[
https://issues.apache.org/jira/browse/OOZIE-15?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13164977#comment-13164977
]
[email protected] commented on OOZIE-15:
----------------------------------------------------
bq. On 2011-12-05 21:32:37, Santhosh Srinivasan wrote:
bq. > Some code formatting and error message comments. The test coverage has
to be improved.
bq.
bq. Santhosh Srinivasan wrote:
bq. I forgot to add my most important comment. This code fix will break
backward compatibility, i.e., what used to work before will now fail. Are there
any options to mitigate the impact to existing users?
We will educate users through updated documentation that this should be
expected behavior and lack of checking earlier was in fact a bug.
bq. On 2011-12-05 21:32:37, Santhosh Srinivasan wrote:
bq. >
trunk/core/src/test/java/org/apache/oozie/command/coord/TestCoordSubmitXCommand.java,
line 99
bq. > <https://reviews.apache.org/r/3018/diff/1/?file=61960#file61960line99>
bq. >
bq. > The test case covers both input-events and output-events. Since
input-events occurs first, the test case is not testing output-events. Can you
add a separate test case for output-events and also test cases where
input-events are specified with separate instance tags and similarly for
output-events if its applicable.
success case of multiple <instance> tags does not apply to output-event
dataset. It can only have single tag and single instance within it. added
success and failure cases for the input-event part
bq. On 2011-12-05 21:32:37, Santhosh Srinivasan wrote:
bq. >
trunk/core/src/test/java/org/apache/oozie/command/coord/TestCoordSubmitXCommand.java,
line 128
bq. > <https://reviews.apache.org/r/3018/diff/1/?file=61960#file61960line128>
bq. >
bq. > Can you also check for the error code?
done
bq. On 2011-12-05 21:32:37, Santhosh Srinivasan wrote:
bq. >
trunk/core/src/main/java/org/apache/oozie/command/coord/CoordSubmitXCommand.java,
line 314
bq. > <https://reviews.apache.org/r/3018/diff/1/?file=61959#file61959line314>
bq. >
bq. > Change 1 to one. Is there a corrective action similar to that of the
input-events?
corrective action for output-event is not the same. Nevertheless, added
information informing about it.
- Mona
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/3018/#review3636
-----------------------------------------------------------
On 2011-12-05 20:15:04, Mona Chitnis wrote:
bq.
bq. -----------------------------------------------------------
bq. This is an automatically generated e-mail. To reply, visit:
bq. https://reviews.apache.org/r/3018/
bq. -----------------------------------------------------------
bq.
bq. (Updated 2011-12-05 20:15:04)
bq.
bq.
bq. Review request for oozie, Mohammad Islam and Angelo K. Huang.
bq.
bq.
bq. Summary
bq. -------
bq.
bq. For the coordinator input events, if there are multiple dates, it uses the
first date to resolve the data
bq. instance, instead of giving warning or error.
bq. For example:
bq. <data-in name="din1" dataset="ds1">
bq.
<instance>2010-06-11T00:30Z,2010-06-11T00:31Z,2010-06-11T00:32Z</instance>
bq. </data-in>
bq.
bq. Similarly for output events.
bq.
bq. This patch will give error for the above case and protect against
submission of job that defaults to the first date without the knowledge of user.
bq.
bq.
bq. This addresses bug OOZIE-15.
bq. https://issues.apache.org/jira/browse/OOZIE-15
bq.
bq.
bq. Diffs
bq. -----
bq.
bq.
trunk/core/src/main/java/org/apache/oozie/command/coord/CoordSubmitXCommand.java
1209756
bq.
trunk/core/src/test/java/org/apache/oozie/command/coord/TestCoordSubmitXCommand.java
1209756
bq.
bq. Diff: https://reviews.apache.org/r/3018/diff
bq.
bq.
bq. Testing
bq. -------
bq.
bq. yes
bq.
bq.
bq. Thanks,
bq.
bq. Mona
bq.
bq.
> Coordinator input event instance should limit 1 instance
> --------------------------------------------------------
>
> Key: OOZIE-15
> URL: https://issues.apache.org/jira/browse/OOZIE-15
> Project: Oozie
> Issue Type: Bug
> Reporter: Mona Chitnis
> Priority: Minor
> Labels: event, oozie
> Attachments: OOZIE-15.patch
>
> Original Estimate: 168h
> Remaining Estimate: 168h
>
> for the input events, if there are multiple dates, it uses the first date to
> resolve the data
> instance, instead of giving warning or error. for example:
> <data-in name="din1" dataset="ds1">
> <instance>2010-06-11T00:30Z,2010-06-11T00:31Z,2010-06-11T00:32Z</instance>
> </data-in>
> Missing Dependencies :
> hdfs://gsbl90390.blue.ygrid.yahoo.com:8020/user/mchiang/coord-input/2010/06/11/00/30/_SUCCESS
> similarly, there is no warning or error when applying dateOffset with count >
> 1.
> <data-in name="din1" dataset="ds1">
> <instance>${coord:dateOffset(baseDate,30,'MINUTE',5)}</instance>
> </data-in>
> 1. dateOffset no longer has 4th argument, count.
> 2. however, there is no error or warning when in coordinator.xml:
> <input-events>
> <data-in name="din1" dataset="ds1">
> <instance>${coord:current(-2)},${coord:current(-1)}</instance>
> </data-in>
> </input-events>
> it resolved to use ${coord:current(-2)} only.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
