Chris, This sounds *very* much like a locking (enqueue) issue. Was the application changed in any way (trigger, packaged code, etc.) to use a table row for generating sequence numbers? I have also seen such problems when the key_value is generated as a select max(key_value) from table and everything queues up behind this.... Look at the output of the following query: Select event, count(*) from v$session_wait Group by event If you see 'enqueue' in this output with a count of 1 (or more), then this theory is correct. You can then look at V$LOCK, V$SESSION, V$SQL and DBA_OBJECTS to find out further information.... (and the list can help you here with this script). Hth, John Kanagaraj <>< DB Soft Inc Phone: 408-970-7002 (W) Listen to great, commercial-free christian music 24x7x365 at http://www.klove.com ** The opinions and facts contained in this message are entirely mine and do not reflect those of my employer or customers ** >-----Original Message----- >From: oracle-l-bounce@xxxxxxxxxxxxx >[mailto:oracle-l-bounce@xxxxxxxxxxxxx] On Behalf Of Chris Stephens >Sent: Friday, February 13, 2004 7:34 AM >To: oracle-l@xxxxxxxxxxxxx >Subject: shot in the dark applications performance problem.... > > >Let me preface all of this by stating that I realize there is >very little >chance of a resolution to this from the explanation I'm about >to give...but >it thought I'd try and describe the problem here and see if it >sparks any >suggestions.... > > >The company I work for has outsourced an applications implementation. >Everything has been running fine. Currently we run 2 call >center apps out >of the thing. This past weekend the company changed our >architecture from a >1-tier system to a 2-tier system. (not sure if that's standard apps >architectures or not) When the reps began taking calls after >that change, >performance slowed to a halt. They have now been taking phone calls on >paper for 5 days now. The company has reverted back to the 1-tier >architecture but the performance problem still exists. The >people testing >the system describe the problem as follows: > >== > >...one person running the call flows=>performance is fine. > > > >...several persons running the call flows=>EXACTLY 1 minute >delays between >screens. > > > >...several more persons running the call flows=>EXACTLY 2 minute delays >between screens. > > > >...several more persons running the call flows=>EXACTLY 3 minute delays >between screens. (weird) > > > >== > >apparently there is a script that can be run that just goes through the >screens without running ANY of the database code and the wait >behavior is >the same. ???? > > > >I checked the machines involved from top and they are barely >being worked. > > > >I checked the waits in the database and the largest event time is from >'sqlnet message from client'. > > > >The company restored the system to a previous incarnation and >it worked just >fine. > > > >Something is very very wrong. > > > >....and thoughts would be greatly appreciated. > > > >Sorry for the vagueness. > > > >...questions welcomed. > > > >Thanks for any info!!!! > > > >chris > > > > > > > > > >---------------------------------------------------------------- >Please see the official ORACLE-L FAQ: http://www.orafaq.com >---------------------------------------------------------------- >To unsubscribe send email to: oracle-l-request@xxxxxxxxxxxxx >put 'unsubscribe' in the subject line. >-- >Archives are at //www.freelists.org/archives/oracle-l/ >FAQ is at //www.freelists.org/help/fom-serve/cache/1.html >----------------------------------------------------------------- > ---------------------------------------------------------------- Please see the official ORACLE-L FAQ: http://www.orafaq.com ---------------------------------------------------------------- To unsubscribe send email to: oracle-l-request@xxxxxxxxxxxxx put 'unsubscribe' in the subject line. -- Archives are at //www.freelists.org/archives/oracle-l/ FAQ is at //www.freelists.org/help/fom-serve/cache/1.html -----------------------------------------------------------------