We have 3 xserves running osx server 10.3.9 connecting through a Qlogic
sanbox fiber switch to Five (5) xraid controllers. We us lun masking
control which server sees which xraid. We have been having a problem
with two of the xserves now. AFP craches due to the a Volume off one
of the xraids is unmounting. We get the below in syslog at the time it
occurs:
Nov 5 00:17:56 dragon kernel: FusionMPT: Notification = 5 (External
Bus Reset) for SCSI Domain = 0
Nov 5 00:17:56 dragon kernel: FusionMPT: External Bus Reset for SCSI
Domain = 0
Nov 5 00:17:56 dragon kernel: FusionMPT: Notification = 7 (Link Status
Change) for SCSI Domain = 0
Nov 5 00:17:56 dragon kernel: FusionFC: Link is down for SCSI Domain =
0.
Nov 5 00:18:04 dragon kernel: FusionMPT:
FindTaskForControllerIdentifier() failed, targetID = 3, context = 0x
There is not link down notices on the xraid. So I was thinking that it
is either the switch, cable, card or xserve.
On one of the servers a couple weeks back when we were having the
problem I swapped cables with no effect. I then swapped fiber card and
it seemed to help. But now one of the other xserves is doing it. Could
I have two fiber cards go bad in a month?
