We have a box that is primarily read at a rate high enough that a particular table has a SELECT outstanding about 99.9% of time 24x7. This box is set up as a replication slave, and the table in question is replicated from another box. It appears as though the slave thread doing the writes has a lower priority than threads doing reads. I see the slave thread in the process list blocked for long periods of time (hours), waiting for access to this one busy table.
I can't see any config option (except perhaps --skip-thread-priority) that would alleviate this problem. I'm not even certain this is the actual problem at this point, since the docs are unclear about the relative priority of the slave thread. Empirically, it appears to be lower, but I'd like some confirmation if anyone can offer any. It certainly makes sense that the slave thread would be lower priority than any other thread, except in cases like this, where it would be better for the slave thread to have the same priority as non-slave writes (i.e., higher than reads). Is there some option I'm overlooking? james montebello --------------------------------------------------------------------- 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
