On Sunday, August 11, 2002, at 12:01 AM, Suresh Ramasubramanian wrote:
> george@??? (George Schlossnagle) [Sunday, August 11, 2002 11:28
> AM]:
>
>> or put the spool disks on a san, or have all spool entries written to
>> two physical places.
>
> Would that massive duplication be necessary?
I guess if you're really worried about hosts dying (whihc netsaint would
tell you about but do little to prevent), you'ld like to make sure you
can recover your first tier spools even if a host 'burns to the
ground'. Duplicating it definitely seems expensive (from a cpu overhead
level), but is cheaper for some installations that are bigger than a
breadbox but smaller than outblaze.
That having been said, having small fcal arrays shared by a couple
relays (not overlapping spools, just co-mounted) seems like a nice and
relatively cost efficient way to ensure spool high-availabilty through
first tier mx failure.
For the maildir access, I completely concur, but I want to make sure the
mail a relay receives to be delivered (less a concern for inbound mail,
more a concern for outbound).
>
> Why not this -
>
> MX (or MXs, with a load balancer)
>
> All MXs forward to a second cluster of hosts, which then deliver to
> maildirs
> hosted on a SAN.
>
> A pop proxy solution like perdition would then proxy pop3 requests from
> the
> appropriate box. Or the pop box would have access to the SAN as
> well ...
>
> -srs
>
>
>
// George Schlossnagle
// Principal Consultant
// OmniTI, Inc http://www.omniti.com
// (c) 240.460.5234 (e) george@???
// 1024D/1100A5A0 1370 F70A 9365 96C9 2F5E 56C2 B2B9 262F 1100 A5A0
