[Linux-cluster] SCSI Error

[Bob Peterson]

On Mon, 2007-07-02 at 15:23 +0800, pastany wrote:
> I am running a 4 node cluster with a fc switch and a fujitsu fc san 
> but i recevie this message ,and some partions dont work
>  
> Jun 27 19:23:49 test1 kernel: SCSI error : <3 0 0 0> return code =
> 0x10000
> Jun 27 19:23:49 test1 kernel: end_request: I/O error, dev sdb, sector
> 668992848

Hi Pastany,

This sounds like a hardware problem to me, not a GFS problem.
It could be a bad drive, bad san or bad Host Bus Adapter (HBA).
Perhaps you should unmount the san from all the nodes, then
from one node, do a simple read test of the san:

1. Check dmesg and maybe clear your dmesg buffer:
   dmesg -c
2. Try reading every sector of the san:
   dd if=/dev/sdb of=/dev/null bs=1M
3. Check your console / dmesg to see if SCSI errors are reported.

You may want to try that separately on a few different nodes just
in case the error was caused by a bad HBA in the node that 
reported the problem.

Regards,

Bob Peterson
Red Hat Cluster Suite