Richard> Okay! I found the problem! Although the file-system had been
Richard> built many times on this device, tha partition table had been
Richard> built with another SCSI Controller! When I executed fdisk,
Richard> it warned that the extended translation was not correct.
Richard> I made new partitions on this drive, made new file-systems,
Richard> then repeated the tests. There was NO file-system corruption.
Richard> Therefore, for the guy that started this thread originally,
Richard> Ingo Molnar, I suggest that he look at the partition table
Richard> with fdisk. He might see the same warning. New paritions fix
Richard> the problem.
It was me who started the thread.
I seem to have another problem than the one you had.
You always got the same number, maybe the number was written
erroneously in your case. I get 4 byte integers which count upwards,
and those only occasionaly.
Another distinction is that I always had the same SCSI Controler, and
I use the whole media for the file system, so I don't even have a
partition table which could cause the trouble. But I also tested it a
while ago with partitions, and there was the same problem.
Does anyone use block devices with a block size of 1024 bytes with
ext2fs bigger than 480M ? The block size is the only thing I see that
distinguishes the mo-drive from a harddisk. And my harddisks run just
fine with ext2fs.
Leo
PS: Is there a newer version of ext2ed than 0.1?