I've got classical eventual consistency symptoms (read after write returns empty result) but there is a surprising twist. The keyspace has replication factor 1 (it's used as a cache) so how can I get a stale result?
Cassandra version 1.2.15. Consistency settings (although I think they should not matter with one-replica case): Read — CL.ONE Write — CL.ALL If you need any additional info I would be happy to provide!