Problem With NIS Slave Server

[Andrew Kaplan]

Hi there -- I went through the motions of adding an additional slave NIS server to an existing NIS domain. The master server is running HP-UX 11.00 and the new slave server is running Fedora Core 2. There is already an existing slave server running Fedora Core 2.

As a test to the slave server's functionality, I configured one of our NIS clients, also running Fedora Core 2, to use the slave server. When I rebooted the client, it appeared to hang when it came time for it to "looking for NIS server..." After I gained control of the client, I reconfigure it to broadcast a request rather than bind to a specific server. After I rebooted the client again, it was able to bind to the NIS domain without a problem. I ran ypwhich and saw that the client bound to the original slave server.

One question that I have is what should the hostname of the slave server be. Specifically, should it be the hostname or the FQDN? If that does not factor into the equation, what configuration settings should I look to change to enable the second slave server be available to the NIS clients?

[Steven E. Protter]

NIS should not need th FQDN, it should work on short names.

I would do this:

tail -f /var/log/messages on the master server.

Then boot the client. Look for errors and issues there.

SEP

[Andrew Kaplan]

I checked the master NIS server for any error messages referencing the new slave server, but did not find any. I then checked the new slave server and found the following lines in the /var/log/messages file:

localhost ypserv[29124]: WARNING: no securenets file found!
localhost ypserv[29124]: Support for SLP (line 20) is not compiled in.

I also checked the NIS client's log file when it was configured to bind to the new slave server and found that while it could connect to the NIS domain, it could not bind to the slave server. This resulted in ypbind not running on the client.

Any other ideas?