Resolution:
Description:
============
Answers to frequent ActiveSpaces configuration questions.

Environment:
===========
All ActiveSpaces versions and platforms

Resolution:
==========
You do not need to list all of the members of the metaspace in the discovery URL,
just enough that you feel it gives you enough fault tolerance. Typically, list only
the seeders in the discovery URL.

If client (seeder or leech) provides its own listen port as the first entry in the
discovery list, the operation starts relatively faster. For every seeder,
the first entry in the discovery URL needs to be it's own listen url:port. This
condition was eventually removed but now works fine irrespective of the order
of discovery URLs.  

The best discovery node candidates are the nodes that should 'always be up' (e.g.,
automatically started when the machine boots up or monitored by say, Hawk,
so that they are automatically restated, etc.). Typically as-agents would be good
candidates for being discovery nodes (as you are likely to have them automatically
started/restated).

Include enough nodes in the list so that you have a good degree of confidence that
at least one of those nodes will be up at all times, two per machine is reasonable.
Listing all of them does not really hurt but may mean extra administrative
work. Remember that all nodes in the metaspace must have the same list of discovery
nodes (though not necessarily in the same order) so if you list all your agents and
decide to add more agents you would have to go and modify the discovery URL for all
the processes that connect to the metaspace.

ActiveSpaces Discovery URLs Configuration Best Practices.