I am looking for a creative way to optimize my web application (largely reads) with a transaction firehose on the data input side. Batch vs. OLTP, if you will.
Is there any technical reason why I could not configure the master database without indices and the slave(s) _with_ indices? The intent being to isolate the inbound transactions from the reads _and_ make the database inserts as simple as possible (e.g. few or no indices). Several slaves could be configured to throttle performance on the web side (more reads, move slaves, etc.), which could gain a performance edge by indexing. For disaster recovery of the master, one slave could be "untouched" and not indexed. I'd be interested to hear comments. Thanks in advance, /Tony --------------------------------------------------------------------- Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail <[EMAIL PROTECTED]> To unsubscribe, e-mail <[EMAIL PROTECTED]> Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
