I am planing to design high-available mysql servers which serve as one
single virtual mysql server to application servers. My criteria is:
1. HA(High Availability)
Due to mission critical services, we hope the mysql clusters can provide
high availability to application servers. In the hardware term,
I think hardware raid or software raid is necessary, but which is
recommended in mission critical case? In the application/software term,
I hope to have load-balancing among mysql servers. Also, I'd like to
have fail-over design in it.
Now we know there's one-way replication in mysql-3.23(ok, maybe two-way
replica. is also possible), how reliable is the built-in replication
function in mysql-3.23? Can I rely on hardware syncronization solutions
like storage appliance solutions?
The service based on it(mysql) is not quite mission critical, but we
just do not want to lose the high availability. I ask here just want to
know how possible can I build a high-available mysql clusters with
Any ideas are hightly welcome.
Patrick Hsieh <pahud@stripped>
GPG public key http://pahud.net/pubkeys/pahudatpahud.gpg
|• High-Available MySQL Servers topology||Patrick Hsieh||27 Jan|