VPN Site Dropping calls after UG to 4.0(2)

[midpenn]

We recently upgraded from 3.3(4) to 4.0(2) have two remote sites connected via VPN with 1760-V routers and FXO ports for local lines at the remote. However since the upgrade there has been a high number of dropped calls at the remotes. Calls are almost always dropped when the remote site puts a caller on hold or park. In some cases calls are disconnected while conversation is going on. These problems did not exist with 3.3(4). Any suggestions / thoughts? Lines are loop start. I've monitored the VPN connection and there does not seem to be a problem with it...

[ssoberlik]

This could happen because of many reasons:

1. ISP dropping packets/traffic at certain times of the day

2. VPN clients behind the PAT device loose their translation on the PAT device

3. IKE Keepalives and/or DPD messages are blocked and thus drop the IPSec SA session

[JOSE TORRES]

What are you using to control the remote ports, MGCP or H.323? Do you have any FXO ports locally not over the VPN and are they experiencing the problem?

[Steven Smith]

How do calls across the WAN behave? If you have problems here and ar running skinny fixup on your firewalls, this could cause a problem. You might check the NIC settings on the callmanager for speed and duplex, but that is a stretch.