NCP connections and Watchdog under OES Linux

A customer of my complained that after migrating from Novell Netware 6.5 to OES2 Linux, the users complained that sometimes the loose their connection with a specific server.

After a trace (how to do show look here) I saw that if a users doesn’t uses a server for more than a hour his connection gets broken. If a users wants to for example get a file list of the specific server the get a connection broken error. If the click OK and the drive again the file list appears. The trace shows that after the connection broken error the Novell Client does a reset of the TCP connection and does a automatic user authentication to the server.
If you wait more than 2 hours the server sends a Keep-Alive packet to the workstation to see if the workstation is still available.
But why does the connection gets broken after a hour?? This was due a firewall that was between the workstation and server. This firewall (Cisco ASA) resets a TCP connection that isn’t used over a hour. They told me this is default.

The solution was to enable watchdog for OES2. Watchdog is a method to see if a workstation is still available. If you go to Remote Manager | Manage NCP Services | Manage Server you can define the option FIRST_WATCHDOG_PACKET. Default this option is 0, the value you define here is the time in minutes that the servers sends his first WATCHDOG packet to the workstation. After that the server will send every minute a WATCHDOG packet to the workstation to see if it’s still available.
I set mine to 50 minutes.

Here you can find information on setting for OES.