Fix handling of state killing on WAN failure
For 2.1, we need to change the state killing on WAN failure to wipe the entire state table. The current state killing only gets the outbound states, leaving the internal states, which causes problems with VoIP failover and any other long-lived connection (especially UDP). Wiping the entire state table is overkill, but will have to suffice for 2.1.
A bit too excessive need to get right.
This reverts commit c59dd719e0a6d9ee8deecaa7bff0d6ee8c76e4ca.
Ticket #3181 do the state flushing only on down gateway detection rather than any time.