Not sure, but I think this is what is going on -- need to dig deeper.
IIRC we call check_syslog during the instance shut down. But we always shut down the instance regardless of whether the test seemed successful or had already failed. If we can, we should probably check whether we're already in a failure cleanup, and if so, skip the syslog check, so that we can see the failed assertion.
Not sure, but I think this is what is going on -- need to dig deeper.
IIRC we call check_syslog during the instance shut down. But we always shut down the instance regardless of whether the test seemed successful or had already failed. If we can, we should probably check whether we're already in a failure cleanup, and if so, skip the syslog check, so that we can see the failed assertion.