This is what is happening:
The data transfer occurs from Machine A (home grown app running Oracle, Apache on Linux), through the broker piece on machine B (Windows 2008 running Weblogic) , to PeopleSoft on machine C (running Oracle on HPUX 11.23). The broker piece is a servlet which is part of PeopleSoft PeopleTools on Weblogic 9.2. The communication seems to bottleneck with a large number of records, where server A times out waiting for a response, but eventually, the end PeopleSoft environment gets the records long after the timeout on machine A occurs
This is what we see:
When we send under 500 records from Machine A, Machine C gets them almost immediately and sends a response back to Machine A. When we send over 500 (last night 912) we see them leaving Machine A but Machine A never gets a response and times out after 10 minutes. Machine C gets the ALL the records about 35 minutes later. It appears that something on Machine B is choking with the volume but eventually gets the records through.
Does anyone have any idea what parameters we can try changing? Anyone know a Weblogic consultant in the NY area? Thank you
thanks for the response. it didn't make a difference. it consistently takes 35 minutes to get the file. it appears the tread goes stuck for the whole duration and then frees up when we get the file. not sure if that helps at all.