[44] in Info-AFS_Redistribution
Re: Setting up AFS -- beware
daemon@ATHENA.MIT.EDU (daemon@ATHENA.MIT.EDU)
Fri Jan 11 19:35:07 1991
From: bob@ibmpa.awdpa.ibm.com (Bob Andrews)
To: kriso@northstar.dartmouth.edu (Kris Olander)
Date: Fri, 11 Jan 91 11:08:48 PST
Cc: info-afs@transarc.com, bernerus@cs.chalmers.se
In-Reply-To: <QbXRb=K1eVXp0FhldT@northstar.dartmouth.edu>; from "Kris Olander" at Jan 11, 91 10:29 am
Kris Olander writes:
>
> Has anyone thought of a reasonable method for getting a find database
> built for their
> cell?
>
Well, whatever you do, you wouldn't want it to run on every client...
that would be too much load on the net and the servers. I run a
single find process, saving the data in a bunch of 1000 line files
in AFS (AFS doesn't perform well with large files). You can then
do a grep on those files to get fast-find functionality. Something
that I found useful was to replace all the components of the path
of a filename that were the same as the previous filename in the
files with a tab. This save a lot of space. The data has to be
reconstructed of course to look at it...but that's not hard...
Bob
