We have only one server (Win 2008 R2) in our site A's LAN with its own domain/forest. This server is a DC and DNS. Site B has its own domain/forest as well.
Between A and B is a site-to-site VPN by a Sonic Wall (in A) and a Cisco ASA (in B). During the most of off-hours when no users access the network resources, I found from B I can NOT ping any A's computer or device using their internal IP. At the same time I can NOT ping the Sonic Wall's internal IP address either while however I can ping its public IP address. (Of course during the work hours when users connect I can ping everything at A from B.)
At this moment, to my surprise I still can use Logmein to access one of the computers inside A. And then from there when I ping any B's host name (not IP, see note below), it will get response normally, and the most amazing thing is, as soon as the ping gets the first response, the above issue gets resolved immediately until hours later (still during off-hours) the issue will occur again. Is it something to do with the server, the Sonic Wall, or some kind of sleep mode, or what? I just totally have no clue. The server doesn't put hard drive to sleep.
Why does this issue happen? Can you shed some light?
Note: If I instead ping any host in B using IP address when Logmein to A as mentioned above, it will get responses normally but it will NOT do any good to resolve the above issue. So strange, right?