> I am running the backups every night, but afterwards when I unmount the > raid array and make a fsck.ext3 over it, every time I get errors (bad > blocks here, some inode errors there and so on). Next night I can use the > array quite fine, but on the day after that I get some filesystem errors > again. >
Bad drives? I don't trust drives until they've run for at least 2 wks after I purchase them - I've had them die in that time. > Now I do not really know where to search for the problem. Is it Linux's > software raid5 implementation, which is buggy? Is it one or more of the > IDE disks? Or is this normal for large (~400gb) ext3 partitions? > No. I can't speak to ext3 partitions, but I've run a few XFS partitions and haven't seen more than 1 drive fail on a 4 disk raid, and 1 out of a 99 drive array (fibre channel). You might want to change the filesystem to RiserFS of something else a bit more robust than ext3. Are these drive WD per chance? If so, I'd suggest looking for some Samsung or Seagate drives. Bob - -- [email protected] mailing list
