Hi all, We have several deployments of our application in different data centers across the world. Most of these are in Vmware ESX deployments. All of these virtual machines are only aware of their private IP (Class C subnet). The hosting provider then maps a public IP on to the private IP with their infrastructure via a 1:1 IP NAT. As a result, when I set up the client in our ESX servers, it can only see the private IP. However when it connects to our hyperic server at the office, it reports the incorrect IP to the hyperic server as it sends the class C IP, not the publicly accessible IP. Is it possible to override the IP of the agent so that it binds to the private IP, but sends the public IP to our server?
It says you can specify, for instance, the address of a firewall doing port forwarding. I'm guessing you could put your public IP in the agent.properties and it would send that instead. You'd have to customize your agent.properties for each agent, but it might be cleaner than the alternative...
... since the only alternative appears to be hacking up the EAM_AGENT.ADDRESS in HQ's db to set the agent IP's manually 🙂 You can use the HQAPI (http://support.hyperic.com/display/DOC/HQApi+agent+command) to get a list of your agents, but they don't appear to support syncing agent information.
Hi Mirko, No I did not run setup again as this is a windows client and I used the msi installer to install the client. I modified the properties manually as per the earlier post to have a bind IP for the agent to bind the socket to, and then just started it. I ran setup and that fixed it, thanks for all your help!