| ev | mpt: http://people.canonical.com/~evand/screenshots/ubiquity/12.04/ | 11:05 |
|---|---|---|
| mpt | woohoo | 11:06 |
| ogra_ | cjwatson, so we still havent properly solved the flash-kernel issue with live-installer (running update-initramfs from console-setup) ... i initially did the diversion from the flash-kernel-installer udeb but that didnt end up in the live-installer.d hooks properly ... on a second look infinity found that setup_dev is actually commented in the code of live-installer (no /dev/mmcblk0 is what makes f-k fail actually) ... | 15:42 |
| ogra_ | cjwatson, do you see a reason to not enable setup_dev (which would also make the extra bindmounting of /dev for grub unneeded) | 15:42 |
| * ogra_ cant really find a reason in the changelog why it was disabled in the first place | 15:44 | |
| cjwatson | Did you do exactly what I said to do in the last conversation we had, or something slightly different? | 15:49 |
| cjwatson | (Do you have a date? I need to refer to it, but don't have convenient IRC logs at the moment.) | 15:49 |
| cjwatson | I'm not keen on messing about with bind-mount handling without great care and unless there's no alternative. That kind of thing is always problematic. | 15:50 |
| ogra_ | seemingly something slightly different, you suggested a diversion of update-initramfs ... which i did from flash-kernel-installer (which didnt work since the udeb is only installed later) it would have to go into live installer | 15:50 |
| cjwatson | That doesn't make sense. | 15:51 |
| infinity | Sure it does. | 15:51 |
| cjwatson | udebs are nearly all installed well before base system installation, and I'm not aware of any special arrangements to anna-install f-k-i later. | 15:51 |
| infinity | flash-kernel is in the squashfs, it has a hook in update-initramfs.d, so anything that triggers update-initramfs breaks. | 15:52 |
| cjwatson | I understand that, but that's not what I'm disagreeing with. | 15:52 |
| infinity | Bootloader-installer packages are on-demand, no? | 15:52 |
| cjwatson | The postinst is run on demand, and apt-installs the deb (not the udeb). | 15:52 |
| ogra_ | i created post-base-installer.d/10flash-kernel-installer in the f-k-i udeb | 15:52 |
| ogra_ | but that didnt end up in the filesystem at live-installer runtime | 15:52 |
| cjwatson | Can I see the diff for this? | 15:53 |
| cjwatson | It doesn't seem to be in the archive. | 15:53 |
| cjwatson | Ah, here we go. | 15:54 |
| ogra_ | flash-kernel-3.0~rc.4ubuntu8 | 15:54 |
| cjwatson | As I thought. You installed the p-b-i script to the wrong path. | 15:55 |
| cjwatson | Download https://launchpad.net/ubuntu/+source/flash-kernel/3.0~rc.4ubuntu11/+build/3678592/+files/flash-kernel-installer_3.0%7Erc.4ubuntu11_armhf.udeb and run 'dpkg -c' | 15:55 |
| cjwatson | -rwxr-xr-x root/root 205 2012-07-24 14:27 ./usr/lib/post-base-installer.d/post-base-installer.d/01flash-kernel-diversion | 15:55 |
| ogra_ | eek | 15:55 |
| ogra_ | oh my, silly me | 15:55 |
| cjwatson | The .install file should've been "post-base-installer.d usr/lib", not "post-base-installer.d usr/lib/post-base-installer.d" | 15:55 |
| ogra_ | but i'm 100% sure there was nothing else than 25live-installer-console-setup in my filesystem | 15:56 |
| ogra_ | i should have seen that subdir in this case | 15:56 |
| cjwatson | Can I see a log of a current installation? | 15:56 |
| * ogra_ hasnt done one in a week, and we work around the issue with a hack that makes f-k exit 0 | 15:57 | |
| cjwatson | It doesn't matter whether it's been worked around. | 15:57 |
| cjwatson | I'm looking for something else. Actually any d-i syslog from the last few months would be fine for what I'm looking for. | 15:57 |
| cjwatson | Well, on the right arch/subarch obviously. | 15:58 |
| * ogra_ can surely find an unrelated bug with a syslog, one sec | 15:58 | |
| ogra_ | hmpf, i only have ubiquity syslogs in the bugs i find | 16:00 |
| ogra_ | gimme 30min and i'll produce a fresh one | 16:02 |
| ogra_ | hmm, or not ... seems i cant proceed beyond locale selection ... serial console stays quiet | 16:14 |
| ogra_ | (weird since our test installs worked in bostaon last week) | 16:15 |
| ogra_ | ah, crap seems i did hit bug 1028905 once again | 16:19 |
| ubot2 | Launchpad bug 1028905 in cdrom-detect "cdrom-detect in quantal omap4 hangs trying to look for install media on an extended partition" [High,Confirmed] https://launchpad.net/bugs/1028905 | 16:19 |
| ogra_ | moving on now with a freshly wiped disk | 16:19 |
| ogra_ | cjwatson, phew, took a bit longer ... http://paste.ubuntu.com/1134597/ | 17:01 |
| cjwatson | OK, that shows what I thought. flash-kernel-installer is retrieved and unpacked at the 'load-cdrom' stage, well before live-installer runs. You can see that by searching for flash-kernel-installer - it's the first hit. | 17:02 |
| cjwatson | So I think you should just try restoring your previous code but fixing the hook location. | 17:03 |
| cjwatson | And if that doesn't work we should debug from there. | 17:03 |
| ogra_ | ok | 17:04 |
| infinity | ogra_: And back out the other hack while you're in there, then. | 17:07 |
| ogra_ | +++ | 17:07 |
| ogra_ | and + | 17:07 |
| * infinity really runs off now. | 17:08 | |
| xnox | cjwatson: am I ok to merge and commit these two branches: | 21:35 |
| xnox | https://code.launchpad.net/~dmitrij.ledkov/ubiquity/crypto-done/+merge/117664 | 21:35 |
| xnox | https://code.launchpad.net/~dmitrij.ledkov/partman-crypto/finish.d/+merge/117657 | 21:35 |
| xnox | cause now lvm bugs are filed, but there are bugfixes in the crypto-done branch.... | 21:36 |
| xnox | e.g. bug 1033985 | 21:36 |
| ubot2 | Launchpad bug 1033985 in ubiquity "unable to install quantal on system running precise with LVMs" [Undecided,Confirmed] https://launchpad.net/bugs/1033985 | 21:36 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!