0

I have OpsCenter 5.0.2 running on a node, and Cassandra 2.1.2 and the Agent 5.0.2 running on another node. I installed both OpsCenter and the Agent using yum/rpm on CentOS 7.

I have no problem using OpsCenter with this agent when I run the Agent's binary directly. However, when I run the datastax-agent service, OpsCenter never sees the connection. Attempting to install the agent from OpsCenter never worked (I always received a "lost connection" error).

I've looked at startup.log and everything looks normal: it loads the confiuration and starts Jetty. agent.log also looks normal: it says it has connected to the cluster, gets to Using partitioner: Murmur3 and that's the end of it. I'm not sure if the logs should say more at this point.

I've also looked for any errors in journalctl as well as systemctl status and don't see anything of interest.

Any ideas on where else to look?

jyoung
  • 101
  • 2
  • When you say "I run the Agent's binary directly", are you running it from `/usr/share/datastax-agent/bin/datastax-agent` or somewhere else? My first guess is that the address.yaml file you have configured that's being used when running the binary directly is different than /var/lib/datastax-agent/conf/address.yaml, which the service should be using – mbulman Jan 03 '15 at 20:31
  • Yes, by running the binary directly I meant running /usr/share/datastax-agent/bin/datastax-agent Through trial and error I determined this binary was using /etc/datastax-agent/address.yaml and ignores /var/lib/datastax-agent/conf/address.yaml, as you said. I have since copied address.yaml from /etc/... into /var/... and started the service again, but it still isn't reporting any data to OpsCenter, and OpsCenter has 0/1 agents connected. Starting the binary again from /usr/share/... reports data normally and shows the agent is connected. – jyoung Jan 05 '15 at 05:57

0 Answers0