This is a query related to the behavior of MSCS and EMC Powerpath.
We have the following
1. Dual Ported Qlogic HBA on a Win 2003 Server running SP 1.
One port of the Qlogic HBA is connected to Fabric A and
the other port to Fabric B.
2. MSCS and EMC Powerpath installed on the Win 2003 Server.
3. Storage is a Dual Ported EMC DMX with one port connected
to Fabric A and the other to Fabric B.
2 LUNs from the DMX are being accessed by the Win 2003
server via 2 paths from Fabric A and B.
When we put the 2 DMX LUNs under MSCS control, we see that
for each LUN when FC traffic is being generated [by the Win
2003 server], the traffic flows in both Fabric A and B. This
means that the traffic is being load balanced across the
two Fabrics.
However we also see that on each path from the Win 2003
server a SCSI-3 Persistent Reservation CMD is issued to the
DMX LUN in order to reserve the path to the LUN. It does
this before sending any READ/WRITE/WRITE BUFFER I/o's.
Is this an expected behaviour?
We have the following
1. Dual Ported Qlogic HBA on a Win 2003 Server running SP 1.
One port of the Qlogic HBA is connected to Fabric A and
the other port to Fabric B.
2. MSCS and EMC Powerpath installed on the Win 2003 Server.
3. Storage is a Dual Ported EMC DMX with one port connected
to Fabric A and the other to Fabric B.
2 LUNs from the DMX are being accessed by the Win 2003
server via 2 paths from Fabric A and B.
When we put the 2 DMX LUNs under MSCS control, we see that
for each LUN when FC traffic is being generated [by the Win
2003 server], the traffic flows in both Fabric A and B. This
means that the traffic is being load balanced across the
two Fabrics.
However we also see that on each path from the Win 2003
server a SCSI-3 Persistent Reservation CMD is issued to the
DMX LUN in order to reserve the path to the LUN. It does
this before sending any READ/WRITE/WRITE BUFFER I/o's.
Is this an expected behaviour?