You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@oozie.apache.org by "Hadoop QA (JIRA)" <ji...@apache.org> on 2011/09/08 06:31:09 UTC
[jira] [Created] (OOZIE-113) GH-78: Option to rerun from failed
node.
GH-78: Option to rerun from failed node.
----------------------------------------
Key: OOZIE-113
URL: https://issues.apache.org/jira/browse/OOZIE-113
Project: Oozie
Issue Type: Bug
Reporter: Hadoop QA
Oozie will provide a mechanism for customers to retry a failed action within a workflow. If the action is successful, the dependent action(s) will be executed as well.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (OOZIE-113) GH-78: Option to rerun from failed
node.
Posted by "Hadoop QA (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/OOZIE-113?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13101828#comment-13101828 ]
Hadoop QA commented on OOZIE-113:
---------------------------------
mislam77 remarked:
This was the original proposal around 10 months back in yahoo: support it into oozie core.
However, the counter proposal was not to support in oozie supported code. Because there is already a work-around where user can follow these two steps:
* At first call, get all the succeeded actions using something like "oozie job -info <id>" and then parsing.
* Secondly call oozie rerun after setting skipnodes as the all successful actions.
Vase on this logic, we discussed to provide a utility for the customer to combine these two steps into one as reference only(i.e. not supportable)
Do you think that logic doesn't hold anymore.
Comments.
> GH-78: Option to rerun from failed node.
> ----------------------------------------
>
> Key: OOZIE-113
> URL: https://issues.apache.org/jira/browse/OOZIE-113
> Project: Oozie
> Issue Type: Bug
> Reporter: Hadoop QA
>
> Oozie will provide a mechanism for customers to retry a failed action within a workflow. If the action is successful, the dependent action(s) will be executed as well.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (OOZIE-113) GH-78: Option to rerun from failed
node.
Posted by "Hadoop QA (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/OOZIE-113?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13101826#comment-13101826 ]
Hadoop QA commented on OOZIE-113:
---------------------------------
anew remarked:
I am not sure why this should be implemented in a client.
I would prefer implementing a "retry workflow" in the Oozie core.
> GH-78: Option to rerun from failed node.
> ----------------------------------------
>
> Key: OOZIE-113
> URL: https://issues.apache.org/jira/browse/OOZIE-113
> Project: Oozie
> Issue Type: Bug
> Reporter: Hadoop QA
>
> Oozie will provide a mechanism for customers to retry a failed action within a workflow. If the action is successful, the dependent action(s) will be executed as well.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (OOZIE-113) GH-78: Option to rerun from failed
node.
Posted by "Hadoop QA (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/OOZIE-113?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13101825#comment-13101825 ]
Hadoop QA commented on OOZIE-113:
---------------------------------
tucu00 remarked:
* When you say 'failed action' you mean an action that ended in ERROR? If so, a workflow could have multiple of such actions and alternate executions paths because of them. How things would be handled then?
* Why a different CLI? Why not a new option for the 'oozie job' command?
> GH-78: Option to rerun from failed node.
> ----------------------------------------
>
> Key: OOZIE-113
> URL: https://issues.apache.org/jira/browse/OOZIE-113
> Project: Oozie
> Issue Type: Bug
> Reporter: Hadoop QA
>
> Oozie will provide a mechanism for customers to retry a failed action within a workflow. If the action is successful, the dependent action(s) will be executed as well.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (OOZIE-113) GH-78: Option to rerun from failed
node.
Posted by "Hadoop QA (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/OOZIE-113?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13099810#comment-13099810 ]
Hadoop QA commented on OOZIE-113:
---------------------------------
mislam77 remarked:
We plan to provide another client utility (at bin/) for this functionality.
Details:
* We will write a Java class that will call to oozie server WS API twice
* First time , to get the job info.
* second time, to rerun with appropriate "skipNodes" option.
* This java class will be invoked from the wrapper script oozie_wf_rerun.sh
> GH-78: Option to rerun from failed node.
> ----------------------------------------
>
> Key: OOZIE-113
> URL: https://issues.apache.org/jira/browse/OOZIE-113
> Project: Oozie
> Issue Type: Bug
> Reporter: Hadoop QA
>
> Oozie will provide a mechanism for customers to retry a failed action within a workflow. If the action is successful, the dependent action(s) will be executed as well.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Closed] (OOZIE-113) GH-78: Option to rerun from failed
node.
Posted by "Roman Shaposhnik (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/OOZIE-113?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Roman Shaposhnik closed OOZIE-113.
----------------------------------
Resolution: Fixed
> GH-78: Option to rerun from failed node.
> ----------------------------------------
>
> Key: OOZIE-113
> URL: https://issues.apache.org/jira/browse/OOZIE-113
> Project: Oozie
> Issue Type: Bug
> Reporter: Hadoop QA
>
> Oozie will provide a mechanism for customers to retry a failed action within a workflow. If the action is successful, the dependent action(s) will be executed as well.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (OOZIE-113) GH-78: Option to rerun from failed
node.
Posted by "Hadoop QA (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/OOZIE-113?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13101829#comment-13101829 ]
Hadoop QA commented on OOZIE-113:
---------------------------------
anew remarked:
The Oozie client is a thin client. Thus this belongs in the server not the client.
> GH-78: Option to rerun from failed node.
> ----------------------------------------
>
> Key: OOZIE-113
> URL: https://issues.apache.org/jira/browse/OOZIE-113
> Project: Oozie
> Issue Type: Bug
> Reporter: Hadoop QA
>
> Oozie will provide a mechanism for customers to retry a failed action within a workflow. If the action is successful, the dependent action(s) will be executed as well.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (OOZIE-113) GH-78: Option to rerun from failed
node.
Posted by "Hadoop QA (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/OOZIE-113?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13101831#comment-13101831 ]
Hadoop QA commented on OOZIE-113:
---------------------------------
bansalmayank remarked:
Closed by 8d5bcbffbc25bbb734ef4113e55cc99621d9c5b9 Option to rerun from failed node.
> GH-78: Option to rerun from failed node.
> ----------------------------------------
>
> Key: OOZIE-113
> URL: https://issues.apache.org/jira/browse/OOZIE-113
> Project: Oozie
> Issue Type: Bug
> Reporter: Hadoop QA
>
> Oozie will provide a mechanism for customers to retry a failed action within a workflow. If the action is successful, the dependent action(s) will be executed as well.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (OOZIE-113) GH-78: Option to rerun from failed
node.
Posted by "Hadoop QA (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/OOZIE-113?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13101830#comment-13101830 ]
Hadoop QA commented on OOZIE-113:
---------------------------------
mislam77 remarked:
So Andreas vote is to support in the server.
In that case, from client side, I prefer to use the current oozie rerun option with an extra flag to rerun from the failed node. So there will be less changes from the WS API side too.
I need to close this discussion soon.
Please comment.
> GH-78: Option to rerun from failed node.
> ----------------------------------------
>
> Key: OOZIE-113
> URL: https://issues.apache.org/jira/browse/OOZIE-113
> Project: Oozie
> Issue Type: Bug
> Reporter: Hadoop QA
>
> Oozie will provide a mechanism for customers to retry a failed action within a workflow. If the action is successful, the dependent action(s) will be executed as well.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (OOZIE-113) GH-78: Option to rerun from failed
node.
Posted by "Hadoop QA (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/OOZIE-113?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13101824#comment-13101824 ]
Hadoop QA commented on OOZIE-113:
---------------------------------
mislam77 remarked:
What will be the good option?
* provide a script in client package in dir bin/. Cons: We might end up supporting/enhancing the code in the long run.
* Provide as a Client utility/example and it is user responsibility to maintain.
* as part of our example package.
Comments?
> GH-78: Option to rerun from failed node.
> ----------------------------------------
>
> Key: OOZIE-113
> URL: https://issues.apache.org/jira/browse/OOZIE-113
> Project: Oozie
> Issue Type: Bug
> Reporter: Hadoop QA
>
> Oozie will provide a mechanism for customers to retry a failed action within a workflow. If the action is successful, the dependent action(s) will be executed as well.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira