On a system that had a range of probes on it, some probes fail to restart after a robot restart, and there are the following entries in the probe logs:
nimSession - failed to create server session on port 0, error code 98
Running a 'netstat -an | grep 480' during the startup may indicate that we are trying to get a port to listen on:
tcp 0 0 #.#.#.#:56846 #.#.#.#:48009 TIME_WAIT
We will ultimately be unable to get the port for listening on and the probe will go into a failed state.
To verify that the problem is not specific to the probe, one of the probes that is running can be shut down (releasing a port) and one of the problem probes should start up fine now with a free port available.
Any UIM/probe Version
Try setting the starting probe port in the controller configuration to the last port that a probe was started on.
IM > double click controller > Advanced tab > First probe port number > Apply
or
AC > robot > probes > click on the ellipsis > configure > setup > First probe port number > Save
logmon example:
Sep 23 14:20:05:977 [140365513185056] logmon: * logmon 4.11
...
Sep 23 14:20:05:978 [140365513185056] logmon: sockCheckLocalPort - close 9
Sep 23 14:20:05:979 [140365513185056] logmon: sockCheckLocalPort - close 9
Sep 23 14:20:05:979 [140365513185056] logmon: sockCheckLocalPort: connect to port 48011 failed 111 (ok - its free)
Sep 23 14:20:05:979 [140365513185056] logmon: sockServer: next available port is 48011
Sep 23 14:20:05:979 [140365513185056] logmon: sockServer: bind failed 98
Sep 23 14:20:05:979 [140365513185056] logmon: nimSession - failed to create server session on port 0, error code 98
Sep 23 14:20:05:979 [140365513185056] logmon: sockClose:0x26652f0:
Sep 23 14:20:05:979 [140365513185056] logmon: nimSession failed
Disabling cdm probe allowed the logmon probe to work without any errors.
Workaround was to set the first_probe_port to 48011 since that was the highest port showing in the list of probes on that robot.