On Linux, how do I determine the number of processes with the exact string "test test 123" anywhere in the process title?

TL;DR: use pgrep -cf "test test 123"

The ps program will list all running processes. Specifically, try:

ps aux 

Now, you could just filter that list using grep to search for your string:

ps aux | grep "test test 123" 

That will print out matching lines. To count them, use grep -c which prints out the number of matching lines instead:

ps aux | grep -c "test test 123" 

The problem with this approach is that the grep process above will also appear in the results. For example, I am currently editing a file named test test 123, but if I run the command above, I'll see both the process of my file editor and the grep itself:

$ ps aux | grep "test test 123" terdon 2461453 22.0 0.2 392944 79796 pts/1 Sl 15:53 0:02 emacs test test 123 terdon 2462354 0.0 0.0 8832 2292 pts/1 S+ 15:53 0:00 grep --color test test 123 

Therefore, the grep -c will return 2 instead of 1:

$ ps aux | grep -c "test test 123" 2 

Which brings us to the right tool for the job, pgrep. This is a tool specifically designed to find processes:

$ pgrep -cf "test test 123" 1 

The -c means "count the matches" and the -f means "search the entire command line, not just the process name".

The other common trick to skip the grep itself is to use a one-character character class instead of the same string so that the grep command line won't contain the string:

$ ps aux | grep "test test 123" terdon 2461453 1.2 0.2 392944 79796 pts/1 Sl 15:53 0:02 emacs test test 123 terdon 2476971 0.0 0.0 8832 2236 pts/1 S+ 15:56 0:00 grep --color test test 123 $ ps aux | grep "[t]est test 123" terdon 2461453 1.2 0.2 392944 79796 pts/1 Sl 15:53 0:02 emacs test test 123 $ ps aux | grep -c "[t]est test 123" 1 

For more on this trick see here. But this really isn't necessary if your system has pgrep as Linux systems do.

echo $(( $(ps aux | grep "test test 123" | wc -l) - 1)) 

should to the trick