Boot fails during fschk of /dev/vg1/var
I am running Lenny with LVM. Last night the system locked completely.
No mouse movement. No keystroke entry. Nothing. I don't like to
power down without a proper shutdown, but I did not seem to have a lot
of choice. This morning, when I tried to boot up I got an fschk on
several partitions. The first few were fine, until I got to
/dev/vg1/var. I got a dma expiry error. I tried several times to
power down and reboot, with the same result.
Next, I booted into single user and tried to fsck /dev/vg1/var. When
I do that, I get the following messages:
/dev/vg1/var has gone 186 days without being checked, check forced.
Pass 1: Checking inodes, blocks, and sizes
[ 191.648009] hda: ide_dma_sff_timer_expiry: DMA status (0x61)
[ 201.648005] hda DMA timeout error
[ 201.648057] hda: dma ti9meout error: status=0xd0 { Busy }
[ 201,648173] hda: possibly failed opcode: 0xc8
[ 201.648224] hda: DMA disabled
[ 201.512005] ide0: reset: success
Then it just sits there indefinitely. Total lock up.
I can reboot, go into single user mode, and mount the partition
without any apparent errors. Only a warning that it has been too long
since it was checked and I SHOULD run fsck. How do I fix this?
Presumably, I could edit /etc/fstab to not mount /var automatically
and mount it manually once the system has booted, but that does not
correct the actual problem. Okay, that doesn't work. The root
filesystem is read-only. Can I remount / as rw while the system is
up? Do I need to get a rescue cd and make the changes that way? I
think that I have enough unused space in my pv to create a new
partition for /var and copy everything from the old partition. Would
this be the best option, or is there something else to do? If this is
my best option, then I guess I need a rescue CD that understands LVM
to boot into and work from. Is this correct?
All help appreciated.
Marc Shapiro
Reply to: