Last updated on JULY 05, 2017
Applies to:Oracle Coherence - Version 18.104.22.168 and later
Information in this document applies to any platform.
Coherence nodes using new G1 GC algorithm that is introduced in the Java 7 could see this problem.
- Created distributed cache running on 4 storage nodes.
- Loading data 300,000 records from a database using JPACacheStore.
- Using JConsole to monitor memory usage of each nodes. After the initial load is finished, all nodes use around 500MB
- Using a script to update data by modifying value in the cache with around 300 requests/sec over a 60 minute period.
- Memory of each node increases while the script is running.
- After 60 minutes, JConsole shows that each node is using memory around 1GB.
- Leaving the cluster for one day with no active workload, then checking with JConsole shows the memory of each nodes is still around 1 GB.
The question is why memory usage does not reduce to around 500 MB after script finish when cache size still 300,000 objects?
Sign In with your My Oracle Support account
Don't have a My Oracle Support account? Click to get started
Million Knowledge Articles and hundreds of Community platforms