Duke University
Department of
Computer Science

Setting up a Jumpstart Server

The Jumpstart Server performs three separate functions, which can be peformed by a single machine or can be spread out across serveral machines, depending on your network.

  • Boot Server

    • Uses RARP and bootp to set the basic network parameters for the machine. Can also use DHCP.

    • Uses tftp to load a boot kernel to perform the more complex task of mounting the appropriate directories used to install the Solaris packages

    • Boot server must exist on same network as client; once client has loaded it's boot kernel, it can access an Install server across routers

  • Install Server

    • Contains Solaris packages, copied from the Solaris installation CD's, to be installed.

    • Contains a Solaris miniroot the client mounts via NFS. The OS install is performed while running from this miniroot.

    • Multiple Install servers can be used to distribute the load.

  • Configuration Server

    • Contains site-specific information used for a custom jumpstart installation

    • sysidcfg file used to set basic network configuration; this is needed to perform an unattended install. A different sysidcfgfile is needed for each architecture and OS release.

    • Typically only have a single Configuration server for administrative ease. Load is rarely a factor.

[Previous| Next]

   © 2001, Duke University Computer Science Department. All rights reserved.
   Comments to braun@cs.duke.edu