The
Simple Network Management Protocol (SNMP) is typically
used to monitor network health, and performance and hardware, as well as to
find and solve network problems. SNMP consists of two main components:
- SNMP agents, which are software components that reside on managed devices
and collect management information (using Management Information Bases or
MIBs). SNMP agents issue traps when SNMP events occur. These traps are sent
through User Datagram Protocol (UDP) to an SNMP Manager.
- An SNMP manager, which is a network management application (for
example, IBM® Tivoli® NetView®) that
monitors and controls devices on which SNMP agents are running
and can receive SNMP traps.
In SAN File System, each metadata server generates SNMP traps in response
to certain events. SNMP traps are not issued from the operating system, hardware,
or the administrative agent.
Tip: The RSA II cards
can be set up to generate hardware traps as well.
You can configure which severity levels of events (informational,
warning, error, or severe) should generate SNMP traps and you can define which
SNMP managers in the SAN environment are to receive the traps. When an event
occurs with a severity level that causes an SNMP trap, SAN File System sends
the trap, and logs the event in the cluster log.
Note: SAN File System supports asynchronous monitoring through traps but does not
support SNMP GETs or PUTs for active management. The SNMP Manager cannot manage
SAN File System.
Not all events in SAN File System generate traps. Examples of events that
might generate SNMP trap messages include:
- When a metadata
server executes a change in state
- When a metadata server detects that another metadata server is not active
- When the size of a file set reaches a specified percentage of its capacity