Zoning for high‐availability systems

Zoning a high‐availability system is slightly more complex than zoning a standard system, due to the need for redundant paths between initiators and targets. Once again, each SAN zone can have only one initiator and one target. But the total number of zones you need depends on whether the SAN is soft‐zoned (by World Wide Port Name) or hard‐zoned (by port number). See:

“WWPN zoning (soft zoning)” on page 16

“Port zoning (hard zoning)” on page 17.

WWPN zoning (soft zoning)

A soft‐zoned SAN maps initiator to target using a logical World Wide Port Name (WWPN), rather than a physical hardware address. This name‐to‐name zoning establishes a logical route that may traverse varying physical ports and varying physical paths through the SAN. To accomplish failover, we thus need only a single zone for the client initiator, the active VTL target, and the standby VTL target.

See the figure below shows a soft‐zoned SAN before VTL failover:

Active VTL server node A

Standby VTL server node B Standby VTL server node A

Active VTL server node B

SAN

Client A

WWPN2 Zone A WWPN1

WWPN3

WWPN5

Client B

I

WWPN6 Zone B WWPN4

During failover, the zone still contains only one initiator and one target at a time. But the target WWPN is remapped from a port on the failed server node to a physical port on the standby server. The standby physical port spoofs the WWPN of the failed port, so zoning does not change. The figure below shows a soft‐zoned SAN after VTL failover, with a standby port spoofing the WWPN of the failed port:

16 VTL User Guide • May 2008

G • 96267

Page 30
Image 30
Sun Microsystems Virtual Tape Library manual Zoning for high‐availability systems, Wwpn zoning soft zoning