Developer Curb Appeal: SolrClould doc improvements

There are some things that need cleanup in the doc.

• Clarification about single machine multiple SolrCloud instances vs. multiple machines each with a single SolrCloud instance

• Clarification about which command line switches are needed for each environment.

• Clarification about the "shard" copies you have to make, and their names, and how shards are allocated upon first startup, and what happens if they're not empty, etc. (assuming we don't remove the need for doing this)

• Clarification of the -Dbootstrap_confdir and -Dcollection.configName options in the getting started guide

• Explanation of Zookeeper's persistence of shard number allocation and config files cache on the local node

• Discussion of Firewall issues and debugging that can look like other network problems, for example all 3 of these can be a firewall issue:
  "Connection Refused"
  "No Route To Host"
  "Permission denied"
  Also a reference to using telnet to a machine and port, use of the ZK "stat" command, and links to references on installing telnet client and fixing firewall issues on Windows, Linux (CentOS), Mac, and maybe others.

And alignment of http://lucidworks.lucidimagination.com/display/solr/Getting+Started+with+SolrCloud and http://wiki.apache.org/solr/SolrCloud (and other doc?)