Small issue with clustering about priam HOT 5 CLOSED

It looks like it is an interface issue. nodetool -h localhost gossipinfo is showing that it is using localhost for gossip. The cassandra.yaml created by Praim has listen_address and rpc_address set to null. Is this correct?

from priam.

Ok I have discovered the issue I think. Cassandra is being loaded using the yaml file located in the JAR and not the one located at /etc/cassandra/conf/cassandra.yaml. Is there a way to prevent this?

from priam.

You need to reverse the CLASSPATH ordering in /etc/init.d/cassandra
Out of the box it reads $cp:$CONFDIR
Change that to $CONFDIR:$cp to make it read the /etc/cassandra/cassandra.yaml and NOT the cassandra.yaml that is included in the priam.jar
Had the same issue! Let me know if you could get it sorted!

from priam.

You got it. That has fixed my problems.

Thank you very much! 👍

from priam.

I seem to be having the same problem as Mbarrin , Ive tried with both dse 2 and dse 1.2 with same result cassandra comes up as two separate nodes but no cluster is present. I'm running amazon linux and the Etc/init.d does not have a class path entry however I did find a entry in the cassandra.in.sh however changing made no difference . Mbarrins description is same as mine to a tee , sdb works, s3 works however hosts show only themselves bound to 127.0.0.1

Additionally the curl get seeds returns no value so I'm thinking perhaps this is the hosts file configuration .

To echo is there a additional step I need to apply ? Or alternate to /etc/init.d/cassandra ????

from priam.