Reverse Invoke on the Windows 2000 Platform

Is anyone running Reverse Invoke on the Windows 2000 Platform?

If so, do you have any issues, or does it appear to be stable?

We use RI on Windows 2000, 2 RI servers in the DMZ, each connected to a WM server inside our firewall. Loading is fairly light, as most partners use FTP or SMTP for integrations.

We had some initial problems with RI connections failing, but have resolved by scheduling

(1), running every minute.

(2) a custom service running every 15 minutes to detect and restore failed connections, using services,,


Can you give me more information on how you did #2? That sounds like something we might need!


We are running RI on Win2K for quite some time now with no real issues apart for the (2) that Martin describes. This custom service is scheduled to run every 15 minutes on our internal IS to check the connections and restore any failed connections, if needed.

Saurabh Maheshwari had uploaded a package on this site that shows how one would write the services to detect and restore RI connections. Please read the following thread for more details:




It is no longer needed to run a custom service to recreate dropped connections to the RI Server.

In the 6.0.1 IS, webMethods has added the code to recreate the connections to which is scheduled automatically.

We’re still at 4.6, so it looks like we need the custom service.

I recall that someone uploaded a custom package in the shareware area that handles this keepAlive issue.

We have that package and are working on customizing it. Thanks!

Another question, does anyone have RI implemented WITHOUT any customized code to re-register the connections? Does it work without it? Is it normal to have such frequent connection problems?

For what it’s worth…We are using reverse invoke in a win2000 environment with IS 4.6. We are using but do not have any other services scheduled. We use the reverse invoke connections for EDI and process 30-50 transactions a day over it. The only time we lost our registered connections it was because the firewall detected an anomoly that caused a switchover to an alternate firewall which would not support the reverse invoke connection. The firewall team worked some magic in the alternate firewall which allowed these connections to remain established, and we haven’t had a problem since…


Jon has a point. IMHO, the need to have this code to re-register connection to the proxy server is dependent on the firewall separating the two servers. Our firewall has been notorious for dropping unused connections after a given time and instead of approaching them, we chose to “fix” it ourselves.



Thanks for the responses!

Let me also add that we are running 4.6.

Is RI more reliable in version 6.01?

we receive FTPed files from external systems to our DMZ server, to move the files from DMZ server to the internal webM IS server we want to use FTP?

Did anyone have issues in the above approach previously?


You can successfully FTP between these two wm IS servers provided your DMZ environment’s security policy allows you to open an FTP hole (port)in the inner firewall.


Hi All,

What is best approach(http,https,ftp,other) to get the files from DMZ server to internal webMethods server? We need have the scheduler on the internal server to pick the files from DMZ and write on to the internal webMethods server disk?

We have different file types, and max file size upto 8mb !!