After successfully completing a NSM 11.2 Event Agent install on a UNIX/Linux system, the Event Agent fails to start. An error message similar to the following appears when executing "unistart all" or "unicntrl start all":
Unable to reach CCI on node NNNNNNNN.
Executed CA-opr start...................................FAIL 4
where NNNNNNNN corresponds to the server machine where the Event Management policy database resides (Event Manager) which was specified during the agent install. This is the value of the $CAI_OPR_REMOTEDB environment variable.
This error typically occurs because the hostname entered when prompted for your Event Manager does not match the name that this server is using for CCI communications. In some cases, the problem is related to CCI configuration or communication problems. To troubleshoot this problem, check the following:
- If you are not currently at the SP2 release level of NSM 11.2, first upgrade to this release level then try starting the Event Agent again. Continue with step 2 if OPR continues to fail to start.
- On the Event Agent, execute the "ccii" command and locate the entries that correspond to the Event Manager. These entries are similar to the following:
Oid(NNNNNNNN,CAI_OPR_DAEMON ) Did( , ) type(R)
Oid(NNNNNNNN,EMSRVC_ROUTER_U ) Did( , ) type(R)
where NNNNNNNN is the registered Event Manager CCI name.
- Make sure that CAI_OPR_DAEMON and EMSRVC_ROUTER_U are displayed as active CCI components for the Event Manager server. If one or both of these are missing from the ccii outputs, then the Event Manager must be started on the Event Manager server:
unicntrl start all
Once Unicenter NSM is started on the Event Manager server and the required components are displayed in the ccii outputs on the Event Agent, try to start the Event Agent. Continue with step 4 if OPR continues to fail.
- On the Event Agent, check the value of the CAI_OPR_REMOTEDB environment variable:
If the hostname does not match the name reported for the Event Manager in the ccii outputs on the Event Agent then make the following configuration file changes after shutting down all of UnicenterNSM (unishutdown all):
- Modify the $CAIGLBL0000/cci/config/<hostname>/ccirmtd.prf file and change the REMOTE entry for the Event Manager. Replace both occurrences of the Manager's hostname with the name reported by the ccii outputs on the Event Agent:
REMOTE = NNNNNNNN NNNNNNNN 32768 startup
- Modify the $CAIGLBL0000/scripts/envset environment setup script and modify Manager's hostname for the following variables replacing the current value with the hostname reported by the ccii outputs on the Event Agent:
# Set the name of the Event Agent Policy Node
- Restart all of UnicenterNSM. Continue with step 5 if OPR continues to fail to start.
- On the Event Manager server, check for a REMOTE entry for the Event Agent system in the CCI remote configuration file (ccirmtd.rc in the ..\CA\CA_APPSW directory on Windows, $CAIGLBL0000/cci/config/`ca_uname`/ccirmtd.prf on UNIX/Linux). If a REMOTE entry exists, make sure that this entry matches exactly the LOCAL entry in the ccirmtd.prf file on the Event Agent server. If it does not, then make the appropriate change to the CCI remote configuration file on the manager then recycle CCI on the Event Manager server. Once the CCI remote service is restarted, recycle all of Unicenter NSM on the Event Agent.
- To recycle the remote CCI service on a Windows Event Manager:
ccicntrl stop rmt
ccicntrl start rmt
- To recycle the remote CCI service on a UNIX/Linux Event Manager:
- If OPR continues to fail to start, then there are CCI communication problems between the Event Agent and the designated Event Manager.
- First make sure that you can ping both ways by hostname (from Event Agent to Event Manager and from Event Manager to Event Agent). Correct any network configuration issues if you cannot ping in both directions.
- Then check that CCI communication is working both ways using the ccir/ccis utilities. Run ccir on one of the two hosts then on the other host run:
ccis <remote hostname> 3
You should see 3 test messages on the remote host where ccir is running. Repeat this process in reverse by running ccir/ccis in the opposite direction. If CCI is not working check that double check the CCI ccirmtd configuration files on both sides and make sure that they are correct.
- If you are still experiencing CCI communication problems, or OPR continues to fail to start, then it will be necessary for you to open an issue with CA Support Online (support.ca.com).