This is a bit of a cop-out, but it works cleanly almost every time: If bringing your system down is not a problem, it would probably be easier (read: less time consuming) to reboot off the emergency disk (Debian 1.2) or the Boot disk (Debian 1.1) and go through the setup again. Instead of repartitioning and re-initializing, mount your previously initialized swap and root partitions, and configure the drivers again. Under the driver configuration menu, under file system drivers, is the ncpfs module. Select it for installation and add it to the kernel.
One caveat on using the ncpfs/ipx modules (and no disrespect to the authors; they have done a great job of working around the non-disclosure limitation from Novell): they are not full implementations, and cannot sense disconnected sessions. If your Novell server ends or disconnects the logged-in account used to mount the volumes in mid-write/read, the ncpfs module will hang and not be able to sync the mounted volumes. I've been able to kill the process conducting the write/read by using the kill -9 option if I catch it early enough, but leaving it unattended may result in a memory-space full of uncleared zombie processes, eventually leading to a failure of the entire system, cold reboot and filesystem repair. I have tried to limit this by keeping the Novell sessions short and handling autonomous (crontab) processes through a central control script. If anyone has an alternative way of mounting Novell servers (nfs maybe), I would love to hear it. Rgds, BK
