Discussion Locked

NT Route Problem

0 helpful
12 replies
Closed to new repliesThis discussion is closed. It stays readable, but it takes no new replies.

I am setting up CheckPoint Firewall on an NT 4.0 server. The server has 3 NIC's with the following addresses, 172.16.92.2 private and 172.16.93.2 private and the other NIC is going out the public area (Internet)with a 134.#.#.2 address.
From a PCwith an address of 172.16.92.7, I can ping any other 172.#.#.# address and the 134.#.#.2 address on the server but no other 134. address.
I have the reversed problem from a 134.#.#.7 addressed PC. I can ping any other 134 address and all NIC's on the server (172.16.92.2, 172.16.93.2 and 134.#.#.2) but no other 172 addressed device.
This seems like a simple thing to do but it has me baffled. I have been playing around with adding just about every route I can come up with.

Thanks for any help,
Larry

12 replies

by hstahl

This may be the obvious answer, but have you already gotten CheckPoint running? In other words, is it already processing the traffic between the groups? If so, IMCP (ping) is most likely not allowed. You just need to add a rule to allow it.

-Heath

by larcat

I stopped Checkpoint until I resolve this problem. I think with the firewall stopped I should be able to ping all devices on the network. Maybe I am wrong.

Thanks for your help.
Larry

by stillatit

What are you doing for NAT (Network Address Translation)? If you let those 172.16.x.x addresses out onto the internet, they will not work (your ISP's routers will block them). Presumably you have setup, either in Checkpoint or somewhere else, the facility to route the 172.16.x.x addresses to the internet while translating their addresses a valid (134.x.x.x ?) address. NAT shuffles port numbers to keep track of who is talking to whom so that replies go back to the station which sent the originaltraffic.

The problem with NAT is that you cannot usually originate traffic from the internet to a NAT'ed station, since there is no way the outside station can address the inside station (clearly the 172.16.x.x address will NOT work). For servers, some NAT implementations allow you to specify that traffic for specific ports (e.g., 80 for http, 25 for smtp) go to specific internal addresses (your servers). Typically this will NOT be set up for ping. If NAT is set up correctly, this explains why the 134.x.x.7 address cannot ping the 172.16.x.x addresses. The fact that it can ping the NIC's on the router suggests that NAT is not set up properly.

Go over your Checkpoint setup. You should have an external network interface defined, which presumably is the 134.x.x.2. (Note that this means that ALL stations on that segment and beyond are the untrusted world.) Double-check your Checkpoint documentation to make sure that you can have two separate trusted networks. In the Checkpoint GUI,set up rules to permit all traffic between the 172.x.x.x segments, and allow whatever traffic you wish to and from the 134 segment, with NAT (if Checkpoint is providing NAT). If the Checkpoint is NOT providing the NAT, you must have a NAT router between the Checkpoint and the final internet router.

Good luck.

If the 134.x.x.x range is a range of addresses you are using internally, you need to

by larcat

I do not think this is a CheckPoint problem. I have CheckPoint disabled. CheckPoint installation instructions says I have to have these networks routed in order for the firewall to work properly and I should be able to ping one device with another. NAT will come in after the firwall is started.

Thanks Anyhow

by stillatit

Looked at as a straight routing problem, the following immediately comes to mind:
What mask are you using on the various NIC's? 172.16.x.x is a Class B address, and the default mask would be 255.255.0.0, but that will NOT work, since it would place172.16.92.x and 172.16.93.x on the same segment. You should be forcing a class C mask, 255.255.255.0 on these subnets. Your routing commands to NT will need the MASK parameter.

Print the routing table on the NT box and check it. You should have the the following, more or less:

0.0.0.0 going to gateway router (default)
172.16.92.0 mask 255.255.255.0 gateway 172.16.92.2
172.16.93.0 mask 255.255.255.0 gateway 172.16.93.2
134.x.x. mask 255.255.255.0 gateway 134.x.x.2

If you see a maskof 255.255.0.0 on the 172 segments, that is part of the problem.

Good luck.

by larcat

This is my Route Print.
Thanks for the help.
Larry

Active Routes:
Network Destination Netmask Gateway Interface Metric
0.0.0.0 0.0.0.0 134.241.92.253 134.241.92.2 1
127.0.0.0 255.0.0.0 127.0.0.1 127.0.0.1 1
134.241.92.0 255.255.252.0 134.241.92.2 134.241.92.2 1
134.241.92.2 255.255.255.255 127.0.0.1 127.0.0.1 1
134.241.255.255 255.255.255.255 134.241.92.2 134.241.92.2 1
172.16.92.0 255.255.255.0 172.16.92.2 172.16.92.2 1
172.16.92.2 255.255.255.255 127.0.0.1 127.0.0.1 1
172.16.92.2 255.255.255.255 134.241.92.2 134.241.92.2 1
172.16.93.0 255.255.255.0 172.16.93.2 172.16.93.2 1
172.16.93.2 255.255.255.255 127.0.0.1 127.0.0.1 1

by bkduke

You may have 1 of 4 different problems here.

1. The firewall definitely needs to be started and rules setup to allow ICMP between the specified source and destination using a specific rule or by general firewall policy. Firewall-1 usually gets installed such that it will protect and not pass traffic untill it is up and running.
2. Make certain your interfaces are properly defined. Your 172.16.x.x interfaces need to have been setup with netmasks of 255.255.255.0 rather than the default 255.255.0.0 that the 172.16.0.0 network defaults to.
3. Routing. Make certain you have a route for the 134.x.x.x network. Do not assume that one is generated.
4. NAT. If you are using NAT for your 172.16.x.x addresses, make certain you have
a proxy arp address and route setup for their static or hiding address.

Hope this helps.

by stillatit

The route print looks ok.

Since the Checkpoint software is turned off, do you have IP routing turned on, or better yet RRAS? You need one or the other to have the NT act as a router.

Also, on the stations other than the server, make sure that you have the mask correct. You may also need to give each station explict routes to the 172 networks, since the default mask will mess them up too.

If a non-server box does not have the correct mask it will assume that all 172.16.x.x addresses areon its local segment and not send cross-segment traffic to the server.

Looking at one non-server computer (pick any one), check the following:
a) is its mask 255.255.255.0 in tcp/ip properties for the NIC?
b) if you do a ROUTE PRINT, what is the mask on its route to any 172 addresses? (if not 255.255.255.0 that is a problem)
c) can it ping all other computers on its segment? (if not, this needs to be solved before the server problem)
d) can it ping any computers on the other segment?

Note that if there are not a lot of machines on one leg or the other, you might solve this problem more easily by renumbering all boxes on one leg to 192.168.x.x.

Good luck.

This discussion is closed to new replies.