(I'd much prefer to work on turbo support, but...)

Can anybody recommend some tools to verify/stress test a hard disk,
preferably with NS3.3?

Background: I recently started setting up a local CVS repository to follow
-current (on a turbo slab with NS3.3).

However, it turned out that the imports failed because of corrupted rcs
files in the repository.
Failure rate is about 1-2 corrupt files per syssrc import (65MB). Files
concerned are reproducible, as is the type of corruption (interspersed NUL 
characters aligned at 8k multiples, possibly followed by data from a
different file). 

Some lengthy CVS debugging produced nothing (even removing/changing all
8k-sized buffers did not change things). 

However, it came out the problem only happens with my trusty IBM
DCAS-34330 (otherwise working normally), other disks worked fine.

Writing the same test cases to this same disk over NFS (from remote or
even via a local NFS export/mount) does work, however. (Because it's so
slow?)
I've seen no error messages, and reformatting the drive did not help.

All in all it now looks like some fs/driver/disk interaction to me (8k
fs buffer size?). 

However, I can't completely rule out disk hardware problems, hence the
question about stress test software.

Any ideas how to debug/solve this? 
The disk in question should hold the -current sources and the netboot
filesystem for the client...

Regards,
Timm
-- 
Timm Wetzel     	<twetzel@gwdg.de>
Biomed. NMR GmbH        Tel +49 551 201-1091    FAX +49 551 201-1307
Am Fassberg 11, D-37077 Goettingen, Germany