| === JanC [n=janc@lugwv/member/JanC] has joined #ubuntu-boot | ||
| === archis [n=archis@unaffiliated/archis] has joined #ubuntu-boot | ||
| === archis [n=archis@unaffiliated/archis] has left #ubuntu-boot ["Ex-Chat"] | ||
| === penguin42 [n=dg@tu006.demon.co.uk] has joined #ubuntu-boot | ||
| penguin42 | hmm pretty empty | 09:22 |
|---|---|---|
| penguin42 | I've got a 6 minute hang during edgy boot - is there anything I can do to get the boot process to spit out any diags? | 09:23 |
| penguin42 | (Yes minute!) | 09:23 |
| makx | remove quiet from grub boot args | 09:24 |
| makx | also splash | 09:24 |
| penguin42 | already done that ; it shows nothing during the pause that happens just after detecting a usb flash card reader and between starting RAID - and it still happens if I remove the flash card reader | 09:25 |
| penguin42 | I've got a sysrq-t in the bug report I filed: https://launchpad.net/distros/ubuntu/+bug/68888 | 09:26 |
| makx | hmm seems more like an mdadm bug than a kernel one | 09:27 |
| makx | but you might want to ask on a week day on #ubuntu-kernel | 09:27 |
| makx | also you could add a set -x to the mdadm boot script and see what's taking so much time | 09:28 |
| penguin42 | makx: I'm confused how far the boot has got at that point - am I out of initrd at that point? | 09:29 |
| makx | can't tell from your snipps | 09:29 |
| makx | if you see a call to init before the pause yes | 09:30 |
| makx | Running /scripts/init-bottom is the last message from initramfs | 09:31 |
| penguin42 | I'm not sure I do - but let me see what I've got in the logs | 09:34 |
| penguin42 | makx: I don't see anything like that in the /var/log/messages - and that trace still shows a lot of busybox calls which make me wonder if its still in the initrd | 09:38 |
| makx | the messages from the initramfs don't land in /var/log/messages | 09:44 |
| makx | what i told you can see it on boot :) | 09:44 |
| penguin42 | yeh but only if I'm quick and watch very carefully! | 09:44 |
| penguin42 | but what about the fact that busybox is still running - that surely sounds like it is the initrd? | 09:45 |
| makx | busybox is not running it's not a damon, what do you mean? | 09:48 |
| makx | s/damon/daemon/ | 09:48 |
| penguin42 | makx: Look at my sysrq-t trace - it shows some busybox processes running | 09:49 |
| penguin42 | I assume that busybox is used by the initrd but nothing else? | 09:49 |
| makx | yes it's inside | 09:49 |
| penguin42 | ok, are there any good ways of debugging a bad initrd or is it just a case of scattering echo's around the scripts? | 09:50 |
| makx | you should have an mdadm script in /usr/share/initramfs-tools/scripts/local-top/ just add an set -x and see what is the troubeling cmd | 09:51 |
| makx | then the next time you can boot with break=top and try to issue the same to double check.. | 09:52 |
| makx | ah when you change that mdadm script don't forget to rebuild initramfs with update-initramfs -u | 09:53 |
| penguin42 | ok, I think I can see it possibly - there is a thing called 'slumber' in local-top/md and its something to do with sleeping but I can't quit esee what its trying to do | 09:53 |
| penguin42 | makx: Can you see that script - can you see htf it isn't supposed to wait for a few minutes? | 09:55 |
| makx | the slumber is waiting for your devices to appear afais | 09:56 |
| penguin42 | ok - but how does it get out of that loop when it does? I can't see anything that breaks out of the while | 09:56 |
| makx | are you booting with lilo? | 09:56 |
| penguin42 | no | 09:56 |
| makx | ok | 09:56 |
| === penguin42 can't see how that script is supposed to work quickly | ||
| makx | +while [ "$configureduuids" ] | 09:58 |
| makx | this seems the wrong way | 09:58 |
| penguin42 | yeh there is something then which removes the uuid's if it found it | 09:59 |
| penguin42 | and that script provides *no* feedback for something that could wait for 10 minutes | 10:00 |
| makx | i need sleep, pls add feedback to bug report :) | 10:01 |
| makx | only a guest here.. | 10:01 |
| penguin42 | hey thanks :-) | 10:01 |
| penguin42 | hmm I think I know why this happens - that script assumes the UUID= is the last thing on the line in the mdadm.conf | 10:02 |
| === penguin42 sets the package as mdadm and suggests a fix | ||
| === penguin42 goes | ||
| === penguin42 [n=dg@tu006.demon.co.uk] has left #ubuntu-boot ["Leaving"] | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!