From: Rick James Date: March 29 2012 5:20pm Subject: Re: mysqld got signal 6 (problem why bigger than I initially realised) List-Archive: http://lists.mysql.com/mysql/227073 Message-Id: <4F7499C5.9070904@yahoo-inc.com> MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit I'll disagree with your disagreement. Many web servers maintain persistent connections. This leads to eventually exceeding max_connections, even if connections are rare. Once he hits 41, he will be issuing 404s, or other bad stuff. Also, having 1000 http threads hanging around is a waste when you rarely have 10 active at once. Or do your threads somehow take a long time without taking much in resources (CPU, I/O, MySQL, etc)? If each of 1000 threads took only 1% of the CPU, the CPU would be overloaded, even on an 8-core box. On 3/27/12 10:38 PM, Reindl Harald wrote: > > Am 27.03.2012 23:24, schrieb Rick James: >> MaxClients 1000 >> could overwhelm max_connections = 41 . Strongly recommend you decrease >> MaxClients to less than max_connections >> (not the other way around). > totally wrong! > > not every http-connection implies a myql-connection > a website typically contains a lot of static files > like images, javascripts, stylesheets and you bring > down your httpd if MaxClients is set too low and > you are unable to serve waiting clients > > additionally if KeepAlive is used there are open > httpd-connections with NO db-connection after > serving a dynamic page using one out of MaxClients > for a timewindow based on KeepAliveTimeout > > MaxClients on httpd should never be lower than "max_connections" > "max_connections" is primary a question of RAM and typical load > > mysqltuner.pl: > [--] Total buffers: 2.4G global + 3.2M per thread (200 max threads) > [OK] Maximum possible memory usage: 3.0G (37% of installed RAM) > _____________________ > > however - 1000 is much too high if we are speak > about a preforked server on most hardware - keep > in mind that a worker process with mod_php consumes > some MB of memory in a typical environment > > but this is all not mysql related > -- Rick James - MySQL Geek