Bruno, your guess was correct. I downloaded the ISO Seatools from the
seagate website and booted of a CD. The drive failed the diagnostics by
reporting alot of 'bad sectors'. The utility suggested to format the
drive using zero fill method... I am not sure how to do this.. but
again I havent googled around either.
What was interesting though was when both drives where connected I was
not able to hot add sda[1..14] to the raid array, but when I physically
removed sbd and placed sda on sata0 it was able to boot. Even though it
booted and recovered some journals it had not synced data for the
previous 4 days. Running the utility again revealed that the drive
still had uncorrectable data errors..thus needing a "Zero fill" format.
I am returning the drive to the vendor as they have the ability to
replace it for me within 24hours... which suits me just fine,, but it
would be nice to know how it is possible to format using the above
mentioned method. Bruno you mentioned dd.. I am not sure what this does.
George
Bruno Wolff III wrote:
On Thu, Nov 10, 2005 at 10:05:28 +1000,
George Pappas <george@xxxxxxxxxxxxxx> wrote:
/proc/mdstat reveals that the Raid Array has lost the sda drive, this
had occured earlier but only dropped part of the sda drive not all. I
This is normal. Partitions that have problems get dropped.
was able to add the drive back by using mdadm... but 2 days later this
has happened again. Has anyone else had this problem? or is able to
guide to further diagnosing my situation.
My guess is that you had some errors on the disk. Are you using smartd to
monitor the drives? If not, you should be. You can use smartctl to see
what the drives have to say for themselves. If there are some bad sectors,
you will generally need to rewrite them so that the drive can remap them.
One way to do this is to pull the partition containing the bad sectors out
of the arrary and use dd to rewrite the bad sector(s) and see if smartctl
says that things have been cleaned up. Once they have add the partition back
into the array.
