Hi,
I'm having some serious hard disk problems and I was hoping someone could help.
I Installed three new IBM disks a while back to replace a Quantum that
crashed. With the new drives installed I was using a Quantum Atlas and two IBM
Ultrastar with the third IBM unused (I did try to install NT on it once).
However the Quantum also crashed and I am trying to use one of the new IBMs to
replace it.
The system consists of:
1 Quantum Atlas XP32150W SCSI ID 0
3 IBM Ultrastar 2ES SCSI DCAS-32160 Wide SCSI IDs 1,2,3
1 NCR 815 controller
1 Symbios Logic 875 controller
The Quantum drive is on the 50 pin connector of the Symbios Logic controller
and is terminated. It provides term power to the SCSI bus and has a 68->50 pin
adapter fitted. The three IBM drives are on the 68 pin connector of the
Symbios Logic controller. The last drive is terminated but I can find no term
power jumper for these drives.
The drive I am trying to use is ID 3 but is actually the first one on the 68
pin cable. The last drive on the 68 pin cable is ID 1.
The problem is that whenever I copy a large number of files (100MB+) to one of
the new partitions the partition gets corrupted. This happens if I am
restoring a backup from a tape or copying files from one partition to another.
I have checked all the SCSI connectors and I have ensured that the last drive
on the chain is terminated. I have partitioned the drive with DOS fdisk and
with the Linux fdisk as well as cfdisk and cabaret. I have formatted the
partition with mkfs and mke2fs as well as cabaret and still the problems
continue.
If I mount a clean newly formatted partition and then copy files to it, then
either reboot (after a correct shutdown) or unmount the partition then run
e2fsck -f <partition> I get errors starting with something like:
Pass 1: Checking inodes, blocks, and sizes
Inode 356, i_blocks is 1028, should be 4. Fix<y>?
Inode 357 is in use, but has dtime set. Fix<y>? yes
Inode 283 has illegal block(s). Clear<y>?
Illegal block #12 (2738524101) in inode 283. CLEARED.
Illegal block #13 (3805091013) in inode 283. CLEARED.
Illegal block #14 (644413645) in inode 283. CLEARED.
Illegal block #15 (1876547683) in inode 283. CLEARED.
Illegal block #16 (1975467485) in inode 283. CLEARED.
Illegal block #17 (1552897231) in inode 283. CLEARED.
Illegal block #18 (1966042671) in inode 283. CLEARED.
Illegal block #19 (2875689973) in inode 283. CLEARED.
Illegal block #20 (353359243) in inode 283. CLEARED.
Illegal block #21 (109761999) in inode 283. CLEARED.
Illegal block #22 (719586791) in inode 283. CLEARED.
Too many illegal blocks in inode 283.
Clear inode<y>? yes
Inode 285, i_blocks is 76, should be 74. Fix<y>?
Inode 359, i_size is 208, should be 19456. Fix<y>?
Things go on like this and then once all the Inodes are fixed Duplicate/bad
blocks are found (it seems like thousands of them). There are no bad blocks on
the drive so they all must be duplicates. These are cloned and then a
neumerous files are moved to lost+found.
I'd appreciate some suggestions.
Thanks.
Red Hat Linux 5.0 (all updates applied).
mount-2.7f-1
fileutils-3.16-6
util-linux-2.7-15
--
Eze Ogwuma
--
PLEASE read the Red Hat FAQ, Tips, Errata and the MAILING LIST ARCHIVES!
http://www.redhat.com/RedHat-FAQ /RedHat-Errata /RedHat-Tips /mailing-lists
To unsubscribe: mail [EMAIL PROTECTED] with
"unsubscribe" as the Subject.
