OK, I realized that I did not give the full details of the kernel I was
running. I was running 2.2.14 + raid-patch. I got the crashes to stop
when I went back to the 2.2.14smp that came with RedHat 6.2.
I also backed all that data that was on the raid array and stopped the
array deleted the /dev/sda5 and /dev/sdb5 partitions that made the array
and created new partitions (type=fd). I then did a mkraid on the new /dev/md0
device THEN did a mke2fs -b 4096(?) /dev/md0. I moved the data back on the
raid and everything seems to be fine as of 4am this morning.
I think my problem was that in the old setup I had done
mkfs on the partitions that made up /dev/md0 (i.e. /dev/sda5 /dev/sdb5)
and then mkraid on /dev/md0 and then much later doing a mkraid on /dev/md0
again (I don't know why).
Any way, the raid seems to be ok now.
Later
Vincent Stoessel wrote:
>
> Ever since I've upgraded to RedHat 6.2
> My raided scsi drive has had questionable integrity.
> The drives were setup to store my extremlely busy
> (read and writes) mysql database tables.
> My main system is mounted on an IDE drive.
>
> /dev/hda8 256194 158197 84769 65% /
> /dev/hda1 21929 11399 9398 55% /boot
> /dev/hda6 2656420 55980 2465496 2% /home
> /dev/hda5 2656420 1569692 951784 62% /usr
> /dev/hda7 256194 11023 231943 5% /var
> /dev/sda6 1935132 245372 1591364 13% /usr/local/backup
> /dev/sdb6 1935132 304636 1532100 17% /usr/local/backup2
> /dev/md0 1968464 193220 1675248 10% /usr/local/var
>
> #
> # persistent RAID1 array with 0 spare disk.
> #
> raiddev /dev/md0
> raid-level 1
> nr-raid-disks 2
> nr-spare-disks 0
> persistent-superblock 1
> chunk-size 4
>
> device /dev/sda5
> raid-disk 0
> device /dev/sdb5
> raid-disk 1
>
> Since the upgrade to the redhat 6.2 I constantly
> getting kernel panics by various programs including
> kswapper and raid1d. To quickly fix the problem I booted
> my machine off the Redhat 6.2 rescue disk and
> copied my data from the raided /usr/local/var to the main ide disk
> (/usr/local/var2).
> The difficult part is that there are so many peices that I have to
> look at. I suspect the adaptec 2940uw card or the aicXX driver.
>
> Machine has been running fine since I made the above changes but I really
> want to find the problem(s) so that I can go back to my raid1 setup which
> I was previously happy with.
>
> Sorry for the sparse details, but I will keep everyone up to date as I
> work on this problem.
> Thanks
>
> Below is the last kernel panic caused by raid1d:
>
--
Vincent Stoessel -
