SNMP Configuration and Management Manual

Host Resources Subagent
SNMP Configuration and Management Manual—424777-006
11-7
Initiating Backup Process Takeover
Initiating Backup Process Takeover
The Host Resources Subagent can be run as a process pair to achieve a basic
fault-tolerant level. The support includes a persistent subagent process and
checkpointing of the startup parameters.
To run the subagent as a process pair and have the backup process take over if the
primary process fails, use the -b startup parameter:
RUN HMSAX /NAME $HMSA, NOWAIT/ -b
You can specify a backup cpu number during starup to create a backup process on the
specified cpu. To start the backup process on a particular cpu, invoke HRSA with -b
[backup_cpu_num] option.
If the specified backup cpu is not up and running, backup process will start on any
available cpu.
If the specified backup_cpu_num is not a valid cpu number, then an EMS event
message will be generated and HRSA process will stop.
You can force the backup process to take over and a new backup process to be
created by setting the value of an object in the zhrmSaProcess group.
Monitoring the Open System Services (OSS) File System
To start the Host Resources Subagent to an HP NonStop Kernel Open System
Services (OSS) environment, use the -u startup parameter:
RUN HMSAX /NAME $HRMIB, NOWAIT/ -u
To monitor both Guardian and OSS environments, enter:
RUN HMSAX /NAME $HRMIB, NOWAIT/ -t -u
If you specifiy neither the -u or -t startup parameters, monitoring of the Guardian
personality is enabled, and monitoring of the OSS personality is disabled.
In the NonStop Kernel Guardian environment, there is a one-to-one mapping between
a file system (fileset) and a physical disk partition.
In the OSS environment, a fileset can span multiple physical disks. To provide
monitorability of OSS filesets, the Host Resources MIB gives information about the
filesets in the hrFSTable and in the hrDiskStorageTable. Each disk in an OSS file
system’s disk pool is treated as an HRMIB partition. Detailed information for each
Guardian disk in the pool is available in the hrPartitionTable, hrDeviceTable, and
hrStorageTable.
The file-system size is calculated as the sum of the component disk sizes. If a disk is in
more than one file system, the disk size will be accounted for twice (once in each file
system).