How does a Plume workstation boot?
This explains the boot process on the Plume workstation.
The workstation starts up, then looks for a medium to boot from. It finds the bootrom, or the floppy disk with an Etherboot bootrom image.
It tries to get its network configuration from a BOOTP or DHCP server (probably the Plume server, if you used the simplest, straight-forward setup). It tries to download its kernel image (the one you should have compiled yourself with network boot support) using TFTP, from the location given from the BOOTP or DHCP server. Again, if you used the easiest setup, it would be the Plume server.
The kernel boots and you see a lot of messages scrolling on the screen (we should really add a nice graphical screen to hide these), it initializes all parts of your computer, including the network card (yes, for its support under GNU/Linux).
The kernel asks again the BOOTP or DHCP server for its network configuration, just in case the BOOTP/DHCP server had more information to give than Etherboot could handle.
It mounts read-only the NFS exported filesystem as root filesystem and /dev as Devfs, finds its init program and starts it. init, in turn, does a lot of things [1]. It starts a few daemons and programs each at a time, until it starts the X server in "XDMCP query mode" and using the XFS font server, and you get a graphical login prompt.
[1] look for whatever UNIX manual, more specifically for System V