On Mon, Oct 20, 2003 at 07:03:37PM +0200, Benjamin KRIEF wrote: > hi guys, > > i've just finished implementing replication on 2 production servers.
[snip] > the problem is that it's quite impossible to fine tune the load between > the servers , and my slave is a bit too busy , so the replication thread > is becoming very slow , and sometimes , i can see that the slave is more > than 2 hours behind the master in terms of replication , and this really > is a pb for me. Two hours?!? Wow. > i looked everywhere for a "high-priority-updates" option, but i can't find > it.tell me if i'm wrong but i deduced that i can't ask the slave to wait > until it reached the master before answering selects. Is your bottleneck CPU, I/O, or lock contention? Is your slave running on worse hardware than your master? > so i've started using more regexps to redirect the queries , but > it's quite awful to my eyes. i would really like to input the ratio > of "select" queries i'd like to send on the slave . Ouch. That's clearly not a good solution. > what about , instead of round robin dns , writing a small perl script , > including a socket listening on the port 3306 , this small script would > have to redirect the sql requests on the slave or on the master , but this > time , i could fine tune it , and input for instance 0,3 for the slave , > so that 3 requests out of 10 would go to the slave. Have you looked at SQL Relay? -- Jeremy D. Zawodny | Perl, Web, MySQL, Linux Magazine, Yahoo! <[EMAIL PROTECTED]> | http://jeremy.zawodny.com/ MySQL 4.0.15-Yahoo-SMP: up 39 days, processed 1,493,548,297 queries (439/sec. avg) -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED]