Zathras reports failed tests as successful

[grdumas]

I experienced a test wrapper failure which was reported in the run output, but the final status line reported that all tests succeeded.

Failure detected in output:
TASK [test_generic : We failed] ************************************************
ok: [perf-intel-23] => {
"msg": "//root/workloads/autohpl-wrapper-1.23/auto_hpl/build_run_hpl.sh reported failure."
}

Final output line reports all tests succeeded:
All tests executed sucessfully

The expected behavior is that test failures and successes would be accurately and consistently reported.

[grdumas]

The current version of Zathras continues to report that all tests execute successfully, even when no test ran successfully.

Example:

PLAY RECAP ************************************************************************************************************
localhost                  : ok=29   changed=20   unreachable=0    failed=1    skipped=25   rescued=0    ignored=0   
perf-intel-23              : ok=32   changed=17   unreachable=0    failed=0    skipped=40   rescued=1    ignored=0   

Error: test started, system died.
Finsihed streams,pig,auto_hpl,coremark,coremark_pro,passmark,pyperf,specjbb,speccpu2017 on perf-intel-23
ls: cannot access 'results_auto_hpl*': No such file or directory
ls: cannot access 'results_coremark_pro*': No such file or directory
ls: cannot access 'results_coremark*': No such file or directory
ls: cannot access 'results_passmark*': No such file or directory
ls: cannot access 'results_pig*': No such file or directory
ls: cannot access 'results_pyperf*': No such file or directory
ls: cannot access 'results_speccpu2017*': No such file or directory
ls: cannot access 'results_specjbb*': No such file or directory
ls: cannot access 'results_streams*': No such file or directory
ls: cannot access 'perf-intel-23*/results*auto_hpl_*tar': No such file or directory
All tests executed sucessfully