Shard awarness does not work with batch statements

[gleb-cloudius]

In case of token awareness the driver uses first statement in a batch do decide. Shard awareness should do the same.

[gleb-cloudius]

It may be lwt related. To reproduce run upstream cassandra-stress with our driver like that:

cassandra-stress user profile=cqlstress-lwt-example.yaml n=10000000 ops'(insert=1)' cl=QUORUM -mode native cql3 protocolVersion=3 -rate threads=200

cqlstress-lwt-example.yaml is in Cassandra sources

[gleb-cloudius]

Looks like protocolVersion=3 was the problem. Without it shard awarness works for batches as well. Is this expected?

[mykaul]

I think we use v4 today, so is this still an issue?