We have found that sometimes the LXC hosts used by the Jepsen tests are not always running. This creates a system failure in Evergreen. In order to understand the problem better, when the failure is detected we'll tail /var/log/lxc/n<host>.log to the task log.