Sometimes, the Cassandra driver we're using would get stuck in an infinite loop while attempting to prepare queries. This is what we're observing :
- The driver tries to prepare a query
- Keyspaces responds with a query ID
- The driver tries to execute the query on the same host using the same query ID
- Keyspaces responds with
Unprepared
- Go to 1
We're seeing an infinite loop of such messages:
Unprepared "Prepared query with ID 0d373b6eedea095de8a5f27b50df44e5 not found (either the query was not prepared on this host (maybe the host has been restarted?) or you have prepared too many queries and it has been evicted from the internal cache)"
This bug is a bit hard to reproduce, I have a sample program that triggers it about one time out of 10 runs, always starting with the same host.
My question is, are there any limitations or known issues causing prepared queries not to register as expected ?
