[ 
https://issues.apache.org/jira/browse/NIFI-2032?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15340910#comment-15340910
 ] 

ASF subversion and git services commented on NIFI-2032:
-------------------------------------------------------

Commit d64fe416be6627682c7458ded71ba2ea6e05a899 in nifi's branch 
refs/heads/master from [~ozhurakousky]
[ https://git-wip-us.apache.org/repos/asf?p=nifi.git;h=d64fe41 ]

NIFI-2032 put back bulletinRepo reporting logic

NIFI-2032 polishing (+2 squashed commits)
Squashed commits:
[4549443] NIFI-2032 fixed 'enableControllerServices' logic
added getRequiredControllerServices() operation to ControllerServiceNode
[3d95c8e] NIFI-2032 Removed legacy references to bulletin repository in 
StandardControllerServiceProvider. (+1 squashed commit)
Squashed commits:
[6402eb9] NIFI-2032 fixed 'enableControllerServices' logic
added getRequiredControllerServices() operation to ControllerServiceNode

This closes #541.

Signed-off-by: Andy LoPresto <[email protected]>


> Processors could be started before the Controller Services that they depend on
> ------------------------------------------------------------------------------
>
>                 Key: NIFI-2032
>                 URL: https://issues.apache.org/jira/browse/NIFI-2032
>             Project: Apache NiFi
>          Issue Type: Bug
>          Components: Core Framework
>            Reporter: Mark Payne
>            Assignee: Oleg Zhurakousky
>            Priority: Critical
>             Fix For: 1.0.0, 0.7.0
>
>
> in the StandardControllerServiceProvider, we enable a Collection of 
> Controller Services but do so in the background with a limited number of 
> threads. We need to ensure that all Controller Services have at least become 
> ENABLING before returning from this method. Otherwise, Processors that depend 
> on them could attempt to start. If this happens, the Processor will be 
> invalid because it references a Disabled Controller Service. As a result, the 
> Processor will not start. In a clustered environment, we will end up with 
> inconsistent run states across the nodes.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Reply via email to