Strange HD problem?
 

I recently installed a 160 WD hard drive in my Mandrake 9.1 box. Had problems mounting it one big partition so I split it into two 80's.

Any ways: Sometimes during continual writing to the drive (copying from a cd to the drive or moving stuff over Samba to the drive) the drive switches to "READ ONLY". When I try to 'touch' a file I get a "drive is read only" message (although fstab says its rw). When I try to unmount it, it says "drive is busy". To unmount it, I have to do a 'fuser' and kill the PID of the drive mount (or whatever is using it), unmount it, then re-mount it. Then it becomes writeable again.

Anyone have this before? What would cause this?

TIA,

Brad

Eh this sounds like a 160Gig HD for a 100$ CAN you bought at boxing day correct? :P

If it is, well I havent had the same problem as you did... My big-file copying works fine in fedora (not in mandrake 9.1)

Just my experience

Yell0w_c0w

I have some more information about this but no solution.

The last time it occured, the /var/log/syslog said there was a Journaling error and the OS was 'remounting' as read-only. I will post some of the log messages for a better example when I get home.

Anyone have this?

Found this in another post: Same messages as mine...

Nov 2 18:31:47 storchnet kernel: journal_bmap: journal block not found at offset 3084 on ide1(22,2)
Nov 2 18:31:47 storchnet kernel: Aborting journal on device ide1(22,2).
Nov 2 18:31:50 storchnet kernel: ext3_abort called.
Nov 2 18:31:50 storchnet kernel: EXT3-fs abort (device ide1(22,2)): ext3_journal_start: Detected aborted journal
Nov 2 18:31:50 storchnet kernel: Remounting filesystem read-only
Nov 2 18:32:35 storchnet kernel: EXT3-fs error (device ide1(22,2)) in start_transaction: Journal has aborted


What if I remount as ext2? Problem solved?

FYI: Answer my own post

The drive is DEAD. Last night, about half of the directories on one of my file systems disappeared. I tried to unmount/remount the drive. It wouldn't mount so I rebooted. When I did, I got a MILLION messages that said "corrupt inode". fsck wouldn't run because of a missing inode table (something like that).

The drive is less than a week old so I'm going to RMA it back to newegg.