Fix multiple block insertion issue in distributed tables

Why it matters

This feature addresses issues with inserting multiple data blocks into distributed tables under the configuration insert_distributed_one_random_shard = 1. It ensures reliable and correct insertion behavior, enhancing stability and performance when distributing inserts to a random shard.

How to use it

Enable the setting insert_distributed_one_random_shard = 1 to direct inserts to a single random shard in distributed tables. This improvement is automatic with the new version and requires no further action from the user.