test environment: test --persist-tun with ping-timeout

[cron2]

So what we lack testing for is ping-timeout reconnects with DCO involved

• --client side
• connect
• ping-timeout (30s)
  (triggerable via kill $CN on the server's management interface, or possibly via firewall rule in between - though timing is complicated if you want reconnect to succeed)
• reconnect
• ping-timeout again (30s)
• eventually tls-renegotiate (300s)

and verify that pings through the tunnel are working all the time as one would expect this - fail while the server's instance is dead, but do not fail longer than 30-ish seconds.

Maybe also check the server's idea of "when did the client reconnect?" - which gives a good indication of "did ping-timeout actually work?".

The issue at hand is on FreeBSD, but I guess that having test coverage on Linux and Windows won't hurt either...

[cron2]

so basically I think this needs server-side assistance, read: a script that sits on the management interface and keeps track of a given instance, killing it twice, and then letting it run to a client-side exit. Ideally controlled by a --setenv UV_* from the client side.

Combine this with a long-running ping on the client, correlating "ping status change" (working<->non-working) with log file timestamps...

Doable, but quite a bit of work.