Support Article

Conclusion Cache Memory Leak



User's production server has been running since few days. The System Management Application (SMA) shows Conclusion cache count around 400k and size around 893 m.

User claims that the range of other servers is around 200k and 350m. There server's instances with a shorter life span two to three days, have about 100k as a peak value.

For all their 22 servers one can see that the "field value" peak always goes up over a period of time.

Error Messages

Not Applicable.

Steps to Reproduce

Monitor Field Value conclusion cache in SMA.

Root Cause

A defect in Pegasystems’ code or rules. Users’ app is issuing localization lookups for fields which contain unique identifiers.

These are naturally ‘misses’ in the Field Value Conclusion cache. But user is caching the misses. There is no mechanism to clean out these cached misses, so they cause a memory leak that eventually brings down production servers. 



Apply HFix-25967
User should also specify a new prconfig entry with following values - (Value true means - PRPC does not create dummy field value conclusion objects for nonexistent field values - default Value is false) 
For example: <env name="conclusion/cache/FieldValue/DisableNullConclusion" value="true"/>
Suggest Edit

Published March 4, 2016 - Updated October 8, 2020

100% found this useful

Have a question? Get answers now.

Visit the Collaboration Center to ask questions, engage in discussions, share ideas, and help others.