Hi!
I noticed yet another dying IBM drive in one of my servers, producing IDE
+ "standard" ext3 error messages (happened during the start of a new
journal methinks):
ext3_reserve_inode_write: aborting transaction: Journal has aborted in
__ext3_journal_get_write_access<2>EXT3-fs error (device ide5(57,1)) in
ext3_reserve_inode_write: Journal has aborted
which clearly shows that there's a "\n" missing in the printk in
~/include/linux/ext3_jbd.h, line 97. (or the string printed misses one,
don't know).
ext3 remounted the drive readonly; I did a badblock test on it (which
returned zero errors, *sigh*, I love flaky hardware), e2fsck'ed it
(complained only about a missing lost+found an some wrong free block
counts, nothing major), and then tried to remount it rw again but got the
following error:
root@yadda:~# mount -oremount,rw /dev/hdk1
mount: block device /dev/hdk1 is write-protected, mounting read-only
root@yadda:~#
which was accompanied by the following printk's:
ext3_abort called.
EXT3-fs abort (device ide5(57,1)): ext3_remount: Abort forced by user
Do I have to reboot the server to be able to mount the FS rw again or
should a normal umount/mount suffice?
best regards,
michael
