You are talking about pure failover only right?
I was talking about load balancing for mysql which is pretty much undoable.
Pure failover might be possible, just make sure the slave becomes the
master on failover, and when the old master get back up it starts being
a slave, and replicates from the new master.
Francois JEANMOUGIN wrote:
Jan Klopper :
Sent: Wednesday, March 02, 2005 11:56 AM
To: LinuxVirtualServer.org users mailing list.
Subject: Re: Mon + LVS for a mysql cluster
True,
But as soon as you start doing updates to the database you will horribly
screw up your data, because mysql can't propogate the changes made on
slaves back tot the master and so on.
Are we talking about the same thing? I'm not talking about master/slave or
active-slave method using old MySQL replication process. I'm talking about
NBD cluster:
http://dev.mysql.com/tech-resources/articles/mysql-cluster-for-two-servers.ht
ml
http://dev.mysql.com/doc/mysql/en/mysql-cluster-basics.html
I'm currently evaluating this solution, which as passed the "case study"
test. I will implement it at the end of this month. I would like to know if
anyone here have any (bad or good) experience with such MySQL setup.
Thank you,
François.
P.S.: Is it or not out of topic?
P.P.S.: Coda technologies have already been discussed a lot here, but I is
clearly not available for production.
|