Sun Microsystems T3 用户手册
Chapter 3
Configuring Partner Groups
25
Any controller unit will boot from the master controller unit’s drives. All
configuration data, including syslog information, is located on the master
controller unit’s drives.
configuration data, including syslog information, is located on the master
controller unit’s drives.
How Partner Groups Work
If the master controller unit fails and the “heartbeat” between it and the alternate
master stops, this failure causes a controller failover, where the alternate master
assumes the role of the master controller unit. The new master (formerly the
alternate master) takes the IP address and the MAC address from the old master and
begins to function as the administrator of the storage system. It will also be able to
access the former master controller unit’s drives. The former master controller unit’s
drives will still be used to store syslog information, system configuration
information, and bootcode. Should it become necessary to reboot the storage system
while the master controller unit is inactive, the alternate master will use the former
master controller unit’s drives to boot.
master stops, this failure causes a controller failover, where the alternate master
assumes the role of the master controller unit. The new master (formerly the
alternate master) takes the IP address and the MAC address from the old master and
begins to function as the administrator of the storage system. It will also be able to
access the former master controller unit’s drives. The former master controller unit’s
drives will still be used to store syslog information, system configuration
information, and bootcode. Should it become necessary to reboot the storage system
while the master controller unit is inactive, the alternate master will use the former
master controller unit’s drives to boot.
Note –
After the failed master controller is back online, it remains the alternate
master controller and, as a result, the original configuration has been modified from
its original state.
its original state.
In a redundant partner group configuration, the units can be set to do a path failover
operation. Normally the volumes or LUNs that are controlled by one unit are not
accessible to the controller of the other. The units can be set so that if a failure in one
controller occurs, the remaining one will accept I/O for the devices that were
running on the failed controller. To enable this controller failover operation,
multipathing software, such as VERITAS Volume Manager, Sun StorEdge Traffic
Manager software, or Solaris Alternate Pathing (AP) software must be installed on
the data application host.
operation. Normally the volumes or LUNs that are controlled by one unit are not
accessible to the controller of the other. The units can be set so that if a failure in one
controller occurs, the remaining one will accept I/O for the devices that were
running on the failed controller. To enable this controller failover operation,
multipathing software, such as VERITAS Volume Manager, Sun StorEdge Traffic
Manager software, or Solaris Alternate Pathing (AP) software must be installed on
the data application host.
Note –
In order for a feature such VERITAS DMP to access a LUN through both
controllers in a redundant partner group, the
mp_support
parameter must be set to
rw
to enable this feature. If you are using Sun StorEdge Traffic Manager, the
mp_support
parameter must be set to mpxio. For information on setting the
mp_support
parameter and options, refer to the Sun StorEdge T3 and T3+ Array
Administrator’s Guide.