On 30/06/2010, jz <[email protected]> wrote: > Thanks > > Yes i am also thinking on same lines. It must be connection opening > overhead. e.g. if i do this > > no. of threads: 10 > no. of connections: 10
You should use "Max Number of Connections" = 0 otherwise connections are pooled between threads, which would not happen with independent users in real life. > ramp-up: 0 sec > > So these threads will open all connections. Now when i re-run the above > test, will the physical connections be opened again or taken from pool of > previous run? Connections are not kept between runs. > I will also share, results of your idea. > > Regards, > Jehanzeb Qayyum > > > > > > On Wed, Jun 30, 2010 at 10:52 AM, Sudip Kumar Bhattacharya < > [email protected]> wrote: > > > Can it be related to the overhead of creating the extra connection objects > > and getting the connection to the db? > > You can take the connection creation overhead by running your query in a > > loop of say hundred iteration. > > Then if you average out by taking total time /total iterations, you may get > > the expected output. > > Sent on my BlackBerry® from Vodafone > > > > -----Original Message----- > > From: jz <[email protected]> > > Date: Wed, 30 Jun 2010 10:42:03 > > To: <[email protected]> > > Reply-To: "JMeter Users List" <[email protected]> > > Subject: JMetere JDBC Test Confusing Result > > > > Hi, > > > > I am running JMeter JDBC test. I am getting confusing results which vary > > with no. of threads and no. of connection in pool > > > > Query: select sysdate from dual > > > > Test Case 1: > > no. of threads: 10 > > no. of connetcions: 1 > > > > Avg Sample Time is 5 ms > > > > Now i increase no. of connections to 2, thinking that it would decrease > > sample times. But now avg. sample time moves to 200 ms. I have noticed that > > increasing no. of connections while keeping no. of threads constant > > increases sample times. Please can someone explain this phenomenon? > > > > Thanks > > > > Regards, > > Jehanzeb Qayyum > > > > > --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
