
5–18
HSG80 Array Controller V8.7 Maintenance and Service Guide
Upgrading the Subsystem
6. Set the CACHE_FLUSH_TIMER to 1 second using the following commands:
HSGA> SET THIS_CONTROLLER CACHE_FLUSH_TIMER=1
HSGA> SET OTHER_CONTROLLER CACHE_FLUSH_TIMER=1
7. Disable writeback caching on all units to help minimize the failover time, using
the following command as required for each unit:
HSGA> SET unit-name NOWRITEBACK_CACHE
8. Determine if all data has been flushed from the cache module, using the following
command:
HSGA> SHOW THIS_CONTROLLER
The following text is only a portion of the resulting display—an arrow points to the
pertinent status indication:
IMPORTANT: Repeat step 10 on both controllers (THIS_CONTROLLER and
OTHER_CONTROLLER) until no unwritten data remains in either cache module memory.
If unwritten data is present after several minutes, verify that WRITEBACK_CACHE was
disabled on all units using the SHOW UNITS FULL command. For any units with
WRITEBACK_CACHE enabled, return to step 9 and proceed.
9. Shut down Controller B using the following command:
HSGA> SHUTDOWN OTHER_CONTROLLER
IMPORTANT: Disregard any messages pertaining to misconfigured controllers or failover
status.
NOTE: After Controller B shuts down, the reset button and the first three LEDs turn O
N
(see Figure 5–3).
Proceed only after the reset button stops F
LASHING
and remains O
N
.
10. Verify that all units failed over to Controller A by using the following command to
show the status of each unit:
HSGA> SHOW UNITS FULL
The following text is only a portion of the resulting display—an arrow points to
the pertinent status indication:
Cache:
256 megabyte write cache, version 0022
Cache is GOOD
No unflushed data in cache
CACHE_FLUSH_TIMER=1 SECOND