With replication working the next step to gaining a very high
performance system is to have an effective method to divide the queryies
to the many replicated servers. Lets say I have 4 slaves replicating
off of 1 master and 3 webservers that I want to use to pull queries on
the database cluster. Is there any existing way to make those 4 slaves
look like one very high performance database server (by sending
alternating queries to different servers)? The linux virtual server
project can do this for other tcp connections (ie. http, ftp, ...). But
will it work for MySQL? Has anyone done it? It sounds like using the
direct routing method it should work great. But I haven't done any
thing like this before so I am only guessing. If this won't work will
any of the hardware load balancing solutions like cisco's localdirector
work?
Eric
Eric Thelin erict@stripped
AZtechBiz.com: Where Arizona Does Tech Business
Voice: 480-377-6743 Fax: 480-377-6755
