We have some LWTs that worked well before we migrated to the latest version of the driver 4.9.0. After it, we started to see lots of messages about timeouts during CAS operations. Timeout on the CAS operation is configured and this happening earlier then timeout could be exceed.
Moreover, we followed this issue https://issues.apache.org/jira/browse/CASSANDRA-9328 and we're doing serial read after unsuccessful write, but we have an empty record at the end.
Could you give a hint if driver migration could be the issue of this problem?