Mark Payne created NIFI-5361:
--------------------------------
Summary: Upon restart of NiFi, some processors show Active Threads
but never actually run
Key: NIFI-5361
URL: https://issues.apache.org/jira/browse/NIFI-5361
Project: Apache NiFi
Issue Type: Bug
Components: Core Framework
Reporter: Mark Payne
Assignee: Mark Payne
Upon restart, if there are a large number of processors that take a long time
to start up (i.e., their {{@OnScheduled}} methods take a long time to
complete), then other processors may also timeout while waiting for their
{{@OnScheduled}} method to complete. It looks like this is due to the fact that
when we schedule many processors we determine the 'timeout timestamp' before
submitting the background task to call {{@OnScheduled}}. As a result, if
several processors don't finish or take a long time to finish, then other
processors may just sit, waiting for their chance to have {{@OnScheduled}}
called and eventually timeout before they even have a chance to invoke the
method.
When this happens, the result is that the Runnable to call {{@OnScheduled}}
never is called before the corresponding {{Future}} is canceled. So the
Processor is never scheduled. However, its Active Thread count has already been
incremented. So it appears as if the Processor has an Active Thread but it will
never make progress.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
