Skip to Content.
Sympa Menu

grouper-users - RE: [grouper-users] Out of memory during loader start with full sync = true

Subject: Grouper Users - Open Discussion List

List archive

RE: [grouper-users] Out of memory during loader start with full sync = true


Chronological Thread 
  • From: Gagné Sébastien <>
  • To: "Bryan E. Wooten" <>, <>
  • Subject: RE: [grouper-users] Out of memory during loader start with full sync = true
  • Date: Fri, 10 May 2013 14:02:09 -0400
  • Authentication-results: sfpop-ironport01.merit.edu; dkim=neutral (message not signed) header.i=none

You can set a system variable for the memory, here in .bashrc we have :

export MEM_MAX=2000m

 

The only time we got an outofmemory error was in the initial seed of all the grouper groups in AD. I think there might be a memory leak when doing ldap create/modify since when I’m doing a bulkSync with few changes (<5000 group create) I don’t get any error.

 

You could try to set more memory and simply run the loader another time. That was our solution since the second run won’t have to create all the previously created groups.

 

PS. The “omit” parameters are set to true here.

 

De : [mailto:] De la part de Bryan E. Wooten
Envoyé : 10 mai 2013 10:00
À :
Objet : [grouper-users] Out of memory during loader start with full sync = true

 

We renamed and moved some folders and groups. So I decided that a full sync was in order. I set all my Load LDAP groups cron entry to run next Wednesday and made the same schedule for the PSP change log consumer.

 

While it was running (this can take a long while with many groups having over 10k members, some 100k) I started to notice this exception (it occurred well over 100 times):

java.lang.RuntimeException: Active PITGroup with sourceId=6fe2f5c77a7b47509e86ead620ed2c2d not found,

Problem in HibernateSession: HibernateSession: isNew: true, isReadonly: false, grouperTransactionType: READ_WRITE_NEW

        at edu.internet2.middleware.grouper.internal.dao.hib3.Hib3PITGroupDAO.findBySourceIdActive(Hib3PITGroupDAO.java:100)

        at edu.internet2.middleware.grouper.changeLog.ChangeLogTempToEntity.processGroupUpdate(ChangeLogTempToEntity.java:286)

 

And then after about an hour running I got an out of memory error so I stopped the loader.

 

I see in the gsh start script this:

 

if [ "$MEM_MAX" = "" ]; then

MEM_MAX=750m

Fi

 

I guess I need to set this in setenv.sh?

 

But do I also need to make these settings in grouper-loader.properties:

 

# Omit diff responses from bulk response to conserve memory.

# changeLog.psp.fullSync.omitDiffResponses = true

 

# Omit sync responses from bulk response to conserve memory.

# changeLog.psp.fullSync.omitSyncResponses = true

 

Thanks,

 

Bryan




Archive powered by MHonArc 2.6.16.

Top of Page