IT-VAS
asked on
How to setup rexec with forefront TMG
Hi,
We are implementing a solution where we are placing a forefront TMG machine between our clients
and our server and the clients are on a separate VLAN.
Now we are experiencing problems when the clients try doing a Rexec against our server and we believe
it has something to do with the fact that Rexec starts on port 512 but later on switches communication
to the stderr port and when this happens our server don't get a correct SYN ACK back from the client and
eventually the server times out the connection since no SYN ACK has reached the server.
Have anybody implemented a solution with forefront TMG and rexec that knows something about this?
We are implementing a solution where we are placing a forefront TMG machine between our clients
and our server and the clients are on a separate VLAN.
Now we are experiencing problems when the clients try doing a Rexec against our server and we believe
it has something to do with the fact that Rexec starts on port 512 but later on switches communication
to the stderr port and when this happens our server don't get a correct SYN ACK back from the client and
eventually the server times out the connection since no SYN ACK has reached the server.
Have anybody implemented a solution with forefront TMG and rexec that knows something about this?
ASKER
Hi,
Thank you for the answer. We have thought of that solution, but the trouble is that the rexec protocol seems to choose
different stderr ports every timy in the range from 1024 and above so we don't know how to solve it without opening
a lot of ports in the forefront TMG.
It seems very hard to find material related to Rexec through firewall.
Br,
Johan
Thank you for the answer. We have thought of that solution, but the trouble is that the rexec protocol seems to choose
different stderr ports every timy in the range from 1024 and above so we don't know how to solve it without opening
a lot of ports in the forefront TMG.
It seems very hard to find material related to Rexec through firewall.
Br,
Johan
emmmmmm,
you can reduce the risk by opening these ports only to a specific host no to all external.
create a new computer object with the server ip address, and create an access rule from internal to that computer.
If that is not an option for, the application documents should advice which ports are required to make it work correctly.
you can reduce the risk by opening these ports only to a specific host no to all external.
create a new computer object with the server ip address, and create an access rule from internal to that computer.
If that is not an option for, the application documents should advice which ports are required to make it work correctly.
ASKER
Hi again,
We have tried both suggestions, but with no success.
Forefront says Unidentified IP-trafic (TCP:512) Connection Established with result SUCCESS
But after that there is no more information in the logging and still no connection between client and server.
Br,
Johan
We have tried both suggestions, but with no success.
Forefront says Unidentified IP-trafic (TCP:512) Connection Established with result SUCCESS
But after that there is no more information in the logging and still no connection between client and server.
Br,
Johan
I think you are seeing why it is just plain a bad idea to put a firewall between the Clients and their LAN Resources. Firewalls are to protect the LAN from outside networks,...you should not be sticking them in the middle of the private LAN. the fact that the users are a different subnet in this case is a given and the fact that the subnet (segment) was created by VLANing doesn't really matter.
Many applications that are designed to operate within a private LAN use protocols that are too complex to run over a firewall unless the Firewall has an Application Layer Filter to interpret and process the protocol's traffic pattern,...and there is no way you are going to come up with an Application Filter for a custom application such as this.
Many applications that are designed to operate within a private LAN use protocols that are too complex to run over a firewall unless the Firewall has an Application Layer Filter to interpret and process the protocol's traffic pattern,...and there is no way you are going to come up with an Application Filter for a custom application such as this.
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
Not the perfect solution, but it only requires minor change in the clients to make it work.
You can define a new protocol using a primary port and secondary port from the new protocol wizard.
then add this protocol to that access rule.