On 10/01/2013 10:59, shthead wrote: > On 10/01/2013 6:43 PM, Vincent Hoffman wrote: >> looking though the log lines such as >> Jan 10 07:17:13 delphi pdns[31017]: 5046 questions waiting for database >> attention. Limit is 5000, respawning >> >> seem to be disturbingly common, any tuning suggestions? this a pretty >> powerful box with reasonably tuned mysql but plenty of spare resources. > > Hello Vince, > > I ran into a similar issue at one stage which turned out to be when > the server was getting backed up. It might be worth enabling the slow > query log (if it isn't already enabled) and checking that to see if > any queries happen to be running for a long time when that happens. > Also make sure if the tables are InnoDB that you have a large enough > buffer pool to fit the entire thing in memory, that gives a nice > boost. Maybe if you want a quick overview of how your server is going > give this Perl script a go: > Good Idea with the slow query log. Its a read only slave of a master DB, sadly not an InnoDB DB only myisam. The server has 16Gigs of RAM and I already have key_buffer_size = 1G I may look into changing the engine to innodb though to allow better caching (but that opens a whole new can of worms ;) > http://mysqltuner.pl > I already use this as part of my mysql tuning. > Does it happen periodically or is it a random occurance? > random as far as I can tell. I'm looking for correlations with network traffic/backups etc but cant see anything at the moment.
Thanks, Vince > Thanks > > _______________________________________________ > Pdns-users mailing list > [email protected] > http://mailman.powerdns.com/mailman/listinfo/pdns-users > _______________________________________________ Pdns-users mailing list [email protected] http://mailman.powerdns.com/mailman/listinfo/pdns-users
