After a lot of screwing around, I called Connor to get a replacement for my
1 gig hard drive, which I though was going. They told me that the Fast
SCSI 2 specs require AT LEAST ONE FOOT of ribbon cable between each of my
SCSI devices. New longer cable, and no more kernel panics.
>> * From: trink%/etc/HOSTNAME (Trink Andreas)
>>: My Linux system has started crashing about once or twice a day, never in any
>>: particular pattern. It always starts with the message:
>>: Kernel panic: EXT2-fs panic (device 8/17): ext2_read_inode: unable to read
>>: inode block -inode=????, block=?????
>>: The question marks are always different numbers. Read is sometimes write.
>>: The damage to the file system reported by e2fsck is always different, also.
>>: I'm running kernel 1.2.11, and using a DPT 2120 SCSI controller on two
>>: hard drives and a CDROM.
>>: Anybody have any ideas?
>>I get the same message some times the day, but I have an idea why: My harddisc
>>seams to be a little bit broken, that means, the harddisc switchs off (you
>>hear the noise) and immediatly after it switchs on (you hear the noise also).
>>Thats takes approx. 10 secs. During this time there is no harddisc available
>>and if the kernel wants to read or write a block, it is'nt possible and
>>results into a kernel panic message.
>I had the same problem, except that this didn't result in a kernel panic;
>instead it resulted in the process being put in an uninterruptible sleep.
>Usually, the process that bit it was syslogd...or init. Which made it
>difficult to bring the system down properly....
>The solution (for me): Jiggle the HD's power cable. Mine was just a LITTLE
>bit loose, so it was on occasion breaking contact, causing the drive to power
>down. I re-seated it, and haven't had any problems since.
>Scott Johnson -- Graduate Slave, ECE Department, Oregon State University
>Check out my new (but unfinished) Web page--http://www.ece.orst.edu/~johnsos
>Philosophy is useless, theology is worse.
> -Dire Straits