Hi all,

This has happened to me on quite a few occasions already but i've yet to find a solution for it. hopefully someone here can help solve the problem or at least point me in the right direction.

I'm running a stress test with 100 users on a web application with 3 physical boxes (2 web/app win2k servers load balanced by a hardware LB) and a db unix (solaris) server.

I've got rational agents setup on all three boxes and my test manager is able to ping them just fine using direct IPs.

When I run my test however, i can collect server resource stats from the db server and just one of the web/app server. The only hint of the problem is during the initialisation of the run, the status in test manager will show the test manager connecting to the problematic server and it will time out after about 1 or so min (the rest takes barely 3 secs to connect and initialise).

I'm using the latest Test Manager and agent build v2003.06.14. I've also made sure that its not a firewall/network problem by pinging the servers from the test manager and running the netstat command on the test manager during the run (netstat shows that connections have been established with all 3 servers).

I've also ruled out problems with the LB by bringing down the web/app server that is responding.

I'm at wits end now. hopefully can pick up a few tips from here on how to proceed with debugging this issue further.