[ 
https://issues.apache.org/jira/browse/NIFI-1301?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15064801#comment-15064801
 ] 

ASF subversion and git services commented on NIFI-1301:
-------------------------------------------------------

Commit bef3fc8b40cec21fd7ce86fde3e812fd348f1341 in nifi's branch 
refs/heads/master from [~markap14]
[ https://git-wip-us.apache.org/repos/asf?p=nifi.git;h=bef3fc8 ]

NIFI-1301: Ensure that when creating site-to-site connection, if remote 
instance is applying backpressure that we do not block indefinitely waiting for 
the connection to be made

Signed-off-by: joewitt <[email protected]>


> RPG can become hung when input port on target NiFi does not accept files 
> because of backpressure
> ------------------------------------------------------------------------------------------------
>
>                 Key: NIFI-1301
>                 URL: https://issues.apache.org/jira/browse/NIFI-1301
>             Project: Apache NiFi
>          Issue Type: Bug
>          Components: Core Framework
>    Affects Versions: 0.4.0
>            Reporter: Matthew Clarke
>            Assignee: Oleg Zhurakousky
>             Fix For: 0.4.1
>
>         Attachments: 
> 0001-NIFI-1301-Ensure-that-when-creating-site-to-site-con.patch, 
> dumpRPGHang.log, make_badness_happen_with_site-to-site.xml
>
>
> If i have an RPG sending to a input port on another NiFi and that input port 
> is no longer accepting files because of backpressure on the input port 
> connection, the sending NiFi starts logging that the destination port is full.
> The potential problem is that it create a log message for every file that is 
> queueing up on the connection(s) feeding that RPG destination input port.  
> This logging seems excess.
>  
> The critical issue:
> Also when i select "disable transmission" I get the spinning wheel but thread 
> does not complete. Entire instance is now unresponsive.
> Even though it became hung on request to disable transmission, on restart it 
> started running again. Hung again. 
> The only options for regaining control is to either restart with auto resume 
> state set to false or take back pressure off destination input port so thread 
> can complete.



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

Reply via email to