From: Michael Widenius Date: June 20 2002 11:21pm Subject: POOLED experiment results List-Archive: http://lists.mysql.com/internals/3912 Message-Id: <15634.25493.520496.258242@narttu.mysql.fi> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hi! >>>>> "roberto" == roberto writes: roberto> Hi to all, roberto> I have builded a beta object for mysql pooling. roberto> The results (with 1000 queries serialized) are still very good: roberto> without pooling a simple SELECT and retrieving data need of around 16 msecs: roberto> ~63 queries for second (only) roberto> with pooling a simple SELECT and retrieving data (same previous routine) roberto> need of 1.17 msecs: ~850 queries for second roberto> (on a celeron2 1000 Mhz).Consider that the profiler isn't real timing...more roberto> probably the 850 qxsecond can be higher without debugging code. roberto> www.dellapasqua.com/mysql_nopool.gif roberto> www.dellapasqua.com/mysql_pool.gif roberto> I have set the timeout on 86400 seconds (1 day), Which timeout variable is this ? roberto> does Mysql accept this high number as timeout ??? Most timeout is in seconds and stored in an integer; In this case this timeout should be ok. roberto> Please if someone know the my_sql_option options, can suggest me if this roberto> setting can be right? If this is the 'connect_timeout' variable, to mysql_options() this only affects the initial connection to the server, not how long the connections stays up (which is the server 'wait_timeout' variable). The wait_timeout value can be set to several years, if you so desire... Regards, Monty