On Mon, 14 Mar 2005 23:20:42 +0100 (CET), Maciej ?enczykowski
wrote:
> Have you tried switching to a different HDD cable and/or
> drive? [You have of course fsck'ed the filesystem...]
Tried different drive, but not a different cable. I don't have
physical access to that server, it's only a remote server. How could
the cable give these type of errors? I thought that since it does
the physical connection, that should be ok.
>
> Although if you're getting the same error every time...
> Weird... Cheers, MaZe.
No, not weird, just a problem with the journaling in ext3 and that's
what freezes the server. Problem reported in Kernel 2.6.10.x [forgot
which] according to Google but Centos4 is 2.6.9
Other weird thing that happened in the log. When restarting the next
morning the server, it added some entries from the previous day of
the crash. So the ext3 journaling did it's thing: recovered and
synced the drive.
>
>> Already done :
>> memtest86+ [not memtest86] for 11.5 hours
>> Mersenne prime test for almost 3 hours [I think that's what
>> you mean by mprime]
>>
>
>>>> The error is:
>>>> Kernel: Assertion failure in __journal_file_buffer at
>>>> fs/jbd/transaction.c 1947: "jbd_is_locked_bb_state(bh)"
>>>>
