It was found that in some circumstances, the client can fail to update to the changed cluster topology if there was a failure and either the instance wide client timeout or the operations the client is doing have a short (i.e., less than 700ms or so) timeout. JCBC-88.
It was found that in the dependent spymemcached client library that errors encountered in optimized set operations would not be handled correctly and thus application code would receive unexpected errors during a rebalance. This has been worked around in this release by disabling optimization. This may have a negilgable drop in throughput but shorter latencies.
If using the CouchbaseConnectionFactoryBuilder, authentication could be ignored by the client library.
Ensure the configuration is updated even if the node the client is intended to receive updates from fails without closing the TCP connection. Make sure these configuration checks are reliable and do not consume significant resources.
As a workaround to a reliability under rebalance issue, optimization has been disabled in this release. This can cause a negligible drop in throughput but have better per-operation latency.