Hi,
Today one of my machine got the following error and the filesystem was
remounted as read-only automatically:
Aug 3 14:24:09 localhost kernel: EXT3-fs error (device md2): ext3_readdir:
bad entry in directory #48170704: rec_len is smaller than minimal -
offset=0, inode=0, rec_len=0, name_len=0
Aug 3 14:24:09 localhost kernel: Aborting journal on device md2.
Aug 3 14:24:09 localhost kernel: ext3_abort called.
Aug 3 14:24:09 localhost kernel: EXT3-fs error (device md2):
ext3_journal_start_sb: Detected aborted journal
Aug 3 14:24:09 localhost kernel: Remounting filesystem read-only
Aug 3 14:24:09 localhost kernel: __journal_remove_journal_head: freeing
b_committed_data
Aug 3 14:24:09 localhost kernel: __journal_remove_journal_head: freeing
b_committed_data
My machine is running with a AMD x86_64 CPU, 3GB Ram and 10 Seagate harddisk
in MD raid 5 mode. Kernel is 64bits 2.6.17
The MD raid 5 did not report any problem and my server log does not contain
any error about libata and md. It just show the above ext3 errors.
The affected filesystem is around 1 month old and never have any error and
crash before.
I need to umount the affected ext3 volume and remount it to take it back to
read write mode.
Please cc me if possible, I'm not subscribed. Thanks.
Kyle