[ 
https://issues.apache.org/jira/browse/NIFI-1857?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15318534#comment-15318534
 ] 

ASF GitHub Bot commented on NIFI-1857:
--------------------------------------

Github user ijokarumawak commented on a diff in the pull request:

    https://github.com/apache/nifi/pull/497#discussion_r66078443
  
    --- Diff: 
nifi-nar-bundles/nifi-framework-bundle/nifi-framework/nifi-site-to-site/src/main/java/org/apache/nifi/remote/StandardRootGroupPort.java
 ---
    @@ -236,7 +235,8 @@ private void onTrigger(final ProcessContext context, 
final ProcessSession sessio
                 return;
             }
     
    -        session.commit();
    +        // TODO: Comfirm this. Session.commit here is not required since 
it has been committed inside receiveFlowFiles/transferFlowFiles.
    +        // session.commit();
    --- End diff --
    
    HTTP Site-to-Site has to split the transferring process to two HTTP 
requests. The first request calls StandardRootGroupPort's onTrigger() as Socket 
Site-to-Site does. But HTTP version has to hold the transaction, and the second 
HTTP request actually does the session.commit. Whether it calls 
`session.commit()` is determined by each transport protocol inside 
`receiveFlowFiles` or `transferFlowFiles`.
    
    This `session.commit` doesn't let HTTP version to hold the transaction, and 
it was redundant since it has been committed inside 
receiveFlowFiles/transferFlowFiles.
    I just wanted to confirm that this doesn't break anything. Would you 
confirm it @markap14 ?


> Support HTTP(S) as a transport mechanism for Site-to-Site
> ---------------------------------------------------------
>
>                 Key: NIFI-1857
>                 URL: https://issues.apache.org/jira/browse/NIFI-1857
>             Project: Apache NiFi
>          Issue Type: Improvement
>          Components: Core Framework
>            Reporter: Koji Kawamura
>            Assignee: Koji Kawamura
>   Original Estimate: 480h
>  Remaining Estimate: 480h
>
> We should add support for using HTTP(S) for site-to-site to be an alternative 
> to the current socket based approach.
> This would support the same push based or pull based approach site-to-site 
> offers now but it would use HTTP(S) for all interactions to include learning 
> about ports, learning about NCM topology, and actually exchanging data. This 
> mechanism should also support interaction via an HTTP proxy.
> This would also require some UI work to allow the user to specify which 
> protocol for site-to-site to use such as 'raw' vs 'http'. We also need to 
> document any limitations with regard to SSL support for this mode and we'd 
> need to provide 'how-to' when using proxies like http_proxy or something else.



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

Reply via email to