[
https://issues.apache.org/jira/browse/CONNECTORS-1180?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14497892#comment-14497892
]
Karl Wright commented on CONNECTORS-1180:
-----------------------------------------
FWIW, if you are using Zookeeper with multiple agents processes, it is possible
that the underlying cause of this problem for you is a misconfigured cluster.
If, for example, you have multiple instances of Zookeeper and they are not
talking to one another, then each cluster member essentially operates
independently. For performance reasons, ManifoldCF's stuffer thread uses
Zookeeper-based locking to prevent an active document from winding up in more
than one queue. If that locking is not working, then a document can be
processed by more than one worker thread at the same time, leading to precisely
this kind of result.
> Jobs abort with "Error: Unexpected jobqueue status - record id 1427880204801,
> expecting active status, saw 4"
> -------------------------------------------------------------------------------------------------------------
>
> Key: CONNECTORS-1180
> URL: https://issues.apache.org/jira/browse/CONNECTORS-1180
> Project: ManifoldCF
> Issue Type: Bug
> Affects Versions: ManifoldCF 2.0.2
> Environment: Using the Windows Share repository, Solr output
> connectors.
> Using Zookeeper.
> Reporter: Andreas Baumann
> Assignee: Karl Wright
>
> Jobs abort out of the blue with
> "Error: Unexpected jobqueue status - record id 1427880204801, expecting
> active status, saw 4"
> I had the impression that I could force it by clicking in the admin interface
> and for instance when I pressed refresh in "Status and Job Management".
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
