[jira] [Commented] (NIFI-2776) When node connects back the to the cluster intermittently processor does not return to the same state as its in the cluster

["Mark Payne (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/NIFI-2776?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16276849#comment-16276849 ] 

Mark Payne commented on NIFI-2776:
----------------------------------

I was able to replicate this bug by performing the following steps:

1. Create a 2 node cluster and start it.
2. Create a simple flow: DebugFlow -> UpdateAttribute.
3. Configure DebugFlow with the "@OnStopped Pause Time" property set to 20 seconds
4. Start the flow
5. Go to the cluster menu and disconnect one of the nodes
6. Navigate to the node that was disconnected (this is best done in a separate browser tab)
7. Stop DebugFlow processor on the disconnected node
8. Quickly navigate back to the other node and re-connect the disconnected node.
9. Once the node has reconnected, there will be a bulletin with the error message:
{code}
Can not start 'AtomicReference' since it's already in the process of being started or it is DISABLED - STOPPING
{code}

Note that the error message is also erroneous, indicating that 'AtomicReference' could not be started.

Will look into this.

> When node connects back the to the cluster intermittently processor does not return to the same state as its in the cluster
> ---------------------------------------------------------------------------------------------------------------------------
>
>                 Key: NIFI-2776
>                 URL: https://issues.apache.org/jira/browse/NIFI-2776
>             Project: Apache NiFi
>          Issue Type: Bug
>          Components: Core Framework
>    Affects Versions: 1.1.0
>            Reporter: Arpit Gupta
>            Assignee: Mark Payne
>
> Here is the scenario
> 1. Create a flow and start a processor
> 2. Disconnect a node
> 3. On the disconnected node stop the above processor
> 4. Connect the above node to the cluster
> 5. Wait 30s.
> 6. Check if the processor started on the node that was connected in #4.
> Very intermittently we see that the processor does not get into running state.
> When we query the processor status on the node we get the following bulletin
> {code}
> "bulletins": [{
> 		"id": 0,
> 		"groupId": "275e45f8-0157-1000-ffff-fffff191c079",
> 		"sourceId": "275e4abc-0157-1000-0000-00005740dd0c",
> 		"timestamp": "06:24:35 UTC",
> 		"nodeAddress": "host:port",
> 		"canRead": true,
> 		"bulletin": {
> 			"id": 0,
> 			"nodeAddress": "host:port",
> 			"category": "Log Message",
> 			"groupId": "275e45f8-0157-1000-ffff-fffff191c079",
> 			"sourceId": "275e4abc-0157-1000-0000-00005740dd0c",
> 			"sourceName": "putES",
> 			"level": "WARNING",
> 			"message": "PutElasticsearch[id=275e4abc-0157-1000-0000-00005740dd0c] Can not start 'PutElasticsearch' since it's already in the process of being started or it is DISABLED - STOPPING",
> 			"timestamp": "06:24:35 UTC"
> 		}
> 	}],
> {code}



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)