Try a restart on that guy with and with the JVM option -dcassandra.load_ring_state=false (see cassandra-env.sh).
If that does not fix it…. http://www.datastax.com/docs/0.8/troubleshooting/index#view-of-ring-differs-between-some-nodes Cheers ----------------- Aaron Morton Freelance Cassandra Developer @aaronmorton http://www.thelastpickle.com On 19/08/2011, at 5:49 AM, Ian Danforth wrote: > All, > > One node in my cluster returns a different answer for nodetool ring. It > seems to be ignorant of one of the other nodes. I've verified that all nodes > can ping each other without difficulty. > > Here is the proper response: > > Address DC Rack Status State Load Owns > Token > > 127605887595351923798765477786913079296 > 10.123.7.163 datacenter1 rack1 Up Normal 22.21 GB 25.00% > 0 > 10.223.7.129 datacenter1 rack1 Up Normal 19.75 GB 25.00% > 42535295865117307932921825928971026432 > 10.124.139.202 datacenter1 rack1 Up Normal 20.63 GB 25.00% > 85070591730234615865843651857942052864 > 10.90.229.212 datacenter1 rack1 Up Normal 20.56 GB 25.00% > 127605887595351923798765477786913079296 > > And here is the out of sync response from .129: > > Address DC Rack Status State Load Owns > Token > > 127605887595351923798765477786913079296 > 10.123.7.163 datacenter1 rack1 Up Normal 22.21 GB 25.00% > 0 > 10.223.7.129 datacenter1 rack1 Up Normal 19.76 GB 25.00% > 42535295865117307932921825928971026432 > 10.90.229.212 datacenter1 rack1 Up Normal 20.56 GB 50.00% > 127605887595351923798765477786913079296 > > How can I get .129 to acknowledge the proper state of the ring? > > Thanks in advance, > > Ian