|
Dunno if this helps, but here is my setup...
[BGP INTERWEB]
|
+---------------+
| |
[KEEPALIVED1] - [KEEPALIVED2]
| |
[100 Mbit Switch]
|
+--------+--------+--------+
| | | |
[NODE1] [NODE2] [NODE3] [NODE4]
| | | |
+--------+--------+--------+
|
[1000 Mbit Switch]
|
[1.5TB NAS]
LVS:
----
keepalived using LVS-NAT
MTA:
----
Qmail (modified)
VpopMail
Web:
----------
sqWebmail (heavily modified)
qmailAdmin (modified)
vqAdmin (modified)
Anti-Virus:
-----------
clamAV
qscanq
Anti-Spam:
----------
SpamAssassin
RBL (spamhaus, ordb, rfc-ignorant etc..)
SPF (Qmail linked with libspf)
Maildrop
Other:
------
MySQL3 w/ replication
ReiserFS
NFS3
30,000 mailboxes
Each node is identical, in fact, even the LVS nodes could participate in
an emergancy. They are identical in that they can all handle all the
aspects of our mailsystem, but don't necessarily. The director
specifies which nodes will handle what and manipulates this according to
load.
Cheers,
James
--
James Couzens,
Programmer
-----------------------------------------------------------------
http://libspf.org -- ANSI C Sender Policy Framework library
http://libsrs.org -- ANSI C Sender Rewriting Scheme library
-----------------------------------------------------------------
PGP: http://pgp.mit.edu:11371/pks/lookup?op=get&search=0xBD3BF855
|
