I want my test plan to be able to vary its behavior slightly from server to
server (in a controlled fashion) when running a test plan against multiple
jmeter servers simultaneously. So far I've done this by defining
-Jserver_num=X on each jmeter-server on startup with a unique incremented value
for X...then in my test plan I make decisions based on ${server_num}.
Is there a better way to do this? It's a little annoying to keep in sync, and
since I don't really care if the server_num stays the same from run to run it'd
be better if the jmeter client could just assign a server_num to each remote
server before it runs a workload and send it to them (along with a server_count
variable for cases where I want to do decisions like i % server_count ==
server_num).
Just one example...perhaps I have 100,000 users with unique usernames and I
want to ensure that each user sends exactly one request, but I want to spread
the load across 5 jmeter servers. Server 0 should run user[0..19,999], server
1 should run user[20,000.. 39,999], and so on.
Thanks,Chris
