After years of looking at disk failures, they have educated me that anything can fail and usually does. The simple disk failures are head crashes or other bad spots on the disks, but the worst are controller failures. Sometimes, the electronics in the disk will lockup the entire SCSI bus, taking everything with it. Other times, the controller reads without error but never writes anything even though the status indicates success.
So the lack of syslog error messages indicates that the driver is happy with the responses but in no way validates the health of the disk. Inchecking the health, I would start with the basics: ioscan -f and diskinfo will test SCSI ID comands, while dd will read the raw disk regardless of content. If those don't work or return strange values, no need to try anything else.
One thing: Is the c6t0d0 disk internal or a simple external JBOD disk, or is it part of a SAN? Improperly configured access or LUNs on the SAN can create strange errors like this.
Bill Hassell, sysadmin
