This is a discussion on Memory Footprint Growing within the Informix forums, part of the Database Server Software category; --> Running AIX 5.2, Informix 9.3.UC7 User sessions monitored from onstat -g ses show a memory footprint that is growing ...
| |||||||
| FAQ | Members List | Calendar | Search | Today's Posts | Mark Forums Read |
| ||||
| Running AIX 5.2, Informix 9.3.UC7 User sessions monitored from onstat -g ses show a memory footprint that is growing over time. This is an OLTP environment with approx. 300 users doing things such as Invoicing, Sale Entry, Job Costing/Estimating, etc. When the users are on the programs for extended periods of time, the memory footprint is growing unexpectedly. On some data-entry programs, the footprint stablizes and remains within a consistent window, but applications that are batch type processing do not. This was causing our system to run into a bug on 9.3 where the engine does not handle the 2GB limit correctly and crashes the engine rather then sending the -116 error (no more memory available) to the process. We have worked around this by lowering our number of BUFFERS to allow for more virtual memory. In the mean time, we need to determine why the memory footprints are growing so we don't have to constantly monitor these types of applications to ensure that users are periodically leaving the application and restarting. Any ideas?????? Thanks |
| |||
| What is the app written in? Look for unreleased cursors in the code etc Rob Burba wrote: > > Running AIX 5.2, Informix 9.3.UC7 > > User sessions monitored from onstat -g ses show a memory footprint > that is growing over time. This is an OLTP environment with approx. > 300 users doing things such as Invoicing, Sale Entry, Job > Costing/Estimating, etc. When the users are on the programs for > extended periods of time, the memory footprint is growing > unexpectedly. On some data-entry programs, the footprint stablizes > and remains within a consistent window, but applications that are > batch type processing do not. > > This was causing our system to run into a bug on 9.3 where the engine > does not handle the 2GB limit correctly and crashes the engine rather > then sending the -116 error (no more memory available) to the process. > We have worked around this by lowering our number of BUFFERS to allow > for more virtual memory. In the mean time, we need to determine why > the memory footprints are growing so we don't have to constantly > monitor these types of applications to ensure that users are > periodically leaving the application and restarting. > > Any ideas?????? > > Thanks -- Paul Watson # Oninit Ltd # Growing old is mandatory Tel: +44 1436 672201 # Growing up is optional Fax: +44 1436 678693 # Mob: +44 7818 003457 # www.oninit.com # |
| ||||
| The app is all 4gl (7.31 toolset) We are looking at unprepared cursors and unfreed cursors/statments. I am trying to figure out if anyone has any other areas to look for this problem or if they have seen this problem in their applications. I notice that the memory usage grows more substantially when views are involved but I don't know if that is pertinent to views or because the particular view that I am looking at is somewhat complex. Paul Watson <paul@oninit.com> wrote in message news:<404374C6.E5706298@oninit.com>... > What is the app written in? Look for unreleased cursors in the code > etc > > Rob Burba wrote: > > > > Running AIX 5.2, Informix 9.3.UC7 > > > > User sessions monitored from onstat -g ses show a memory footprint > > that is growing over time. This is an OLTP environment with approx. > > 300 users doing things such as Invoicing, Sale Entry, Job > > Costing/Estimating, etc. When the users are on the programs for > > extended periods of time, the memory footprint is growing > > unexpectedly. On some data-entry programs, the footprint stablizes > > and remains within a consistent window, but applications that are > > batch type processing do not. > > > > This was causing our system to run into a bug on 9.3 where the engine > > does not handle the 2GB limit correctly and crashes the engine rather > > then sending the -116 error (no more memory available) to the process. > > We have worked around this by lowering our number of BUFFERS to allow > > for more virtual memory. In the mean time, we need to determine why > > the memory footprints are growing so we don't have to constantly > > monitor these types of applications to ensure that users are > > periodically leaving the application and restarting. > > > > Any ideas?????? > > > > Thanks |