Hi,
I wonder if anybody could shed some light on this. I am stumped.
Brief overview, checkpoint gateway (R77.30 Gaia) in a DC routing traffic from inbound policy based vpn out of route based vpn to AWS. We have several of these all working fine. This one involves NAT, none of the working vpns are doing NAT.
Connection to a web server in AWS works from Endpoint connect using the "real" ip of the server.
SmartView Tracker shows vpn routing
Fw monitor shows my client connection to aws host.
o shows vpnt15 which is the tunnel interface to aws eth4 is the internet interface.
O shows eth4 - this is correct and works with this connection and other external vpns routing to aws
[vs_0][fw_0] eth4:i[52]: 172.21.254.52 -> 172.29.146.13 (TCP) len=52 id=49031
TCP: 57971 -> 80 .S.... seq=1ef280d9 ack=00000000
[vs_0][fw_0] eth4:I[52]: 172.21.254.52 -> 172.29.146.13 (TCP) len=52 id=49031
TCP: 57971 -> 80 .S.... seq=1ef280d9 ack=00000000
[vs_0][fw_0] vpnt15:o[52]: 172.21.254.52 -> 172.29.146.13 (TCP) len=52 id=49031
TCP: 57971 -> 80 .S.... seq=1ef280d9 ack=00000000
[vs_0][fw_0] eth4:O[52]: 172.21.254.52 -> 172.29.146.13 (TCP) len=52 id=49031
TCP: 57971 -> 80 .S.... seq=1ef280d9 ack=00000000
Fw Monitor below shows connection from the customer end to nat address which is translated on checkpoint to real destination address. Source address remains as original.
SmartView Tracker shows Decrypt and nat translation by the configured by the nat rule. Not vpn routing.
Note - o shows vpnt15 but O also shows vpnt15 - traffic does not arrive at AWS (verified by aws flow log)
[vs_0][fw_0] eth4:i[72]: 172.30.21.144 -> 10.150.194.16 (TCP) len=72 id=17563
TCP: 40203 -> 80 .S.... seq=c909e140 ack=00000000
[vs_0][fw_0] eth4:I[72]: 172.30.21.144 -> 172.29.146.13 (TCP) len=72 id=5445
TCP: 40439 -> 80 .S.... seq=03d3dd1b ack=00000000
[vs_0][fw_0] vpnt15:o[72]: 172.30.21.144 -> 172.29.146.13 (TCP) len=72 id=5445
TCP: 40439 -> 80 .S.... seq=03d3dd1b ack=00000000
[vs_0][fw_0] vpnt15:O[72]: 172.30.21.144 -> 172.29.146.13 (TCP) len=72 id=5445
I have added the subnet being used for nat translations to the encryption domain on checkpoint and the access rule allows traffic from the customer to both the nat address and the real address. Initially i had only the nat address but added the real one to see if it made any difference. It did not.
VPN column is set to any and disable nat inside vp community is not ticked.
The nat subnet is not in use anywhere else.
The aws host is behind an nlb - there is no security group on the nlb so all traffic is allowed.
AWS route tables have been updated with customer networks.
I can provide more information if needed.
Any suggestions would be appreciated.
Regards
Mark