Nuxeo 5.6 / Memory leak executing nuxeo.workManagerCleanup

As we transition to the new Hyland Community, some post links may be temporarily disrupted as we finalize redirects. We apologize for the inconvenience and are working to resolve this as quickly as possible. Details.

Renaud_Rousset

We are using Nuxeo 5.6 in production and we see a memory leak each time nuxeo.workManagerCleanup internal job is executed.

This job is executed using Quartz every few minutes.

In the code of the Nuxeo plateform, the fireEvent method of EventServiceImpl (nuxeo core / event) raise an event for the job to be executed, it adds a new event in a map retained in memory by a ThreadLocal (compositeBundle) : CompositeEventBundle b = compositeBundle.get(); b.push(shallowEvent);

It is not removed afterward, the methods handleTxRollbacked and handleTxCommited are never called.

I traced this behavior debuging the server from Eclipse.

I checked the code history and it has changed since Nuxeo 5.6, but does anyone know if there is a Jira for this bug and in witch version it is solved.

Does someone know how to solve this isssue without changing version of Nuxeo ?

Florent_Guillau

No idea if that's really it but https://github.com/nuxeo/nuxeo/commit/b75e587d3ccb18dd0f4d54bd0c0a908031659c38 is a likely candidate for this fix.