Problem:
Another thing which went wrong porting my Debian etch installation to a VirtualBox image is grub not handling 256 byte inode size of ext3 filesystems. I used a Debian lenny ISO to partition the virtual HDD, so per default it was created with an inode size of 256 bytes. Then I dumped the data. Previously, Debian etch used standard 128 byte inode size. I noticed this when booting the ported image:
Searching for GRUB installation directory … found: /boot/grub
The file /boot/grub/stage1 not read correctly.
Solution:
This is a known bug in older Debian installations. grub 0.97 cannot support inode sizes greater than 256 bytes. So I had to either recreate the partition with proper settings or update grub. I ended up with upgrading Debian to lenny, and everything is working smoothly, at least for a while.
Problem:
Migrating my good old Debian etch installation to VirtualBox guest I stumbled across a resume script error on boot:
resume: could not stat the resume device file
Hitting Enter boots normally, ignoring the missing resume image, but it’s a no-go for production use.
Solution:
The problem occurs when the userspace suspend cannot find the swap partition due to configuration errors. In my case it was quite straightforward as the migration process involved some partitions being merged and other kind of disk juggling.
Fixing methods in this case are multifold:
- remove uswsusp package (that’s what I’ve done – with VirtualBox savestate, guest hibernation is a bit redundant)
or
- correct the /etc/uswsusp.conf file – make sure the ‘resume device = /dev/hdX’ line is correct
- dpkg-reconfigure uswsusp
If you have problems with swap partitions not being found – try using labels (read more here). One more thing to note – uswsusp doesn’t handle swap files graciously, so if you don’t have a real swap partition (I use swap files for my VirtualBox guests) removing it can be the easiest way.
Recent Comments