We run an 8 node Cassandra v2.1.16 cluster (4 nodes in two discrete datacentres) and we're currently investigating a problem where by restarting Cassandra on a node resulted in the filling of Eden/Survivor/Old and frequent GCs.

http://imgur.com/a/OR1dk

This hammered reads from our application tier (writes seemed okay) and until we determine what the root cause is, we'd like to be able to start Cassandra with gossip disabled.

Is this possible?
Thanks
.bN

Reply via email to