup" on a 3TB RAID6 array, with 2GB of RAM. [00:20] It tried to create a *two TERAbyte* swap partition! [00:20] cjwatson: ping, are you awake by any chance? [00:26] This was with d-i, not ubiquity, of course. Using the 8.04.2 Server CD. [06:35] ubiquity: superm1 * r3013 ubiquity/ (7 files in 5 dirs): [06:35] ubiquity: merge w/ mythbuntu-ubiquity. [06:35] ubiquity: Remove page skipping, inspect.stack and other hack filled code. [06:38] cjwatson, ^ that should be a majority of the "weird" code gone now in mythbuntu_ui. [08:10] ubiquity: superm1 * r3014 ubiquity/ (11 files in 5 dirs): [08:10] ubiquity: - Just ask all questions instead. [08:10] ubiquity: - Remove themes and plugins pages and all associated debconf templates. [09:23] superm1: cool, thanks [10:38] debian-installer: cjwatson * r1038 ubuntu/ (4 files in 2 dirs): Move mainline architectures to 2.6.28-7 kernels. [10:40] debian-installer: cjwatson * r1039 ubuntu/debian/changelog: releasing version 20081029ubuntu16 [13:31] ubiquity: evand * r3015 ubiquity/ubiquity/frontend/gtk_ui.py: Handle a non-existant selection in the timezone city drop down box. [13:35] cjwatson, evand: Would you guys be comfortable with me filing bugs about stuff that doesn't fit on say, a Q1U, per the UDS discussion with evand about it. [13:35] StevenK: Absolutely [13:36] though do note the UI is still in a bit of flux at the moment, and I have yet to make a concerted effort to reduce the space taken up by new UI elements [13:36] but bug reports are always welcome [13:41] What hardware detection does d-i do that impacts the install? (target disk size for sure, laptop-mode or something?) [13:43] hard to summarise briefly, best to look at the hw-detect source package for most of it [13:43] thanks [13:43] laptop-mode yes, detects processor and installs libc6-i686, may install packages for certain other types of hardware, etc. [13:52] im pretty sure the answer is 'no' but to be sure: does the installer alter initrd? [13:54] Trying to figure out how portable a drive is once the install is done. I know I can move it to just about anything, within reason (don't think I can move x86 to ppc) [13:56] ubiquity: evand * r3016 ubiquity/ubiquity/timezone_map.py: Iterate through the closest five points to the mouse on consecutive clicks from the same location in the timezone map. [14:01] CarlFK: err. the installer creates the initrd; the contents of the initrd depend on the set of packages you have installed [14:01] CarlFK: it is generally a design goal that it should be portable, but there may be bugs [14:05] cjwatson: FYI I verified your size changes to d-i and these are the same I would have done [14:05] (for NSLU2) [14:06] cjwatson: There's only a weird change to a comment: you changed "size of partition - 16" in a comment to "size of partition - 21" [14:06] cjwatson: But it's really "- 16" (and that's the value you used) [14:09] - util/pad $(TEMP)/initrd.gz.nslu2 6291440 # size of partition - 16 for header [14:09] + util/pad $(TEMP)/initrd.gz.nslu2 5636080 # size of partition - 21 for header [14:09] Yes [14:09] I thought "size of partition" was the whole thing, and thus that comment would need to be changed in step with the actual value? [14:09] cjwatson: Size of partition is the whole thing [14:09] Sorry [14:09] whole thing> kernel plus initramfs [14:09] The whole thing is size of partition - 16 [14:10] one of us is confused [14:10] cjwatson: The partitions start with a 16 bytes header [14:10] oh [14:10] I thought it meant 16 blocks [14:10] And are multiple of blocks (131072 bytes) [14:11] cjwatson: Your size adjustment is exactly 5 blocks [14:11] right, that's why I changed the comment because I thought it meant 16 blocks [14:11] 43*131072 - 16 == 5636080 [14:12] debian-installer: cjwatson * r1040 ubuntu/ (build/config/armel/ixp4xx/netboot.cfg debian/changelog): [14:12] debian-installer: Fix my confused comment change in nslu2 initrd padding code (thanks, [14:12] debian-installer: Loïc Minier). [14:12] Thanksely [14:12] cjwatson: No I have no idea why it wont work, but we'll rebuild the kernel with the right endianess first (flipping endianess manually didn't help) [14:12] *Now [14:13] discussion is on #ubuntu-arm, but I'm just keeping you up-to-date [14:13] ogra is chasing this; I don't have a NSLU2 so I can't [14:16] cjwatson: design goal w/bugs <- perfect [14:17] CarlFK: I think normally the main changes in the initrd will be to do with partitioning rather than hardware; for example if you install with LVM then you get LVM bits in your initrd, etc. [14:17] there may be exceptions that escape my memory [14:17] perhaps the location of the swap partition, if it isn't UUID-ified [14:18] cjwatson: "design goal" is more than enough [14:19] I am looking at a 300 line bash script that does an install: debootstrap sarge /root/$ROOTFSDIR $MIRROR > /tmp/debootstrap.out 2>&1 & [14:19] *backgrounding* debootstrap? [14:19] I suppose it might then go on to parse the .out for progress information ... [14:20] while [ -e /proc/$PID ] do echo -n '.' sleep 5 [14:20] and give me a headache ... [14:20] all this is going away. d-i + preseed = install happiness. [14:26] ubiquity: evand * r3017 ubiquity/ubiquity/frontend/gtk_ui.py: Properly split the timezone names. [14:37] evand: will the map for kubuntu remain as is or will that change to the newer map too? [14:38] as is for the foreseeable future [14:39] okay cool [14:46] cjwatson: how many hours have you spent working towards the "design goal that it should be portable" ? [14:46] * CarlFK tries to keep a straight face... [15:02] we got it right from the start O:-) [15:10] I'm planning on releasing a new ubiquity in roughly two hours. If anyone still has things they want to check in and don't think they'll make it by then, please let me know and I'll wait. [16:40] ubiquity: evand * r3018 ubiquity/debian/po/ (79 files): debconf-updatepo [16:41] ubiquity: evand * r3019 ubiquity/ (d-i/manifest debian/changelog): [16:41] ubiquity: Automatic update of included source packages: partman-auto [16:41] ubiquity: 83ubuntu3, partman-base 128ubuntu6, partman-partitioning 64ubuntu3, [16:41] ubiquity: partman-target 58ubuntu3, user-setup 1.23ubuntu10. === cjwatson_ is now known as cjwatson [17:22] ubiquity: evand * r3020 ubiquity/debian/changelog: releasing version 1.11.7 [23:09] cjwatson: Have you heard about regressions in Hardy where d-i during partioning when using preseed ? (something to do with the partition file system) ? I have been reported that by two different people at the office today (I plan to have a look at it tomorrow) [23:10] stgraber: I'm not sure I can parse that let alone say whether I've heard about it :) [23:10] stgraber: it doesn't sound familiar, no ... [23:10] a "fail" was miissing in the first sentence :) [23:11] "partition file system" EPARSE [23:11] anyway, even if I make guesses it's not familiar to me, so feel free to send over more details when you have them and I can take a look [23:34] cjwatson: "No root file system is defined" with http://www.stgraber.org/download/test.seed (light version of our preseed for debuging) [23:37] http://www.stgraber.org/download/ltsp/ltsp-cluster-raid.seed (OpenVZ + RAID1) does the same (and I'm 90% sure we used it a few months ago to install our new servers ...) [23:51] evand, http://pastebin.com/f1789919d [23:51] that's coming up with the latest ubiquity upload. where is rsvg supposed to come from? [23:56] looks like python-gnome2-desktop i think. that's troublesome for non gnome