 | | | Just in case...help on a SCSI issue please | Just in case...help on a SCSI issue please 2005-01-07 - By Brent M. Clements
Back
I have pretty much ruled that the problem lies with the Brocade 3800 Fibre
Channel switch, but I want to see if anyone else has had problems with
Enterprise 3, LSI Fibre Channel cards and Brocade 3800 Silkworms.
Here is a description of our problem:
We have a brocade 3800 switch
We have the following simple configuration
Host system running RHEL 3(latest updates including kernel)
and LSI Logic HBA 929X card
|
|
3800 silkworm running 3.0.2c
|
|
infortrend SATA storage subsystem
When we connect the infortrend storage subsytem directly to the host, we
can format partitions, read and write, etc without any problems.
When we connect the host and the sub system back to the switch and try to
format or perform any i/o on the subsystem, we receive SCSI Timeout
issues, though the operation usually continues and everything is fine.
The same goes for when we try to mkfs.XXX any partition on the storage
subsystem, though again, the format completes fine and we can then mout
the partition as normal. It's just that we receive a large number of SCSI
timeout issues when connected to the switch. The switch ports which are
connected to the host and subsystem show alot of loss of sync errors when
the brocade diagnostic command porterrstat is run. We have swapped
cables, connected the cables to different switch ports, etc yet nothing
seems to work.
We've pretty much eliminated the storage subsystem since it works fine
directly connected to the HBA on the host.
Anybody have any clue what the possible problem could be?
The odd thing is this. We have another brocade 3800 running 3.0 and while
it shows a large number of loss of sync errors, we never experience the
SCSI timeout issue.
Thanks,
Brent
--
Taroon-list mailing list
Taroon-list@(protected)
http://www.redhat.com/mailman/listinfo/taroon-list
Earn $52 per hosting referral at Lunarpages.
|
|
|