The management and monitoring subagent

The GPFS™ SNMP management and monitoring subagent runs under an SNMP master agent such as Net-SNMP. It handles a portion of the SNMP OID space.

The management and monitoring subagent connects to the GPFS daemon on the collector node to retrieve updated information about the status of the GPFS cluster.

SNMP data can be retrieved using an SNMP application such as Tivoli® NetView®. NetView provides a MIB browser for retrieving user-requested data, as well as an event viewer for displaying asynchronous events.

Information that is collected includes status, configuration, and performance data about GPFS clusters, nodes, disks, file systems, storage pools, and asynchronous events. The following is a sample of the data that is collected for each of the following categories:

Cluster status and configuration (see Cluster status information and Cluster configuration information)
- Name
- Number of nodes
- Primary and secondary servers
Node status and configuration (see Node status information and Node configuration information)
- Name
- Current status
- Type
- Platform
File system status and performance (see File system status information and File system performance information)
- Name
- Status
- Total space
- Free space
- Accumulated statistics
Storage pools (see Storage pool information)
- Name
- File system to which the storage pool belongs
- Total storage pool space
- Free storage pool space
- Number of disks in the storage pool
Disk status, configuration, and performance (see Disk status information, Disk configuration information, and Disk performance information)
- Name
- Status
- Total space
- Free space
- Usage (metadata/data)
- Availability
- Statistics
Asynchronous events (traps) (see Net-SNMP traps)
- File system mounted or unmounted
- Disks added, deleted, or changed
- Node failure or recovery
- File system creation, deletion, or state change
- Storage pool is full or nearly full

Note: If file systems are not mounted on the collector node at the time that an SNMP request is received, the subagent can still obtain a list of file systems, storage pools, and disks, but some information, such as performance statistics, will be missing.