Details
-
Bug
-
Status: Closed
-
Top
-
Resolution: Fixed
-
2.20.09, 2.22.04, 2.23.02-alpha
-
None
-
None
Description
To reproduce (and see it):
- Start a java profiler (yourkit)
- Connect to your running CMS
- Open the memory view (with class list) and filter to include the following classes:
org.hippoecm.repository.impl.DocumentManagerImpl
org.hippoecm.repository.impl.WorkflowManagerImpl
- Log in to the CMS. The number of objects for the classes mentioned above increases by one.
- Open, then delete a document
- Log out
- In the profiler, force garbage collection a few times
Expected behavior: The number of objects for the classes mentioned above decreases by one.
Actual behavior: The number of objects for the classes mentioned above doesn't decrease.
This only seems to occur when deleting a document. When publishing, taking offline, moving, renaming or copying a document, no memory leak is observed.
Doing the logging in, deleting and logging out about 300+ times, brings your memory usage up to almost 500MB.
Attachments
Issue Links
- relates to
-
REPO-320 Unhandled exception in the internal of Hippo Repository when delete event workflow configured.
- Closed