Description of problem: Seems that *AVC subtest* looks for audit messages generated also before the test case run. As a results after first AVC denial all following AVC checks fail. Version-Release number of selected component (if applicable): https://beaker-stage.engineering.redhat.com/jobs/12589 How reproducible: always Actual results: each *-avc.log contain all AVC denial messages Expected results: each *-avc.log contain just AVCs reported during test case run Additional info: Blocks RHEL7 Alpha-1 testing
Marian, Let me know if there is anything you need for this. Systems are available on beaker-stage to reproduce this.
This is an interesting bug, but as far as I can tell beah/rhts is doing everything correctly. Here is what is happening.. 1 - Hardware clock is off by one hour or so. 2 - System boots and spews some selinux errors that are in the "future" 3 - ntpd starts and resets the time back to the correct time slot 4 - beah runs and does ausearch using correct start time and is picking up those boot up errors until we progress past the corrected time offset. I was able to "fix" this by setting the systems hardware clock to the current time and then re-running the test. Interestingly the original selinux error is no where to be found now since it happened before we started testing. So its good that this happened and means we should probably have the install task or some other task check for selinux errors from before we ran. I'm going to close this as "not a bug". Thanks!
*** Bug 797177 has been marked as a duplicate of this bug. ***
http://gerrit.beaker-project.org/#/c/1235/
Beaker 0.9.2 has been released.