On 15 Jan 2000, Russ Allbery wrote:
RA>Russell Nelson <[EMAIL PROTECTED]> writes:
RA>
RA>> Because every such installation I've ever seen has used NFS. I'm not
RA>> talking about what's good, or what's right. I'm talking about what's
RA>> possible to do tomorrow. Yes, it might be that a specialized mail
RA>> transfer protocol would work better; convince Netapp to support it.
RA>
RA>Um, how do you think we're scaling our mail system right now? (And we
RA>don't need Netapps to do it.)
RA>
RA>If it's a matter of making it work tomorrow, I can do it faster with a
RA>farm of lighter-weight servers than with anything based on NFS. NFS is
RA>the complicated and expensive solution. If people are currently all doing
RA>it that way, it's either because they don't know better and are too used
RA>to throwing NFS at a problem or it's because they're using the *other*
RA>features of Netapps (snapshots, reliability, etc.) and therefore are stuck
RA>having to use NFS whether they like it or not.
To an extent, but keep in mind that the maildir/nfs solution is _simple_.
Now, you can do things to make it more "robust" (read: complex) to add
functionality, but if you want "simple and scalable", maildir and nfs is a
clear winner.
RA>Give me a bunch of machines with DiskSuite and a couple of internal disks
RA>each any day.
Ack. Veritas :-).
(nothing wrong with disksuite until you start into 0+1 and the like...,
IMHO)
--
--Matt Schnierle
--mgs at stargate dot net
--Stargate Industries, LLC
--#include <std/disclaimer.h>
--"It's not that simple."
