OpsCenter with Cassandra spanning multiple EC2 regions.

(5 posts) (2 voices)

• Started 1 year ago by
• Latest reply from
• This topic is resolved

Tags:

No tags yet.

1. 

   Anonymous
   Unregistered

   I'm running onto an issue with OpsCenter. We have a Cassandra environment spanning 2 EC2 regions. In order for them to all be reachable from the OpeCenter host, I added them using the public address. Unfortunately the agent is reporting the host info based on the local private address. The public address is not an interface on the local host but a NAT translation.

   All updates from the agents are tagged with the Private IP which isn't mapped a host in OpsCenter.

   Is there a way to map these similar to the cassandra idea of listen, RPC, and broadcast address?

   If not would there be an issue setting up a second for other issues?

   Posted 1 year ago #
2. 

   mbulman
   Key Master

   There are several configuration options you can utilize for the agent in address.yaml.

   local_interface - This is the IP the agent will report back to opscenter. It should match whatever you see in nodetool ring (we attempt to auto detect this if not explicitly set)

   agent_rpc_interface - This is the interface the agent will listen on for its http api. This would need to be the IP that is accessible from the central opscenter machine.

   thrift_rpc_interface - This is the IP the agent will attempt to connect to via thrift for writing metric data.

   Posted 1 year ago #
3. 

   Anonymous
   Unregistered

   Not 100% clear here.

   I think I got local_interface but seem to be not quite getting the other 2.

   Assuming the node is behind a layer of NAT between the OpsCenter host and it:
   Should the agent_rpc_interface be the local address on the host or the translated address?
   Should the thrift_rpc_interface be the (public) address of the OpsCenter host?

   On a related note is there a list of available options for the address.yaml file?

   Posted 1 year ago #
4. 

   mbulman
   Key Master

   agent_rpc_interface should be the translated address. It's the IP that the central OpsCenter machine will use to connect to the agent.

   thrift_rpc_interface is the address that the agent will try to connect to Cassandra/DSE on via thrift. This should be the same as rpc_address in cassandra.yaml. The agent should be able to auto detect this, so you probably don't need to mess with it.

   Unfortunately right now there is a not a comprehensive list of options.

   Hope that helps

   Posted 1 year ago #
5. 

   Anonymous
   Unregistered

   Thanks,

   That got me going.
   Rik

   Posted 1 year ago #

RSS feed for this topic

Reply

You must log in to post.