You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@oozie.apache.org by "Purshotam Shah (JIRA)" <ji...@apache.org> on 2016/04/21 21:09:25 UTC

[jira] [Commented] (OOZIE-1402) Increase retry interval for non-progressing coordinator action using exponential backoff concept

    [ https://issues.apache.org/jira/browse/OOZIE-1402?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15252482#comment-15252482 ] 

Purshotam Shah commented on OOZIE-1402:
---------------------------------------

{code}
     <property>
+        <name>oozie.service.coord.input.check.requeue.interval.increase</name>
+        <value>30000</value>
+        <description>This value will be added into oozie.service.coord.input.check.requeue.interval and resulting value
+will be the requeue interval for the actions waiting for a long time without any input.
+        </description>
+    </property>
{code}
Can you set it to 0? If somebody want to use it, they can override.
Time should be in second, I guess nobody is going to configure in ms.
Can you also change the JIRA title? We are not doing exponential backoff.

> Increase retry interval for non-progressing coordinator action using exponential backoff concept 
> -------------------------------------------------------------------------------------------------
>
>                 Key: OOZIE-1402
>                 URL: https://issues.apache.org/jira/browse/OOZIE-1402
>             Project: Oozie
>          Issue Type: Improvement
>    Affects Versions: trunk
>            Reporter: Mona Chitnis
>            Assignee: Satish Subhashrao Saley
>            Priority: Minor
>             Fix For: trunk
>
>         Attachments: OOZIE-1402-1.patch, OOZIE-1402-2.patch
>
>
> Currently every coordinator action is retried to check data directory in the
> next minute. 
> We could make it better by waiting longer for coordinator action that is not
> progressing (i.e. find no new directory) for repeated retries
> The waiting time should start from 1 minute for X retries. Then the action
> should wait for 2 minutes. After X retries it should wait for 3. The same way
> it will go to some max-wait-time and stay there until timeout



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