Template:FAQ:Agent 20
This is most commonly seen when performing an "snmpwalk" on an agent
that's either using a default "vendor provided" configuration
(typically access to the 'system' group only), or which is trying
to restrict access for individual users or communities to a subset
of the whole OID tree.
The agent implementation of "GetNext" processing is relatively inefficient when dealing with inaccessible objects, and it is quite easy for the clients to time-out and retry a request, while the agent is still trying to process the original. If this happens continually (as is typically the case with an snmpwalk), the agent can get swamped by this backlog.
The 5.0.x line has now addressed this, starting with the 5.0.7 release. The 4.2.x line still suffers from this problem, and it is unlikely that this will be fixed. (The 5.0.7 approach relies on some of the new features in the 5.0.x line, and it has not proved possible to apply this to the 4.2.x code base).
