Redis running jobs are not suspended in case of crash

When there is brutal crash (without a shutdown) the running jobs stay in the nuxeo:work:run queue.

On startup the running job are not relaunched and they are displayed as running in the Admin Center/Activity/Background job.

On normal shutdown the running job on the instance are moved to the suspended queue nuxeo:work:prev and re launch at startup time.

The current work around is to watch in the admin center/activity/background job for running job with old schedule time. Then remove these job by hand from Redis (or drop the redis db when application is down).