Start background process with ssh, run experiments script, then stop it - bash

Question

I am running client-server performance experiments on several remote machines. I am trying to write a script to automate the experiments. Here is how it looks like (in a simplified way) for the moment.
for t in 0 1 2 3 4 5 6 7 8 9; do
cmd1="ssh user#${client1} runclient --threads=${t}"
cmd2="ssh user#${client2} runclient --threads=${t}"
$cmd1 &
$cmd2 &
wait
runclient connects to a server that I have started manually. It works fine, but I would like to automate starting and stopping the server as well. That means
Start the server in the background at the beginning of experiments
Run all the experiments
Stop the server at the end of experiments
I have found several suggestions but I am not sure which one is good for me exactly. Some recommend nohup, but I am not sure how to use it, and I don't understand why I should redirect stdin, stdout, and stderr. There is also the maybe the "-f" option to ssh to start a background process. In that case, how can I stop it later?
Edit: in response to the comments, the server is part of the performance experiments. I start it in a similar way to the client.
ssh user#${server} runserver
The only difference is that I want to start the server once, run several experiments on the clients with different parameters, and then stop the server. I could try something like that
ssh user#${server} runserver &
for t in 0 1 2 3 4 5 6 7 8 9; do
cmd1="ssh user#${client1} runclient --threads=${t}"
cmd2="ssh user#${client2} runclient --threads=${t}"
$cmd1 &
$cmd2 &
wait
But as the server does not stop, the script would never go past the first wait

Answer 1

Track your PIDs and wait for them individually.
This also lets you track failures, as shown below:
ssh "user#${server}" runserver & main_pid=$!
for t in 0 1 2 3 4 5 6 7 8 9; do
ssh "user#${client1}" "runclient --threads=${t}" & client1_pid=$!
ssh "user#${client2}" "runclient --threads=${t}" & client2_pid=$!
wait "$client1_pid" || echo "ERROR: $client1 exit status $? when run with $t threads"
wait "$client2_pid" || echo "ERROR: $client2 exit status $? when run with $t threads"
done
kill "$main_pid"