We recently had an issue where the clock on our Jenkins server was ahead by about 15 minutes which caused the test result files (that are generated on other machines and shared over NFS) to appear 15 minutes old to Jenkins. In one case where the test suite takes less than that 15 minutes it correctly guessed that something was wrong and displayed a message like:

10:20:01 Test reports were found but none of them are new. Did tests run? 
10:20:01 For example, /gscmnt/gc13001/info/model_data/apipe-ci/workspace/0-Genome-Perl-Tests-Quick/PERL_VERSION=5.10/PERL_VERSION/5.10/test_results/Model/Tools/Validation/Sciclone.t.junit.xml is 15 min old

However, for another test suite that takes about 40 minutes it appears to have just silently ignored old files which caused it to report all tests as passing despite one of the tests failing.

From the console log:

...
18:40:08 ***** ./Model/Command/Services/WebApp/Core.t STDERR *****
18:40:08 <<Waiting for dispatch ...>>
18:40:08 <<Starting on blade10-2-11.gsc.wustl.edu>>
18:40:08 Job <5231505> is submitted to queue <short>.
18:40:08 *********************************************************
...
19:07:37   <testsuite failures="1"
19:07:37              errors="0"
19:07:37              time="0.0159821510314941"
19:07:37              tests="3"
19:07:37              name="Model_Command_Services_WebApp_Core_t">

But then the test report make no mention of this test. It also appears to be missing about 6k tests from the normal 16k tests run. Allowing a test failure to be reported as success seems absolutely undermining. Is it possible I could have configured Jenkins to prevent this problem? Or is this possible a critical bug (that has a low? chance of occurring)?