Coda File System

Re: *****SPAM***** Newbie Question - hoarding files

From: Jan Harkes <jaharkes_at_cs.cmu.edu>
Date: Tue, 4 Feb 2003 14:33:25 -0500
On Tue, Feb 04, 2003 at 04:10:38PM +0000, paul.rogers_at_mowlem.com wrote:
> SPAM: -------------------- Start SpamAssassin results ----------------------

Hmm, I wonder where my filter broke. It is supposed to forward
suspicious mails to the mailing list only to my address and not the world.

> I have a couple of questions relating to hoarding files.
> 
> 1.  Having created a hoard file using one of the techniques in the
> manual, where is this hoard file actually situated, in order to make
> it work?  Is this configured somewhere?

It is just a file, mine is in /coda/usr/jaharkes/hoardlist, that way I
can get at it from any client.

> 2.  Once hoarding begins, where do the hoarded files reside?  I\'d
> like to mount a drive to give me enough room to store all the hoarded
> files on the server.  What should be the mount point for the drive?
> Is it the coda root on the clients?  If so how is this handled when
> the server is a available ie is mounted on the coda root?  If not how
> does the mapping occur when the server becomes unavailable?

Ehh, store hoarded files on the server? mount point? I don't think you
understood hoarding.

Think of the hoard daemon as a really dumb user who once every 10
minutes uses your client to access all the files in the currently loaded
hoard profile. The end result is that if the file was updated on the
server, the latest version is pulled into the Coda client cache.

If there was no hoard daemon, cached files might be old versions,
because the server callbacks only tell the client to invalidate the
local copy. The next access will fetch the updated version, but that
won't work if the client has no network. So the hoard daemon is used to
pro-actively refetch files.

If the user knows he will disconnect, he can also run 'hoard walk' to
trigger immediate hoarding.

Jan
Received on 2003-02-04 14:40:04