README

This test tests the appsock exhausted code, which attempts to retry on different nodes various times before erroring.
Previously when receiving an appsock exhausted error, cdb2api would not retry on a different node which could accommodate the connection.
This test only queries one node from the cluster and just checks that a connect error is received now instead of an appsock error.

We cannot test this using CDB2_RANDOM on a cluster and see if there exists a node that hasn't reaced the maximum number of connections since the old code could also get to the maximum number of connections on each node by connecting to a random node each time. Would need to use CDB2_RANDOMROOM and have a cluster with nodes in different rooms.