messed up: swsusp doesn't let me boot my system anymore

[pjv]

I kinda seem to have messed up good this time :
I was trying out swsusp for suspend-to-ram on my 2.6.1-gentoo-r1 laptop. I did /bin/hibernate (using that script). It started to suspend and then crashed with an error that said I should run /sbin/depscan.sh. I manually rebooted and then found out that my system just wouldn't boot anymore. The first time it just froze after the "Generating service dependencies ... [ ]"-line (I'm not too sure about this). On all subsequent manual reboots it stops on the line before that: "Activating (possibly) more swap [ok]". I've tried booting with: resume=/dev/hda7(myswap) and noresume. I used the LiveCD and reset the swap partition (mkswap). I also tried booting with: noresume2 and resume2=swap:/dev/hda7. (Why o why did I only now read about these resume2 settings?? They were thus not used on first reboot.) No luck at all! Sometimes it just hangs and sometimes it reboots itself. Seems most like the problem listed in the faq of the swsusp homepage, unfortunately the noresume2 solution does not solve it for me!
Can anyone plz help me out? For now I'd just like to get my system back, I miss it so much...

[dvc5]

Sounds like your boot image might have been corrupted. You might wanna try backing up your /boot/ partition and using the livecd to chroot in and install a new kernel.

-Loz

[pjv]

I've tried booting with other (and older) kernels like 2.6.1-vanilla and 2.4.22-ac4 and older. They all stop at the same line (and then sometimes reboot). So I don't guess it is kernel-related. Maybe it has something to do with a corrupted swap partition? Or does it get written in a file on the harddrive that it needs to resume next time? Any chance I can delete that?

For now my problem most ressembles: http://swsusp.sourceforge.net/FAQ-5.html#ss5.3

Thx in advance

[dvc5]

If you're able to use the livecd to chroot into your install, your root partition should be okay. Since you've reformatted your swap (mkswap), the resume image should be erased from the disk. AFAIK, nothing gets written to your root paritition having to do with hibernation/suspend. Have you tried booting without any resume flags in your kernel line? Maybe chroot into your install using the livecd, compile a test kernel w/out acpi support, and see if that will boot.

-Loz

[pjv]

I've chrootted with my LiveCD now and it has shown a problematic root partition. I've used reiserfsck to fix it (had to rebuild the table). So now I'm left with corrupted data.
It's not looking very good . Guess, I'll now have to rebuild the whole system (maybe even bootstrap)?! Jeezes, things are really messed up!

[dvc5]

Ouch, sorry about your data. I dunno if you've already nuked your drive, but if you check in the /lost+found/ folder, you might be able to recover some of your data. I was able to recover my /home and /etc folders by searching around in there. Good luck.
_________________
#define NULL rand() /*heh heh heh */

Green Is Good

[pjv]

Well ... I'll think twice next time I read the words "dangerous" .

I'm still hoping to get it back in an easy way. Is there any way to get your system files back without need for a rebuild from scratch? Could one perform a bootstrap without too much influence to the rest of the system?

Thanx for the tips so far!

[dvc5]

Depending on how borked your other system files are, running bootstrap *shouldn't* hurt anything. However, if you've updated packages like glibc, gcc, or binutils and had problems in the past, running bootstrap might cause some unforseen library/linking problems. When my drive crashed a while ago, I wound up running "emerge -e world" which rebuilds every package in the system. 24 hours later, I was back up and running. I would recommend trying bootstrap, but make sure you backup your /etc /boot and /home etc. directories in case you need to do a clean sweep and start from scratch. Good luck.
_________________
#define NULL rand() /*heh heh heh */

Green Is Good

[pjv]

Thx for your advice