Currently, when a VM test case completes (either successful or failed), the associated VMs are destroyed. All associated information, including any log files that were generated during the test and that go beyond what's captured in the console, log.html and output.xml, are deleted.
Test runs should store pertinent log data on some sort of persistent storage. Options include
- writing logs to a directory that is NFS mounted,
- sending logs via syslog to a remote server,
- other suitable options for storing logs outside the TG/SUT itself.
Ideally, a common strategy should work for VIRL VMs, hardware testbeds, Vagrant VMs, and LF Nested VMs.