I ramp up the users by 25 users.  Started with 25 and increased 25 
gradually to reach 150 for simultaneous and started with 25 and 
increased  the same 25 to reach 75 for concurrent.

- Shan

At 10:31 AM 7/21/2002 -0400, Dominic Uliano wrote:
>When you observed the high CPU utilization and slow response time, what was
>the ratio of User to System time on the CPU?  Sun typically runs at a 2:1
>ratio under normal load and then will transition to 1:1 and then 1:2 under
>peaks of very high load.  When the switch in the ratio occurs it means that
>the machine is spending more cycles on system resources versus user
>(application) requests.  Depending upon when and what the ratio was it could
>either mean a server bottleneck or application code bottleneck.
>
>   Sounds like you are using Mercury LoadRunner -- did you ramp of the users
>or did you log them all on at the same time?
>
>Regards,
>Dominic Uliano
>
>-----Original Message-----
>From: Shan Gopalakrishnan [mailto:[EMAIL PROTECTED]]
>Sent: Saturday, July 20, 2002 2:03 PM
>To: Jetspeed Users List
>Subject: high CPU utilization
>Importance: High
>
>
>Hello folks
>
>Our team did a quick poc application in jetspeed and did a load test thru'
>load runner.
>Below is the configuration of the Solaris box.   We had just that jetspeed
>application running in that box while
>doing the load test.  We used  Tomcat 4.0.3 as the servlet engine.
>
>System Configuration:  Sun Microsystems  sun4u Sun Enterprise 220R (2 X
>UltraSPARC-II 450MHz)
>System clock frequency: 113 MHz
>Memory size: 2048 Megabytes
>========================= CPUs =========================
>                      Run   Ecache   CPU    CPU
>Brd  CPU   Module   MHz     MB    Impl.   Mask
>---  ---  -------  -----  ------  ------  ----
>   0     0     0      450     4.0   US-II    10.0
>   0     2     2      450     4.0   US-II    10.0
>========================= IO Cards =========================
>       Bus   Freq
>Brd  Type  MHz   Slot        Name                              Model
>---  ----  ----  ----------  --------------------------------
>----------------------
>   0   PCI    33     On-Board  network-SUNW,hme
>   0   PCI    33     On-Board  scsi-glm/disk (block)
>Symbios,53C875
>   0   PCI    33     On-Board  scsi-glm/disk (block)
>Symbios,53C875
>
>
>The results came out to be as follows :
>
>For  80 concurrent users with  ( zero sleep time )  the % CPU utilization
>was  85% in the box where jetspeed was running
>For  150 simultaneous users ( with 30 seconds sleep time ) the % CPU
>utilization was around 78% or so in the box where jetspeed was running.  As
>we spiked the value the performance was going down . The initial response
>time is expected to return within 5 to 6 seconds .
>But during the peak load the response was going down enormously.
>
>We observed the middle-tier and the database back-end was stable and showed
>less than 10% CPU utilization.
>
>We also used  OptimizeIt to track down where the % CPU increases and
>observed some of the jetspeed framework calls
>are pretty high cpu intensive.   Any opinions on this.?
>
>- Shan
>
>
>--
>To unsubscribe, e-mail:
><mailto:[EMAIL PROTECTED]>
>For additional commands, e-mail:
><mailto:[EMAIL PROTECTED]>
>
>--
>To unsubscribe, 
>e-mail:   <mailto:[EMAIL PROTECTED]>
>For additional commands, e-mail: 
><mailto:[EMAIL PROTECTED]>


--
To unsubscribe, e-mail:   <mailto:[EMAIL PROTECTED]>
For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>

Reply via email to