Saturday, August 13, 2005
Tech issue: MySQL clusters versus replication
I'm in the process of designing the server farm, and am looking for MySQL High Availability experience. Traditionally, I have run the database engine on a big, mostly-hot-swappable box (read: expensive) and backed everything up periodically. This is fine when your application's usage periods allow slow / downtime (e.g. most SAP installations -- companies aren't doing a lot of transactional business at 4AM on a Sunday...)
I'm debating using MySQL Clustering this time around. The primary advantage I see is it's built for high availability. The disadvantages are two: 1:memory requirements for the data store side (2x database size plus overhead spread across the total number of data store machines); and 2: no support of blobs yet (at least as of the 4.x series, I haven't seen if they've added blobs to v5) -- update -- I see blobs are now supported -- at least, out on the development branch... RTFM....
Anyone out there running MySQL Clustering who would be willing to chat for a few minutes? Email me at my firstname at gatheroo.com or just reply here.
I'm debating using MySQL Clustering this time around. The primary advantage I see is it's built for high availability. The disadvantages are two: 1:memory requirements for the data store side (2x database size plus overhead spread across the total number of data store machines); and 2: no support of blobs yet (at least as of the 4.x series, I haven't seen if they've added blobs to v5) -- update -- I see blobs are now supported -- at least, out on the development branch... RTFM....
Anyone out there running MySQL Clustering who would be willing to chat for a few minutes? Email me at my firstname at gatheroo.com or just reply here.
Labels: old
Posted by Mike Milkovich at 9:37 AM
Post a Comment