Big-board here, with the annoying questions nobody wants to answer.
I read all about the 3.5 changes on vb.com, but I didn't see anything about the query_write query_read abstraction until just now. Color me floored!
My question is, for us large sites that use MySQL replication functionality, is there anything in place currently to distribute reads among a group of servers which are replicating? I'd love to be able to easily spread reads to our slave server(s) for queries, and have the ability to pseudo load balance queries, or at least have some type of "failover", that is, if our slave server is unresponsive to remove it from the pool of available query_read capable servers.
Obviously this would be a very rudimentary form of clustering, but I'll take what I can get at this point (It's more MySQL's fault than anything else, not having master<->master replication sucks).
|