Yesterday I move some agent to our 10.5.2 SP2 cluster. yesterday till 6 PM I was login to that cluster and making sure the agents are reporting. since morning I am not able to login to that cluster. I restart it but nothing work.
workstation is showing connecting to enterprise manager and after taking 15 min throw "Error attempting to connect to enterprise manager "hostname"
logs are not showing any relevant message only getting following
3/22/18 11:14:55.100 AM EDT [ERROR] [Alarm Pooled Worker] [Manager.SNMPAlertAction] Failed to send SNMP trap due to exception: Socket problem: host=10.205.228.31, port=163, bindAddr=10.215.204.298 Cannot assign requested address (Bind failed)
Ok so this means that port 163 was failed to be bound to. This may have happened because another process has already bound to it. If on Windows, I would check to see if Windows SNMP is running as it uses the same port and interferes with us. Or possibly another outside process has it opened.
Another thing that could have happened is that an alert was opened and sent out via SNMP (could be caution or danger). Every SNMP alert that gets opened has to get closed. When that alert closes (goes below caution or danger depending on how the alert is set up), another SNMP is sent closing it.
So its possible that the SNMP open trap went through but never closed. Thus any subsequent open traps will fail with the bind failure message.
You could search the logs for this, but probably best to open a case for investigation.
Case has been open 00993679
Per Support Case 00993679, the problem was resolved after using the exact same release of 10.5.2 SP2 Workstation as the MOM. The previous problematic workstation release was 10.5.2.15 (Build 990015).
In general the standalone workstation version being used should match the target EM version to ensure compatibility because only that combination would have been tested (using webstart workstation automatically means the workstation version matches EM version because it is downloaded from the EM).