From: eric@tantalus.nrl.navy.mil (Eric Youngdale) Subject: Disk corruption problem (Was: Wierd behavior with 0.98). Date: 1 Oct 1992 14:09:39 GMT
In article <1992Oct01.125825.1150@infko.uucp> hodgen@infko.uni-koblenz.de writes:
>Last night I grabbed the 0.98 sources, added the fast scsi patches posted here
>recently and compiled (Jump table 4.1). Rebooted, start up X and go to xterm
[...]
>I ran fsck -v /dev/sdc3 (my /usr) and something had written about 15MB worth of
>Zone XXX marked as used but not used" type messages. Any ideas? My hardware
>
>486 33MHz 256KB cache
>8MB RAM
>Adaptec 1542B
>2 x Quantum 210
>Seagate ST2096N
[...]
Yes, I think I have an idea about what is going on. A week ago, 3
people reported an inode corruption problem with scsi disks, and the one thing
that they all had in common was that they each had more than one scsi disk
drive. From the looks of it, I can now say that we have a fourth case. This
problem has been observed both with an Adaptec 1542 and a ST02.
Currently I think that there is a bug of some kind which has to do with
more than one scsi disk drive on the system at one time. What I do not know is
if a multi-scsi-disk system is always unstable or not. Basically I would like
to hear from anyone who is using more than one scsi disk without problems.
Also, I would be interested in knowing if there are any users with
two IDE drives, and I would like to know if that configuration is stable
(especially if you are using at least part of each drive for linux).
(My recollection is that there are people using one scsi disk with one IDE disk
without trouble).
For those of you experiencing the trouble, try cutting back to a single
scsi disk system, and see if that helps.
I hope to be able to come up with a patch once I have identified the
problem.
-Eric
-- Eric Youngdale eric@tantalus.nrl.navy.mil