[overlord-issues] [JBoss JIRA] (RTGOV-574) Investigate 'index already exists' exception when previous command indicates does not exist

[Gary Brown (JIRA)]

    [ https://issues.jboss.org/browse/RTGOV-574?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13004228#comment-13004228 ] 

Gary Brown commented on RTGOV-574:
----------------------------------

Thanks for the clarification - could you raise a bug regarding the context classloader and we can see whether this can be addressed and still support fuse :)

> Investigate 'index already exists' exception when previous command indicates does not exist
> -------------------------------------------------------------------------------------------
>
>                 Key: RTGOV-574
>                 URL: https://issues.jboss.org/browse/RTGOV-574
>             Project: RTGov (Run Time Governance)
>          Issue Type: Task
>            Reporter: Gary Brown
>            Assignee: Gary Brown
>
> When updating to Elasticsearch 1.3.2 (RTGOV-568), noticed that when the EAP server was restarted, it resulted in an IndexAlreadyExistsException from ES.
> However this exception was generated from code called to create the indexes - however prior to that call it checks with the ES node whether the indexes already exist.
> Ivan McKinley found the following reference that may be relevant: http://stackoverflow.com/questions/23883110/elasticsearch-index-exists-not-working-reliable
> He also suggests:
> "If this is the problem then we may have an impact from running “embedded” mode.  The more data we have the longer response  from the async process perhaps. 
> externalising the es process may resolve this.. and another test is to remove the entire contents of the index/type/  to further prove that we are affect the amount of data.
> The various scenarios here are definitely interesting for us regarding  documentation"



--
This message was sent by Atlassian JIRA
(v6.3.1#6329)