On Mon, Apr 26, 2004 at 01:27:08PM -0700, MySQL wrote: > > On Mon, 26 Apr 2004, Jeremy Zawodny wrote: > > > > > Yeah, we've been seeing this bug a bit too. I'm trying to isolate it > > > > and figure out if it's a FreeBSD related problem or something more > > > > MySQL specific. > > > > > > > > Are you using FreeBSD's native threads or LinuxThreads? > > > > > > FreeBSD native threads. > > > > Intesting. > > > > Are you using wildcard hosts in you grants, like "%.example.com"? > > Yes.
Okay. I have a hunch that it's somehow corrupting entries in the acl_cache. I'm working to see if I can prove this, since it's a bit hard to reproduce on demand. > > > Is LinuxThreads becoming stable enough to become production worthy? > > > > Heck yes. That's all we've used for the last 1.5 - 2 year at Yahoo. > > What would you say is the greatest performance gain, or most positive > result of this? Two things: 1. I/O is much faster. 2. MySQL can use all CPUs in a SMP box It's been a really big win for us. > The machine this database is on has 15k rpm SCSI drives which I've seen > spike up to unreal usage (60MB/s), which means that we can't really > improve much on performance without splitting the database onto more > hardware. Just curious why you are placing all your bets on LinuxThreads. We had horrible performance on native threads, instability, and couldn't take advantage of SMP boxes. If you've ever looked into FreeBSD 4.x's poor excuse for "threads" you'd see why. It's all async calls instead of real threading. :-( Jeremy -- Jeremy D. Zawodny | Perl, Web, MySQL, Linux Magazine, Yahoo! <[EMAIL PROTECTED]> | http://jeremy.zawodny.com/ [book] High Performance MySQL -- http://highperformancemysql.com/ -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED]
