cbcollect: --single-node-diag should default to true

By default, cbcollect_info collects diag information from all reachable nodes in the cluster. Note that this is true only when run from the command line.

It's quite common for users to invoke cbcollect_infor approximately simultaneously on all the nodes in the cluster - which can result in all nodes processing diag information for all other nodes at about the same time - which can further result in memory spikes and sometimes memory allocation problems.

We should change cbcollect_info to default to just collecting diags from the node it's run on - which is a more natural semantic anyway.