run_sys_tests: add more clarity about whether it successfully launched the test jobs

There are two related enhancements that we should make to run_sys_tests:

(1) If the jobs were launched successfully, we should print a clear message to that effect. New users of this script find it confusing that there is a bunch of output then no clear message about whether things got launched successfully or not: it is unclear whether the normal output is expected or is signaling some issue. (Alternatively, we could make run_sys_tests less verbose by default, changing the prints in run_sys_tests to logger.info calls, and requiring you to run with `-v` if you want all of the current output. My feeling is that the current output is something that I always want to see, and assumed the same would be true for others, so I have treated this output as part of the normal output of the script and put it in print statements so it always appears. But I am open to hiding the output by default and always specifying `-v` when I run it myself if people prefer that. Then it would be following the Unix standard operation of "silence indicates success", and there would be less of a need for a line saying "success" at the end.)

(2) @glemieux just ran into a situation where the create_test job failed to launch because the cheyenne account was overspent. run_sys_tests did not make this clear unless you ran with the `-v` option. run_sys_tests should probably abort with an error message in this case. I think this would involve changing this:

https://github.com/ESCOMP/CTSM/blob/3dcbc7499a57904750a994672fc36b4221b9def5/python/ctsm/joblauncher/job_launcher_qsub.py#L31-L32

to abort with an error message if an error occurs from the qsub job (and similar for the other job launchers).