Support Article
Server slow post Sandbox upgrade to Pega 7.1.8
Summary
Sandbox has recently been upgraded to Pega 7.1.8. There are frequent thread dumps and requestor lock exceptions. The server hangs frequently with few users accessing it and it needs to be restarted frequently.
Error Messages
2015-07-15 10:42:24,081 [ WebContainer : 9] [ STANDARD] [ APP:01.40.01] (ernal.basic.SystemSettingsImpl) ERROR <ip>|<server> USER - failed to get setting
java.lang.StackOverflowError
at com.pega.pegarules.bootstrap.util.VicariousThreadLocal.get(VicariousThreadLocal.java:157)
at com.pega.pegarules.priv.factory.StringBuilderThreadLocalContainer.acquire(StringBuilderThreadLocalContainer.java:66)
at com.pega.pegarules.pub.util.StringUtils.concat(StringUtils.java:1125)
at com.pega.pegarules.pub.util.StringUtils.concat(StringUtils.java:1001)
at com.pega.pegarules.exec.internal.basic.SystemSettingsImpl.getDynamic(SystemSettingsImpl.java:234)
at com.pega.pegarules.exec.internal.basic.SystemSettingsImpl.getDynamic(SystemSettingsImpl.java:221)
at com.pega.pegarules.data.internal.clipboard.DeclarativePageInstanceContainerImpl.<init>(DeclarativePageInstanceContainerImpl.java:70)
at com.pega.pegarules.data.internal.PRDataProviderImpl.getDeclarativePageInstanceContainer(PRDataProviderImpl.java:965)
at com.pega.pegarules.session.internal.mgmt.base.AbstractDataPageHandler.<init>(AbstractDataPageHandler.java:52)
at com.pega.pegarules.session.internal.mgmt.base.handler.ReadOnlyDataPageHandler.<init>(ReadOnlyDataPageHandler.java:47)
at com.pega.pegarules.session.internal.mgmt.base.RequestorPageDir$ThreadFamilyDirectory.<init>(RequestorPageDir.java:915)
at com.pega.pegarules.session.internal.mgmt.base.RequestorPageDir.getPageDirectoryInternalForFamily(RequestorPageDir.java:868)
at com.pega.pegarules.session.internal.mgmt.base.RequestorPageDir.getPageDirectoryInternalForCurrentFamily(RequestorPageDir.java:880)
at com.pega.pegarules.session.internal.mgmt.base.RequestorPageDir.getPageDirectoryInternal(RequestorPageDir.java:845)
.
.
.
Steps to Reproduce
Have more than 10 users login with the developer portal and execute the application functionality.
Root Cause
The root cause of this problem is a defect in Pegasystems’ code/rules. This stackoverflow eats up all database connections because final blocks are not executed.
Resolution
This issue is resolved by HFix-23174.
Published August 20, 2015 - Updated October 8, 2020
Have a question? Get answers now.
Visit the Collaboration Center to ask questions, engage in discussions, share ideas, and help others.