Clustering: Jobs never get executed again when scheduler instance was aborted

[dannys-janssen]

When clustering is enabled, more than one scheduler instances are running and one scheduler instance is aborted while executing a long running job, the trigger never gets fired again.

After exploring the quartz collection in Mongo DB the "Trigger" document is still in State "Blocked" and there is a "FiredTrigger" document in the collection that relates to the "Trigger" document.
Also the "Scheduler" document of the aborted instance is not there anymore.

See attached file:

quartz_collection_export.zip

To find the failing job apply following filter db.getCollection('quartz').find({"State": { $exists: 1, $ne: "Waiting"}}). Export was done while no job was executing.

When a scheduler instance goes down then the FiredTrigger should be removed from the collection and the state of the Trigger should be "Waiting" or the job should be re-scheduled on a running scheduler instance.

Thanks in advance

Dannys