Cluster chokes on 600+ connections

We have an active director with 256 MB RAM, and two nodes.

When I do a stress test, the client starts getting timeouts when the
number of *inactive* connections hits 600 or so. If I take out a node,
the same number of inactive connections is easily handled by a node, but
it is still a problem for the director.

The nodes and the director are connected on the hub, and the director is
the default gateway. The nodes are also connected to the rest of the
network on the other interface (they need to see the database, etc).

Can hardware limitations cause this problem, RAM in particular?

Thank you,

-- 
'Drey



-- 
Ultra Monkey - http://www.ultramonkey.org/
To UNSUBSCRIBE, email to [email protected], with a body:
unsubscribe ultramonkey-users [email protected]
where "[email protected]" is YOUR email address.