Configuring Diskless Clients

Each diskless client requires an entry in the Host Manager. Use NIS+ to centrally administer the diskless clients.

Add Diskless Clients

1. On the workstation that is going to be the OS server, log on as a user who can assume the admin role.

2. As role admin at label admin_low, open the Host Manager with the NIS+ Naming Service.

1. Add each diskless client as an entry in the Host Manager.

   If the client exists already, delete it and re-create it. A diskfull client cannot be converted to diskless.

   Table 10-3 Diskless Client Information in Host Manager

   Entry	Value
   Host Name
   IP Address
   Ethernet Address
   System Type	Diskless
   Timezone Region
   Timezone
   File Server	(OS server is already entered for you.)
   OS Release	Select the platform for the client.
   Root Path	/export/root
   Swap Path	/export/swap
   Swap Size	> 64 MB

2. Save the changes.

   Files for the client will be created in /export/root/clientname. Adding a diskless client takes from 15 to 30 minutes per client.

Ensure that the Client is Known to the NIS+ Master

1. Log in to the NIS+ master as a user who can assume the role root and assume it.

2. As root, at label admin_low, make sure that the client information in the kernel cache and the tnrhdb table is correct.

   1. Launch a terminal.

   2. Look for the client's IP address or a fallback address in the kernel cache.

      # tninfo -h

   3. Check that the information is in the tnrhdb NIS+ table.

      # niscat tnrhdb.org_dir | more

3. If the client is in the tnrhdb file correctly, but is not in the kernel cache, update the kernel.

   # cd /etc/security/tsol # tnctl -T tnrhtp # tnctl -H tnrhdb

   1. Then check the kernel cache and run the command nistntime.

      # tninfo -h # /usr/lib/nis/nistntime tnrhtp # /usr/lib/nis/nistntime tnrhdb

4. If the client is not in the tnrhdb file correctly, open the Database Manager with the NIS+ naming service, choose tnrhdb, and enter the client or the fallback mechanism for the client's subnet.

   When you exit the Database Manager, the tnrhdb and the kernel cache are updated.

Set up Each Client's Mounts

1. On the OS server, as root at label admin_low, open the Admin Editor from the System_Admin folder, with the file /export/root/clientname/etc/vfstab.

   You will do this once per client.

2. Create an /opt entry in the vfstab file.

   The /opt mount point enables the client to run Solstice AdminSuite. You can add other mount points as well.

   For example,

   <server>:/export/opt - /export/opt nfs - yes     bg,intr,soft
   squirrel:/export/tools - /export/tools nfs - yes   bg,intr,soft

3. Write the file and exit the editor.

4. As root, at label admin_low, create the mount points in the client's root directory.

   # cd /export/root/clientname # mkdir -p export/opt # mkdir -p export/tools

Verify Each Client's tnrhdb Entries

1. On the OS server, as root at label admin_low, open the Admin Editor from the System_Admin folder, with the file /export/root/clientname/etc/security/tsol/tnrhdb.

   You will do this once per client.

2. Correct any entries in the file that are not in the following format:

   ip_address:template
   nnn.nnn.nnn.nnn:template
   

   For example, the following is a correctly formatted sample entry:

   129.150.129.7:tsol

Boot a Diskless Client

When booting for the first time, provide the client with a root password.

1. At the ok prompt, type boot net.

2. When booting for the first time, provide and confirm a root password.

   Result: The diskless client is ready for use by a normal user.

   See Trusted Solaris Administrator's Procedures for the procedure to remove a diskless client.