Shutting Down And Restarting One Subsystem Only (Controlled Failover, Mainframe Connect) - Sun Microsystems StorEdge A7000 Service Manual

7.6 Shutting Down and Restarting One
Subsystem Only (Controlled Failover,
Mainframe Connect)
Note – This section applies to mainframe-connected Sun StorEdge A7000 servers
only. For Sun StorEdge A7000 servers connected to an open system host computer:
unmount the open system host's file systems, if the file systems reside on the failed
component's subsystem
Occasionally, you may need to shut down one subsystem while the Sun StorEdge
A7000 is still in use. This technique is called controlled failover, because it involves
manually putting (or varying) channels serviced by the subsystem offline, from the
mainframe side. During controlled failover, one Sun StorEdge A7000 subsystem is
handling all channel communications.
When one subsystem is handling all channel traffic (that is, failed over), the cache
operates in writethrough mode to protect user data.
In the following example, subsystem dsp1 is failed-over to subsystem dsp2.
The controlled failover and restart includes these steps:
1. Log onto the subsystem dsp1 as root, through the System Console subsystem
window. If the subsystem windows are not open, open them according to the
steps in paragraph 7.4.6.
2. Check the channel device paths and device range addresses the subsystem is
responsible for, and vary them offline at the mainframe computer: To view the
pairs of device paths if the Sun StorEdge A7000 uses ESCON connections:
# /usr/install/simckd/bin/ckdadmin switchinfo
To view the pairs of device paths if the Sun StorEdge A7000 uses Block Mux
Channel (BMC) connections:
# /usr/install/simckd/bin/bmctbar
Chapter 7 Operation 7-33