Consider the following situation:
The sender has .not_terminated file. It gets disconnected. The last trail
file is then terminated without adding new data (this can happen for example
when auditd is being stopped on the sender). After reconnect the .not_terminated
was not renamed on the receiver as it should.

We were already handling similar situation where the sender crashed and the
.not_terminated trail file was renamed to .crash_recovery. Extend this case to
handle the situation above.