To manage multiple templates in HP OMU, we need to dump them first. Dumping to file using opctempl will ease up the process of narrowing variation to gain uniformity. In the end converging to standard will help us tackle problems when faced with hundreds of templates being assigned to hundreds nodes. I tried to make [...]
In a de-installed node, OVCoreID can then be manually added to the management server. After installing the node copy the new $ ovcoreid from the managed node to the management server: opcnode -chg_id cocpeca1.somecompany.lan id=xxx <ovcoreid> The node label cocpeca1 must already exist, check first by using: opcnode -list_nodes node_list=cocpeca1.somecompany.lan It is also better to [...]
In my-client implementation there is standard owner for OV processes, that is opc_op. We can check this on running agents/OV daemons via ps -ef | grep OV. A non-standard my-client will return e.g. (which are owned by root): root 2555 1 0 Mar10 ? 00:00:08 /opt/OV/bin/ovcd root 2569 2555 0 Mar10 ? 00:00:08 /opt/OV/bin/ovbbccb -nodaemon [...]
If you wish to clean all OVO installation in managed node, start deinstallation from other OpenView application i.e. OV Performance Agent (OVPA). This will resolve dependency issues.
…cont’d… More Messy situation usually occur if agent was previously installed (along with activation in management server, node addition, etc.). If an HTTPS agent is deinstalled manually in the managed node, its data stays exist in certificate server & management server. We need to clean this mess. Under complete deinstallation, the following directories are removed [...]
Having pre-installed machine to be used for agent installation experiment is somehow tedious. But, I benefit from the troubleshooting effort. Actually the pointers is written in Arsena – OVO Admin Guide v1.0a, but it wasn’t emphasized visually as being significant (so I missed it). There are three common install situation: Clean. OVCoreID and node mapping [...]