Hi all, I'm working on getting a new cassandra implementation up and functional. We're running cassandra 2.0 on Centos 6.4.
Right now, the issue that we've run into is that counters are vastly different depending on what hosts they're hitting. Our keyspace has a replication factor of 2, which we increased after creating the keyspace. We have row and key cache turned off because we were hitting -- https://issues.apache.org/jira/browse/CASSANDRA-4687. I'm using a python script that goes through and counts a specific key. With read consistency of 1, I get inconsistent counts. I would expect some amount of inconsistency, but here's the counts for one value: 7077 57352 If I raise the consistency level to 2, I always get the higher number. Running node tool repair has made no difference to the results. How can I figure out why there's such a huge difference in results on one node and not on the other?
