I looked up Colin Bloch's documentation, and I managed to setup mysql replication option. I'm running two identical freeradius 0.6 with mysql 3.23.38 radius database on two separate FreeBSD 4.5 servers, and our ISP is proxing their radius authentications to our radius servers, so we can have accounting information, and we can bill to our customers. I have some question regarding mysql replication setup since I believe "radacct" table will not be identical if you are running two separate freeradius servers, and different clients may be hitting two servers randomly.
I used "binlog-do-db=radius" option on my mysql master to replicate only radius database, and I also used "replicate-ignore-table=radius.radacct" on my mysql slave serer to ignore replication on radacct table. When master and slave database are not identical, the slave serer replication process usually failed, so I tried to avoid the situation like this by ignoring radacct table. Also, if you want to have a single accounting data to reference for all radius authentications, running multiple freeradius front end on multiple servers to point a single mysql backend server? Does anyone have any tips regarding how to set up mysql replication with freeradius for redundancy in production environment? Regards, Ken - List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
