Your network can have one or more NIS slave servers. Having slave servers ensures the continuity of NIS services in the event of the master server being down. Before actually running ypinit(1Mnis) to create the slave servers, take several precautions.
First, as you did with the master NIS server, you must check that every slave server's password database contains an entry for the daemon user name and that it precedes other entries with the same user ID.
Also, make sure that the network is working properly before you set up a slave NIS server. In particular, check that you can use rcp(1tcp) to send files from the master NIS server to NIS slaves. If you cannot, follow the procedures outlined in ``Administering Remote Procedure Calls (RPC)'' to permit the use of rcp.
Now you are ready to create a new slave server. Perform the following procedure:
ypinit prompts for the domain name for the slave server.
ypinit prompts for the name of the servers for the domain.
where master is the host name of the existing NIS master server.
Ideally, the named host is the master server, but it can be any host with a stable set of NIS maps, such as another slave server.
ypinit will not prompt you for a list of other servers as it does when you create the master server, nor will it run ypbuild again. However, it will stop executing if you have not used ypinit -c to initialize the list of servers, and it lets you choose whether or not to halt at the first non-fatal error. ypinit then calls the program ypxfr(1Mnis), which transfers a copy of the master's NIS map set to the slave server's /var/yp/domainname directory.
+auto.directThus, whenever the automounter reads this file, it will consult the NIS auto.direct map upon reaching this line.
This step ensures that processes on the slave server also use the NIS services, rather than files in the local /etc directory. In this way, you ensure that the NIS slave server is also an NIS client.