Configuring the voice server

On the system that you want to configure as a voice server:

Log in as root.
Create and mount a new file system to store your voice data.
The size of this file system is determined by the amount of voice data (including voice messages, voice segments, and user greetings) that you expect your single system image to generate.

For information on how to create and mount a file system, see the System Management Guide; Communications and Networks book.

The following instructions assume that you name your new file system /ssi.
Copy the file /home/dirTalk/DIRTALK.vox.tar.Z (which is supplied with Blueworx Voice Response) to the server you are working on.
The following instructions assume that you copy this file to the /tmp directory.
Type the following command:
```
cd /ssi
```
Type the following command:
```
zcat /tmp/DIRTALK.vox.tar.Z | tar -xvf-
```
Type the following command:
```
chown -R dtuser:staff /ssi
```
Type the following NFS command:
```
mknfsexp -d /ssi -c client1, client2,...clientN, database_server -r database_server
```
where client1 through clientN is a comma-separated list of the host names of the client nodes of your single system image, and database_server is the host name of the database server.

This command saves the export definition in the /etc/exports file, so the directory is exported every time the server starts.

Note: The mknfsexp command shown here has the -r flag specified to allow the database server to access files on this partition as a root user. The -r flag is required for saveDT to function correctly. For more information about this command, see the Commands Reference.