I have created a preliminary run_mms_unit_tests.ksh script to periodically run the MMS mgunit test suites out of cron. It's based on the mms_run_all_tests routine in SPEDAS, but duplicates most of the logic rather than calling mms_run_all_tests.
It works, but needs some refinement to be more usable:
[x] Test a fresh bleeding edge copy, rather than the socware installation, which may be inconsistent or out of date
[x] Postprocess test results to identify failed tests, and log/email to SOC developers. It would be helpful if each individual failure included both the name of the failed test, and the suite it came from (so we can just "grep fail test_results" and get an actionable summary of what failed).
[ ] Accept a list of tests to be run via the command line (for quick retests), or default to 'all'
[ ] Add suffix to working directory to indicate which test group (so log rotation will keep the previous run for each group)
I have created a preliminary run_mms_unit_tests.ksh script to periodically run the MMS mgunit test suites out of cron. It's based on the mms_run_all_tests routine in SPEDAS, but duplicates most of the logic rather than calling mms_run_all_tests.
It works, but needs some refinement to be more usable: