Bug in 2.0.29? (was Re: ide timeout)

dave (tel1dvw@is.ups.com)
Thu, 27 Feb 1997 00:03:02 -0500 (EST)

Ok, this is the story. Please bear with me, as I think this is important.
I was getting wierd scsi and ide timeout errors, so I replaced the
motherboard, hoping that would fix the problem. I could reproduce the
problems by writing a bash while loop that copies data from one disk to
another. After I replaced the motherboard, I ran that same script. I
didn't get the timeout errors any longer, but I did get errors to the
extent of:

kernel: EXT2-fs error (device 08:11): ext2_read_inode: bad inode number: 0

which is a different error than I had gotten before I replaced the board.
This is the script that is doing it, and I didn't realize its problems
until I thought about it for a while:

while [ 1 ]
cp -rvf /usr/* /mnt/scsi1/ &
cp -rvf /usr/* /mnt/scsi2/ &
cp -rvf /usr/* /mnt/ide/ &
sleep 60

This was to stress-test the disks. I realized that if the copies didn't
finish in 60 seconds, there is potential for the cp's to start to
overwrite each other.

Does the error message I got indicate a normal condition for such
circumstances? I would think the kernel could handle that.

Please advise, as I am interested in knowing why this happened. Is there
a more idiot-proof way to stress-test the motherboard, to determine if it
is in fact the motherboard?

Dave Wreski