Yes, I would be very interested to see this in more detail. The
alternating auto_increment would solve the unique ID problem. I didnt
think of that, but its pretty slick.
~Rob
Josh Marshall wrote:
Rob Gil wrote:
What kind of options are there for MySQL replication or MySQL
Clustering?
-The obvious problem of writes to slave servers... is there a way to
separate reads and writes?
-Has dbmail been tested on NBD tables?
-Has any work been done with HA mysql applications and dbmail?
I have a dev system about to replace our current mail server system
that uses MySQL replication in Master-Master form. Basically you set
up (for a pair of servers) two slave sessions, each pointing at each
other, and set the autoincrement to use odd values for server1 and
even values for server2. Since we have only two mail servers in this
config I run the mysql servers locally to each machine. If you wanted
to have more than two servers you could either use a MySQL
master-master-master....n ring for the mail servers to connect to
localhost, or have two separate MySQL servers in master-master format
and use LVS to spread the load between the mysql servers (use
keepalived to check the health of the servers and use persistency to
keep the requests on the same machine)
I really don't think you want to use dbmail with MySQL cluster as it
keeps all the data in memory... and dbmail databases are way too big
to keep in memory :)
If anyone wants more details of what I've got setup let me know I'll
try to write something more detailed.
Josh.
_______________________________________________
DBmail mailing list
[email protected]
https://mailman.fastxs.nl/mailman/listinfo/dbmail
_______________________________________________
DBmail mailing list
[email protected]
https://mailman.fastxs.nl/mailman/listinfo/dbmail
