NonStop S-Series Operations Guide (G06.27+)

Processors: Monitoring and Recovery
HP NonStop S-Series Operations Guide—522459-008
11-9
Recovery Operations for a Processor Halt
3. Dump (copy) the contents of its memory to disk or tape unless otherwise indicated.
Dumping the contents of a halted processor (its registers and entire memory
contents) can be a useful diagnostic tool for analyzing and resolving the problem.
If you do not have failure-recovery software, such as the HP Tandem Failure Data
System (TFDS), installed on your system, perform a processor dump:
•
To perform a memory dump of a processor on a running system, follow the
directions in Dumping a Processor to Disk on page 11-10.
•
If the entire system is down (all processors are halted), you can perform a tape
dump on one processor as described in Dumping a Processor to Tape (Down
System Only) on page 11-17. Tape dumps must not be performed on a running
system.
•
If you are advised to perform a memory dump for all processors, refer to
Dumping All Processors in a System on page 11-21.
4. Submit a tape containing the processor dump to your service provider for analysis.
5. Load the operating system into the processor or processors that halted.
•
To reload individual processors, use the TACL RELOAD nn command, where
nn is the processor number of the processor you want to load.
a. From the Startup TACL window, log on to the system as a super-group
user (255,n).
b. At the TACL prompt:
> RELOAD nn, prime
where nn is the number of the processor to reload.
c. Wait until the Processor Status dialog box shows the status for all
processors to be “Executing NonStop OS.”
•
If all processors are halted, you must start the system as described in
Section 16, Starting and Stopping the System
.
6. Restart processes that were running in the affected processors without a backup
process. Generic processes and processes that were running with a backup
process are restarted automatically by the system.