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

Processors: Monitoring and Recovery
HP NonStop S-Series Operations Guide—522459-008
11-8
Recovery Operations for a Hardware Error Freeze
7. Reload the remaining processors.
Send the dumps to your service provider.
Recovery Operations for a Hardware Error Freeze
Contact your service provider. Depending on the circumstances, a hardware error
freeze might require the PMF CRU or a memory unit to be replaced. See Replacing
Processor Memory or a PMF CRU on page 11-25.
When a hardware error freeze occurs, scan strings are automatically saved to
$SYSTEM.ZSERVICE. In OSM, the default file name is ZZPSnnnn. In TSM, the
default file name is ZZSSnnnn. If the PMF CRU or memory unit is replaced due to a
hardware error freeze, back up the ZZPSnnnn or ZZSSnnnn file to tape and submit
it along with the CRU or memory unit being replaced.
If the processors in your system use a NonStop S7000 PMF CRU, S72000 PMF CRU,
or S7400 PMF CRU, you should also submit a memory dump to the Global Customer
Support Center (GCSC) for concurrent analysis. When a hardware error freeze occurs
for a NonStop S70000 PMF CRU, however, dumping memory after the scan string file
is constructed does not provide any useful information.
If you want to suppress the scan string save function for a NonStop S70000 PMF CRU
to enable memory dumps, contact your service provider.
Recovery Operations for an Unacceptable Service State
Contact your service provider. When the service state reported by the OSM Service
Connection or the TSM Service Application is not acceptable, the PMF CRU or a
memory unit might need to be replaced. Refer to Replacing Processor Memory or a
PMF CRU on page 11-25.
Recovery Operations for a Processor Halt
When a processor halts unexpectedly or shows signs of abnormal behavior:
1. Review Identifying Processor Problems on page 11-5.
2. Make sure that a processor dump is the best course of action. Although a
processor dump is useful for most processor or system failures, contact your
service provider before attempting a processor dump if a processor is in a
hardware error freeze state.
Note. After reloading the remaining processors, run your startup scripts if any.
Caution. If an error message alerts you that the freeze is due to an uncorrectable memory
error (UCME), do not take a memory dump. The dump will cause another hardware error
freeze.