/srv/irclogs.ubuntu.com/2007/12/21/#ubuntu-installer.txt

xivulonhttps://code.launchpad.net/~ago/partman-auto-loop/lupin-support00:29
TorgotonHey all. I managed to get Ubuntu to start to install on my ancient laptop. Gutsy kernel crashed every time, but Dapper started the process ok, if VERY slow.01:15
TorgotonWould that be an installer issue, or a kernel issue?01:16
CarlFKTorgoton: "start to install" = installing stuff to HD or just booting the installer ?01:19
Torgotonjust booting the installer, CarlFK.01:19
Torgotonthat is, starting the installer crashes with Gutsy Mini ISO/netboot files (linux and initrd.gz) but the Dapper version starts fine.01:21
CarlFKwhat about the 'check cd" ?  (which also boots a kernel01:21
CarlFKah, that kinda boot01:21
TorgotonYeah. This is a very old 33MHz 486 with 36MB RAM. No CD, no USB.01:22
CarlFKsounding like an installer01:22
CarlFKwow01:22
Torgotonyeah!01:22
CarlFKpretty sure 48 is the minimum01:22
Torgotonwow is an understatement. "Installing the base system" has been going for something like 5 hours.01:23
TorgotonThere's a page that says 32MB is the minimum... looking for it....01:23
CarlFKim surprised the HD is big enough :)01:23
Torgotonheh. The drive that came with the machine is 170MB, but it now has a 5GB drive.01:24
TorgotonHere it is: https://help.ubuntu.com/community/Installation/SystemRequirements01:24
TorgotonThe end has Absolute minimum requirements: 486, 32MB RAM, 300MB disk.01:25
CarlFKum, thats not what I see...01:25
CarlFKoh wait.. that end01:25
Torgotonyea.01:25
CarlFKWhats a "graphical installation" ?01:26
TorgotonI imagine that's one including the X Window System.01:26
Torgotonor maybe... graphical installation.01:26
CarlFK300 isn't enough, cuz 350 gets used during the install01:27
TorgotonYeah. That must be it. I'm doing a text install. Over a serial port, actually.01:27
CarlFKI have a feeling the 32m is out of date too01:27
TorgotonWould you like to see the debug dump from my attempt with Gutsy?01:27
CarlFKheh - I have a 32m stick I have been using to debug a cpu speed problem... just a sec01:28
CarlFKsure01:28
Torgotonhttp://paste.ubuntu-nl.org/48887/01:28
cjwatsonkernel crash => kernel issue01:33
cjwatsonyou're still using the generic kernel though01:34
CarlFKwill now that you put it that way :)01:34
cjwatsonyou sure you're using the 386 build I pointed you to? that should be built against the -386 kernel, not -generic01:34
TorgotonThat is using the two files from the Gutsy Mini ISO.01:34
cjwatsonthere are multiple gutsy mini.iso builds01:34
Torgotonaha!01:35
cjwatsonin /msg, I pointed you to http://archive.ubuntu.com/ubuntu/dists/gutsy/main/installer-i386/current/images/netboot/386/mini.iso01:35
cjwatsonthe /386 in there is significant - you may have missed it01:35
Torgotonyes you did. I thought I tried that one as well, but will try again to be triply sure. <sigh> Time to stop the current install... five hours into it. :)01:35
cjwatsonyou're definitely using something built with the -generic kernel - either netboot/386/mini.iso is misbuilt, or you picked the wrong one :)01:36
cjwatson(not saying the former ain't possible)01:36
cjwatsonfor the record, I think 36MB RAM should work fine in gutsy, though you may get kicked into lowmem mode01:37
cjwatsonshould actually be rather faster than dapper, at least the installer side of things01:37
cjwatsonwe did some serious memory work in d-i over the summer, which shaved off 16MB+01:37
CarlFKTorgoton: what media are you booting from?01:37
Torgotonbooting from hard drive, using Lilo to load the image. Woody is installed on the machine.01:38
CarlFKneat :)01:38
Torgotonthanks.01:39
Torgotonok it's going to take a few minutes to set up.01:39
CarlFKyou might want to try kexec01:39
CarlFKwget linux/initrd.gz, use kexec to boot them01:40
cjwatsonkexec from a woody kernel?01:40
cjwatsonprobably won't work so well01:40
CarlFKwhy would it matter?01:40
TorgotonI've never heard of kexec before.01:41
CarlFKassuming it works at all..01:41
Torgotonholy crap! X started!01:42
Torgotonit never worked before, and it's not pretty. 320x200x1.01:42
cjwatsonCarlFK: the kexec syscalls didn't exist in the woody kernel01:43
cjwatsonkexec is a relatively recent thing01:43
CarlFKhttp://dpaste.com/28713/01:43
cjwatsonas regards disk space, a debootstrap of gutsy takes up 210MB or so01:43
CarlFKah, so much for that idea01:43
cjwatsonthe standard task is another 51MB01:44
Torgotongetting the *386* mini iso...01:45
cjwatsonkexec was merged into Linux in 2.6.1301:46
cjwatsonwoody had, er ...01:46
Torgoton2.2.2001:46
cjwatsonor 2.4.1801:47
cjwatsonbut either way, certainly no dice01:47
Torgotonso should I use lilo or try kexec?01:48
cjwatsondon't bother with kexec, it won't work01:48
Torgotonok. booting...01:48
Torgotoncrash01:50
Torgotonhttp://paste.ubuntu-nl.org/49098/01:51
cjwatsonok, it's at least the right kernel this time01:51
Torgotonbaby steps...01:52
cjwatsonit's having trouble trying to set up the right MMU bits for your machine01:52
cjwatsonalternative_instructions is pretty fundamental01:52
TorgotonThe CPU is a 486SL, if it matters.01:52
TorgotonI think... :)01:52
cjwatsonthis is way out of my field, and squarely in the kernel team's turf01:53
TorgotonYep, 33MHz 486SL01:53
TorgotonHm. Would it still be an Ubuntu issue?01:53
cjwatsoncould be, since it seems to be in the paravirt_ops stuff01:53
cjwatsonwhich IIRC is something we took as an external patch, though I may misremember and it could be it's mainline now01:54
cjwatsonsubject to timezone and holiday vagaries, #ubuntu-kernel should be able to help more01:54
cjwatsonand speaking of timezones, I have to go to bed now01:54
cjwatsongood luck :)01:54
TorgotonThanks so much, cj and carl.01:55
CarlFKqemu with 32m says it isn't enough01:57
CarlFK38 is ok, but low mem mode02:00
banyaohello06:57
banyaoI am new06:58
banyao  8-)06:58
xivuloncjwatson, I have added a banch to partman-auto-loop, I hope that simplifies handling of a couple of patches11:04
xivulonI also need to patch partman/commit.d/*format_* since in case of loopfiles they format the containing device as opposed of the loopfile11:06
xivulonthat though seems scattered around11:07
cjwatsonyeah11:08
cjwatsonthanks for the branch11:09
xivulonwhat is a good place to submit such patch(es)?11:10
xivuloncommit.d/50format_ext3 for instance11:11
xivulonin partman_auto_loop I still have not figured out a good way to skip disk-free-space checks, it's left as TBD11:12
cjwatsonbranches intended to be merged should usually have an associated bug that refers to the branch11:13
xivulonIs that "propose for merging"?11:14
xivulonand/or  link to bug report?11:16
xivulonI'll use both11:16
xivulondone11:17
xivuloncjwatson where should I add a branch for patches against update_grub and /lib/partman/commit.d/50format_*11:25
* xivulon likes LP11:26
cjwatsonupdate-grub is in the grub package11:27
cjwatsonit is not in bzr so you cannot add a branch11:27
cjwatsonuse diff -u and attach the output11:27
cjwatsonthere is no one branch you can use for commit.d/50format_*11:28
cjwatsonseveral different packages are involved11:28
xivulonI have diff at the moment plus patched file to be diffed (forgot to do diff -u yesterday)11:28
cjwatsonyou don't really need to show patches for all of them though, they're all the same and I can just duplicate it across11:28
xivulonwhat's a good place to submit such patches?11:28
cjwatsonpick one of the packages11:29
xivulonok11:29
xivulonI can add a branch in https://code.launchpad.net/partman-ext3/11:32
xivulonbug 17786811:43
ubotuLaunchpad bug 177868 in partman-ext3 "When loopfiles are used mkfs has to target the file and not the containing device" [Undecided,New] https://launchpad.net/bugs/17786811:43
xivuloncjwatson can you please remind evand to follow up with https://bugs.launchpad.net/wubi/+bugs?field.tag=wubi ?12:07
cjwatsonno, because he's on holiday until the new year12:10
xivulonaha I tried to contact him yesterday night12:10
xivulonnp12:10
xivulonps yesterday I had this error in grub-installer, did not have time to dig further, does the syslog ring any bell?12:10
xivulonhttp://paste.ubuntu-nl.org/49146/12:11
cjwatsonno, sorry, there doesn't seem to be a lot to go on there12:13
xivulonnp, I'll do new runs in more verbose mode, that happened late yesterday and had a hard time keeping the eyelids up12:14
xivulonhave to go now, have a nice holiday everyone!12:15
=== cjwatson_ is now known as cjwatson
CarlFKhow can I pass this as a kernel parameter? tasksel tasksel/first multiselect ubuntu-standard, ubuntu-desktop14:24
cjwatsonI'm not convinced you can14:25
CarlFKis this worth trying: tasksel/first=ubuntu-standard,ubuntu-desktop14:25
cjwatsonnot without bug 174557 being fixed14:25
ubotuLaunchpad bug 174557 in tasksel "tasksel/first preseed syntax is inconvenient" [Wishlist,Triaged] https://launchpad.net/bugs/17455714:25
cjwatsonno, that won't work unfortunately14:25
cjwatsoncan't you use a preseed file?14:26
CarlFKyes, but I was trying to only have 1 file14:26
cjwatson(it would have to be tasksel:tasksel/first=blah anyway, but still won't work due to tasksel splitting the value on ", "14:26
cjwatson)14:26
CarlFKi think this worked: tasksel:tasksel/first="ubuntu-standard, ubuntu-desktop"15:17
CarlFKDec 21 14:33:26 frontend: --> SET tasksel/first ubuntu-standard, ubuntu-desktop15:17
CIA-22netcfg: cjwatson * r618 ubuntu/ (base-installer debian/changelog netcfg-common.c):15:54
CIA-22netcfg: * Add an /etc/network/interfaces stanza for lo even if the netcfg menu15:54
CIA-22netcfg:  item wasn't run (LP: #9532).15:54
CIA-22netcfg: cjwatson * r619 ubuntu/debian/changelog: releasing version 1.40ubuntu315:56
=== ebel_ is now known as ebel
=== ebel_ is now known as ebel

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!