Does the code base for v3 support mysql slaves yet?

There is nothing in IPB specific to this. Generally a master/slave setup is handled at the mysql end. You feed IPB an IP address, and when it calls out to that IP to establish the MySQL connection, that server would determine which slave to pass the request to.

Not if you just have simple replication enabled... I think what the OP is requesting is the ability to set one mysql server to write to, and one (or more) mysql servers to read from. :)

It does not have read/write separation at this time but it's something we have experimented with and may make it into a release of IPB in the future.

[quote name='Dan C' date='22 September 2009 - 04:31 PM' timestamp='1253651481' post='1860373']
Not if you just have simple replication enabled... I think what the OP is requesting is the ability to set one mysql server to write to, and one (or more) mysql servers to read from. :)


Yes, slightly different. From my memory (which is spotty, so forgive me if I get something wrong) you can approach this in a couple different manners. The two most common are

Master/slave setup - mostly handled at the mysql end. Application just calls to mysql, which then determines which server to pawn the request off to.

Replication - application has to specifically separate select and update/insert/delete/etc queries to ensure that the "parent" is updated properly.


As Charles said, we've experimented with that second approach, but haven't done large-scale tests, and given the nature of such functionality, it's not something we can just roll out without good testing first. :P