Hey all,

yesterday I went about to get hibernate, or suspend-to-disk (swsusp2) to work. Basically followed the Software Suspend HOWTO (http://www.suspend2.net/HOWTO).

After successfully patching the kernel (2.6.13.2), recompile with filewriter and swapwriter compiled in the kernel and editing the hibernate.conf I gave it a try by typing 'hibernate'. System hibernated but rebooted normal (i.e. without resuming). dmesg complained about not finding the resume2= in lilo.conf so I appended resume2=swap:/dev/hda6 and ran lilo. No problem so far. Reading about possible data loss I copied the init.d/hibernate-cleanup.sh script to /etc/rc.d/rc.hibernate-cleanup (running Slackware-current) and added an entry in /etc/rc.d/rc.S to fire up that script during boot after the file systems are mounted but before swap is mounted.

Then I rebooted, and EVIL struck: immediately after the boot screen (with the manufacturer's brand logo) instead of seeing lilo's menu, I get a black screen with only LI in the left most corner, and there it hangs forever... aaaarrrggghhhhhhhh

I thought cleaning out swap would do the trick (suggested in the howto), so after booting knoppix cd I tried mkswap /dev/hda6 to no avail. Mounting /dev/hda7 (root) and running the hibernate-cleanup script to manually wipe out the swap also didn't help.

Reiserfsck /dev/hda7 (root partition) indicates loads of errors. Will fixing with fsck lead to data loss?

When I inspected lilo.conf, I found that the append="resume2=swap:/dev/hda6" entry was not there anymore.

PLEASE, anyone, what should I do to prevent further harm or data loss, or to further diagnose the exact problem?

Apologies for the long mail, but I am in despair to get my laptop running again without the necessity to reinstall Slack (that's the beauty of linux; never reinstall, it's always fixable.... isn't it?)

Cheers, Leon.

My experience with swsusp is that it works well once correctly setup. Unfortunately, you're not the first one to screw things up while setting it up: I did too

When this happened to me, I went for the safest solution:
- Mount the hard disk read-only.
- Back up all that could be read on disk.
- Unmount and "repair" the filesystem.
- Mounted again, read-write this time.
- See if some more things can be saved.
- Reinstall Linux.

Yves.

Yes, I have backups of the important stuff. After reinstalling lilo at MBR I could again boot up. But then the system prompted me to correct the filesystem with reiserfsck. Am now performing reiserfsck --rebuild-tree and see what that gives.

At the same time I am downloading Slackware-10.2 as I expect a complete reinstall.... Damn, what I love about Linux is that a reinstall isn't needed; everything can be fixed. This time I took things too far I'm afraid

Cheers, Leon.

I reinstalled lilo to MBR which allowed me to boot past the lilo boot menu. Then the boot process stopped at fsck to allow me to correct the file system. I ran reiserfsck, first with --fix-fixable, then with --rebuild-sb and finally went all the way with --rebuild-tree.

Then rebooted and all is well! Saw few messages about files that were damaged/lost but so far everything seems to be a ok!

Cheers, Leon.