The logback SocketServer and SocketServerReceiver component
vulnerability is with the serialization when using network configuration
for centralized log aggregation. The default Cassandra logback
configuration uses files on disk and does not use these features nor
expose logging to the network. If
We tried to extract large volume of data from a 42-node cluster about three
times and in all attempts, client sessions aborts after ~ 60 hours.
Here's what we see in in the client logs
I have reviewed the multiple timeout settings in C*, but none seemed to
relate to the 60 hrs limit.
What is pot
