So we've been having a look at the best way to deploy LUM in a tree that is structured based on the geographical distribution of sites. The tree is essentially made up of a bunch of OU's that represent remote site locations, in which are contained users, groups, local servers, desktops etc

I'm wondering what the best way to deploy LUM in this situation is, given that I think (at this stage) that we will only deploy LUM for use by server administrators (i.e. SSH, sudo etc), not end users.

My main question is about where to locate the UNIX (linux) workstation objects in the tree. My initial assumption is with the NCP server object, but I wonder why Novell allow to you set this to where ever you like. Are there advantages/disadvantages of grouping all of the UNIX workstation objects together somewhere in an OU as opposed to just creating them in the same OU as the host NCP server?

I'm guessing tree walking is not particularly relevant given the NAM cache daemon and that fact that the server has to query for the UNIX config object which is likely to be located somewhere near the top of the tree anyway.

Anyway, just wondering if anyone out there had any thoughts on this (or even better, some real world experience either way).