'Aborting Journal' errors on Ext3

[RumbleFish]

Hey all,
I'm trying to track down some details on a problem that I already have a work around for. But am curious as to why it is happening.

I'm running Debian 4.0 on a Toshiba Tecra 8100 laptop. I have an external USB 500GB hard drive that has 1 500GB partition formated ext3.

Because I don't want the external to spin all the time I am using 'sg_start --pc=5 /dev/sdb' to put the disk in to sleep mode to spin it down. I have a script which watches /proc/diskstats and if the disk has not been accessed in 15 min I spin it down.
This all works perfectly. The disk spins down when idle and spins back up when the partition is accessed.

But every few days I get a 'aborting journal on device /dev/sdb' and then some scsi errors. The disk goes off line and is unreachable until I unmount/mount it. But it has changed device from /dev/sdb to /dev/sdc.
It does this pretty consistently.

Now I converted from ext3 to ext2 with the same setup and it has worked perfectly for about a week now.

Has anyone seen this problem before?
Is it the expected behaviour of ext3 to have problems with a disk that has spun down?
I didn't expect that the journal would have to be read/write on a disk that was not being accessed.

Thanks.

[Lakshmipathi]

Has anyone seen this problem before?

Not really,I'm using ext3 for 5 years now,I never encountered this problem.

Is it the expected behaviour of ext3 to have problems with a disk that has spun down?

Again not sure,May be drop a mail and check with ext3 users,
Ext3-users Info Page