This is in regards to Crowd version: 3.3.5
Last night Crowd went into a hung state due to the:
java.lang.OutOfMemoryError: GC overhead limit exceeded exception
What I noticed after restarting crowd, was that this particular Error has appeared the last 3 days in a row at the same time. The time coincides with the scheduled daily xml backups.
2020-02-13 02:01:05,435 Caesium-2-3 ERROR [scheduler.caesium.impl.SchedulerQueueWorker] Unhandled exception thrown by job QueuedJob[jobId=AutomatedBackup,dead
java.lang.OutOfMemoryError: GC overhead limit exceeded
When I look through the log analyzer it takes me to the following page:
Which recommends increasing the -Xmx size.
I want to know if there is a guideline i can follow that determines the right size to increase that parameter to. Based on number of users, number of applications using this crowd instance as an authenticator, etc..
Currently this is what the setenv.sh is set to:
JAVA_OPTS="-Xms128m -Xmx512m -Dfile.encoding=UTF-8 $JAVA_OPTS"
would increasing from 512m to 1024m suffice?
Also crowd only crashed last night even though the error message was present for the last 3 days.
Thanks in advance