I have several clients setup with mysql replication, one with 6
servers replicated. I have thought about that, but it gets REALLY
messy. The only way I think it would work, is if you had a schema for
a client, that rarely if ever changed. Also, even though mysql rep is
the best I have seen, and I love it, you still have to do selects on
the master. Especially if you do a select immediately following an
insert or update. Even though I have never seen seconds_behind
anything buy ZERO, I don't trust that the data will be there in that
situation.
Also, using the method of turning off autocommit before a large
number of inserts, will make the performance virtually the same as if
you had no indexes.
--
Robert Garcia
President - BigHead Technology
VP Application Development - eventpix.com
13653 West Park Dr
Magalia, Ca 95954
ph: 530.645.4040 x222 fax: 530.645.4040
[EMAIL PROTECTED] - [EMAIL PROTECTED]
http://bighead.net/ - http://eventpix.com/
On Jul 27, 2007, at 10:24 AM, William M Conlon wrote:
I think another speed up is obtainable if you use database
replication, where inserts and updates are done on one database,
without indexes, and the tables are replicated to a second db used
for SELECTS, which does have indexes.
________________________________________________________________________
TO UNSUBSCRIBE: Go to http://www.witango.com/developer/maillist.taf
