IBM E027SLL-H Troubleshooting Guide - Page 178

When the monitoring server is stopped, the internal web server swaps to another IBM Tivoli Monitoring process such as an OS Agent. When the monitoring server starts again, the monitoring server and SOAP are running, but the internal web server is not in the same process. You can determine which process is running the internal web server by starting a browser session to the service console http://server:1920. Ensure the browser View/Status is checked, and then move the cursor over each link. The port involved will be seen in the status line like this: Service Point: cnp -> IBM Tivoli Monitoring Service Console -> IBM Tivoli Enterprise Portal Web Client Service Point: nmp180_hd -> IBM Tivoli Monitoring Service Console In this case the cursor was on the service console link under "Service Point: nmp180_hd" and that was the process running the internal web server. If a firewall rule is in place between the tacmd login process and the server running the hub monitoring server and SOAP, the tacmd login command might fail. The tacmd logic attempts to use the base port connected with the "IBM Tivoli Monitoring Web Services" link. If that is unavailable, it uses the 1920 port. That will fail if the 1920 process is not the same as the monitoring server process. If this condition occurs, it can be resolved by stopping all IBM Tivoli Monitoring processes on that server, starting the hub monitoring server, and then starting up the rest of the IBM Tivoli Monitoring processes. When things are running again, the tacmd login begins to operate as expected. tacmd and SOAP are not able to connect If two instances of the hub Tivoli Enterprise Monitoring Server are started under different user IDs (root plus one other), tacmd and SOAP are not able to connect. When the problem occurs, it is usually in the presence of another process such as the IBM Tivoli Monitoring Universal Agent or the Monitoring Agent for Unix OS. When the problem occurs, both instances of the hub monitoring server are listed on the "IBM Tivoli Monitoring Service Index" web page that is produced on port 1920. The tacmd and SOAP interfaces use the service index as part of their communications, and the extra hub monitoring server entry disrupts them from communicating at all. When the problem occurs, make sure the hub monitoring server started under the non-root login is not running. Then recycle whichever process shows up first on the Service Index page. The system crashes when attempting a bulk import or export command You might be attempting to import from a 0-byte XML file. Ensure that the XML file has content. Also, this can also occur if the file contents are corrupted (if some expected XML elements are missing). Monitoring server fails to start, but then does after a reboot When the Tivoli Enterprise Monitoring Server does not start up properly and you see the following messages in the monitoring server logs, you need to check if anything is using the location server/broker (default is port 1918): (4703AF9A.002B-4:kdcsuse.c,99,"KDCS_UseFamily") status=1c010005, "cant bind socket", ncs/KDC1_STC_CANT_BIND_SOCK (4703AF9A.002C4:kdebpap.c,125,"KDEBP_AssignPort") ip.pipe bound to port 14206: base=1918, 160 IBM Tivoli Monitoring: Troubleshooting Guide