We have a Distributed SpectroSERVER Environment (DSS) with the MLS also used as a trap director for all the devices.
Some of the devices (10 out 2,000), which are modeled on SpectroSERVERs other than MLS are facing issue in receiving traps.
A sniffer trace shows the trap is received at the MLS but not put on the device on the remote landscape, where the device is modeled.
We have also checked the alert_remote_fwd_queue_length attribute for the trap director (MLS) is 0 and is directing the traps correctly in most instances.
If we set the remote landscape as the trap destination on the device, and simulate a trap, using a mib walk request with an incorrect V2c string, the trap arrives at the SS as expected.
As soon as we change the trap destination back to the MLS, no more test traps arrive.
These traps are not seen in the Trap Director cache.
Release : 10.2
Component : Spectrum Core / SpectroSERVER
The devices in a DSS that use Trap Director, should only be modeled on 1 landscape.
If they are modeled on the SpectroSERVER that serves as the Trap Director, the event will be created there and Trap Director, for that event, will not use the Trap Director functionality and thus will not be seen in the Trap Director cache.
All duplicates on the SpectroSERVER where Trap Director is enabled, need to be deleted for this functionality to work