| CPUG | |
| The Check Point User Group | |
| A Resource For The Check Point Community. Fast. Useful. Independent. | |
|
| |||||||
![]() |
| | LinkBack | Thread Tools | Display Modes |
| |||
| I am probably missing something obvious here... I am running NGX R65 on SPLAT and have one external (dynamic) IP address. I have one internal LAN object defined with Automatic NAT, Hide behind Gateway configured. No problems there. I have "Translate destination on client side" checked for both Automatic and Manual NAT in Global Properties. I have a client in the LAN that requires inbound port 9940 mapped to it, so I have defined two manual NAT rules: outbound: [client - Any - 9940 | external IP object(hide) - Original - 9940] inbound: [Any - external IP object - 9940 | Original - client - Original] and then have the two required Security rules: outbound: [client - Any - 9940 - Accept] inbound: [Any - external IP object - 9940 - Accept] So the problem is when I fire up the application on the client that requires inbound 9940, the application reports to it's external connection server that the return port on my external IP is a high NATed port instead of the required 9940 that I think I have manually defined. Any ideas what I am missing? Thanks.. |
| |||
| When I change this manual rule to static: outbound: [client - Any - 9940 | external IP object(hide) - Original - 9940] and try to save or push policy this is what I get |
| |||
| After rebooting both the enforcement module and the smartcenter server, I have now pushed this NAT policy, but the client application still reports to the connection server a high NATed port instead of 9940. Any ideas? Thanks |
| |||
| You need to move the static rules above the Hide rules. Edited to add this: I usually create a group with all "my" networks (assuming I don't want to NAT between my own networks), then I configure a rule to avoid NAT like this (it's my first rule): All my networks | All my networks | any | = | = | = After that one I have all the static NAT rules. I always create them automatically when possible, because I think that is "cleaner". In the end I add a manual hide NAT rule which goes something like this: All internal nets | any | any | Hide fw | = | = I don't like the automatic Hide, because of all the extra rules it creates. Last edited by MarioL; 2008-01-09 at 03:17. |
| |||
| I pushed the following policy but no luck. The app still reports a high NATed port to the connection server instead of 9940. Do you think client side NAT translation has anything to do with it? |
| |||
| I'm getting a little confused here. Is this all correct? Your client is on your LAN and needs to accept connections from the Internet on UDP 9940. Your client on the LAN needs to be able to go outbound anywhere on the internet using UDP 9940. This application is located somewhere on the Internet. Ray |
| |||
| As Ray says, can you please clarify... please detail exactly how the communication between client/server works and where each is, who initiates the connection, etc. PS Translation on the client side only changes when the NAT happens, it doesn't affect the end result "on the packets". |
| |||
| If I understand you correctly then the Server is on your network behind a R65 box with one external dynamic IP address. Do you actually need to initate a connection from the server to the client, other then reply traffic to the clients request. I would look at the SRV_REDIRECT function (look at http_mapped) for how to use this. That way rather than NATting the traffic it just redirects essentially port mapping the traffic rather then NAT as such. VPN-1/FireWall-1 can perform Port Address Translation (PAT), and includes predefined Port Mapping Services. Connections are directed to the firewall module, accepted on a given port and translated to another, then routed to an internal server, when Port Mapping Services are defined and configured. This occurs transparently to users. Procedure: Configuring predefined TCP Port Mapping Services Log into SmartDashboard. Click 'Manage > Services'. On the Services dialog box Click the drop down menu next to 'Show:', select 'User defined services' and choose the desired Port Mapping Service (e.g. http_mapped). Click the 'Advanced' button. In the 'Match' section configure the IP address of the internal server utilizing port mapping, and the mapped ports. Example: Default SRV_REDIRECT(80,0.0.0.0,80) Modified SRV_REDIRECT(80,10.20.1.5,8080) Click 'OK' and close all screens. Configure the rule (see Rule Base configuration). Rule Base Configuration Source: Any Destination: firewall_object Service: Port Mapping Service (e.g. http_mapped) Action: accept Install the Security Policy. |
| |||
| RayPesek and MarioL The app I'm trying to use is iVisit video conferencing. The client app is on a PC behind my R65 firewall. You load the client app and then login to the iVisit connection server on the Internet. It tells that server what external port other iVisit clients can directly connect to you on. There's a network stats window you can open in the app to show what it is reporting for example: "external 10.1.1.1 (10654) : internal 192 168.1.1 (9940)" I can't figure out what R65 NAT rules will allow the client app to detect that 9940 is mapped in to it. Instead it always reports to the server that it can be reached on a random high NATed port. I know the client app correctly reports the outside port because if I login to the internet connection server through a cheap NetGear router, it reports 9940 as the external reachable port. Can't get it to do the same through R65. mcnallym I thought SRV_REDIRECT only worked for TCP and not UDP.. |
| |||
| So far I haven't found any manual NAT rules or SRV_REDIRECT combinations that have solved this problem. Could it be something to do with the fact that I have a dynamic external address? Thanks |
![]() |
| Thread Tools | |
| Display Modes | |
| |