Just adding this as a quick note, in case anyone comes across the same messages in the future.
A client reported that they were swapping the roles of an ITSM 14.1 Conventional Configuration setup.
That is, the Primary Server became the Secondary Server, and the Secondary Server became the Primary Server.
(This is an unusual process. It is not needed in the normal running of an SDM site.)
This had been done successfully previously and run for some time. However, they now wished to put the servers back to the same original roles, and encountered errors.
It might be important to note that they were using DNS names for the servers, and not a direct hostname or IP address.
The failure conditions were these:
02/06 22:16:48.13 MyServer01.abc key_prov_nxd 4240 SIGNIFICANT address.c 474 Unable to find hostname "MyServer01.abc.def.ghi" - setting to local slump id
02/06 22:25:10.93 MyServer01.abc dbmonitor_nxd 4156 ERROR bpobject.c 1998 Found null target node in BPMessage to bpvirtdb_srvr
02/06 22:00:51.685 INFO ? ? WARNING - Check(My_Primary01) does not match current server.
02/06 22:00:51.986 INFO ? ? Process sql_check_db completed with exit code 10
The "netstat" result was unusual, with 18,500 (and counting) "TIME_WAIT" statements against the slump port 2100:
TCP 10.0.254.185:2100 My_Secondary01:37932 TIME_WAIT
TCP 10.0.254.185:2100 My_Secondary01:37933 TIME_WAIT
TCP 10.0.254.185:2100 My_Secondary01:37935 TIME_WAIT
Re-running pdm_configure, even after a server restart, was not sufficient to overcome these issues.
Nor was the issue localised to the usp_servers or dlgtsrv tables.
The client identified that the pdm_configure DNS names had not populated to the NX.env and .tpl files.
After this change was made, all error messages disappeared.