[CentOS] Storage/SCSI Error on our CentOS server

Sat Feb 24 20:17:06 UTC 2007
John R Pierce <pierce at hogranch.com>

Hairul Ikmal Mohamad Fuzi wrote:
> Hi,
>
> Currently we are running CentOS 4.x on a 2-way Opteron machine.
> This machine, through a SCSI host adapter (Adaptec), is connected to a
> 2TB storage unit (an external RAID-5 disk array)
>
> Until our recent unintentional power trip, everything was fine and 
> smooth.
> We have been experiencing complication accessing the storage ( it
> could be either intermittent filesystem error, partition could not be
> mounted in read-write mode, unacceptable writing speed, etc ),
> especially when we start to 'write' on the storage.
>
> After a few check, we are suspecting either :
>
> 1) the storage unit (but the storage control panel did not report any
> disk/raidset failure) is failing or,
> 2) the SCSI host adapter is failing, or
> 3) the filesystem itself is corrupted (we did 'fsck.ext3 -v -f' but it
> turned out it did not find any errors)


or 4) scsi cabling.   I see some scsi transmission errors in there.   
About the only way I know to diagnose something like this would be to 
swap parts... I'd swap the controller card and see if the problems go 
away, then try the cable, then try the storage controller.   if one of 
these things fixes the problem back the other changes out (ie put the 
original card back, etc).