> On Nov. 19, 2013, 1:01 a.m., Rohini Palaniswamy wrote:
> > trunk/core/src/main/java/org/apache/oozie/command/wf/ActionEndXCommand.java,
> >  line 149
> > <https://reviews.apache.org/r/15225/diff/2/?file=381603#file381603line149>
> >
> >     Remove the variable and do check for control node directly

renamed it to execSynchronous to be clearer. Need the variable to be set in 
ERROR condition too.


> On Nov. 19, 2013, 1:01 a.m., Rohini Palaniswamy wrote:
> > trunk/core/src/test/java/org/apache/oozie/action/TestActionFailover.java, 
> > line 126
> > <https://reviews.apache.org/r/15225/diff/2/?file=381609#file381609line126>
> >
> >     Why does this change happen?

because now on invoking ActionStartX.call(), first action - start control node 
- asynchronously calls ActionEndX. This invokes ControlNodeActionExecutor.end() 
:
context.setEndData(WorkflowAction.Status.OK, 
getActionSignal(WorkflowAction.Status.OK));

thereby setting the status to OK, instead of the intermediate DONE before


> On Nov. 19, 2013, 1:01 a.m., Rohini Palaniswamy wrote:
> > trunk/core/src/test/java/org/apache/oozie/command/wf/TestSignalXCommand.java,
> >  lines 197-198
> > <https://reviews.apache.org/r/15225/diff/2/?file=381610#file381610line197>
> >
> >     Can you explain this change?

Since SignalX for first action - start control node - will call ActionStartX 
asynchronously, :start: will go from PREP to OK in this case. Hence the minor 
change to remove :start: from list of PREP actions


> On Nov. 19, 2013, 1:01 a.m., Rohini Palaniswamy wrote:
> > trunk/core/src/test/java/org/apache/oozie/event/TestEventGeneration.java, 
> > line 143
> > <https://reviews.apache.org/r/15225/diff/2/?file=381611#file381611line143>
> >
> >     Can you explain this change?

The change is not have a no-op wf app. Added a 1-node (not just control ones) 
wf app here too. Otherwise we cannot test the intermediate steps such as wf 
going to RUnning, since it directly proceeds to end and SUCCEEDED.


> On Nov. 19, 2013, 1:01 a.m., Rohini Palaniswamy wrote:
> > trunk/core/src/test/java/org/apache/oozie/test/XDataTestCase.java, line 1168
> > <https://reviews.apache.org/r/15225/diff/2/?file=381614#file381614line1168>
> >
> >     Can you explain this change?

calling the method 
Services.get().get(UUIDService.class).generateId(ApplicationType.WORKFLOW), 
produces a new incremental wf id each time. In this test, we need to retain the 
same id, which is also that of the job's wfInstance. The other tests calling 
this method yet are perhaps not affected by this small bug in XDataTestCase 
method.


- Mona


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/15225/#review29084
-----------------------------------------------------------


On Nov. 9, 2013, 4:20 a.m., Mona Chitnis wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/15225/
> -----------------------------------------------------------
> 
> (Updated Nov. 9, 2013, 4:20 a.m.)
> 
> 
> Review request for oozie.
> 
> 
> Bugs: OOZIE-1581
>     https://issues.apache.org/jira/browse/OOZIE-1581
> 
> 
> Repository: oozie
> 
> 
> Description
> -------
> 
> Patch for review updated from the attachments on JIRA.
> 1. Included change to call coordinator action Start synchronously from Ready 
> command.
> 2. Made Decision Action Executor extend from Control Action Executor. That 
> way the synchronous commands can be executed for all such 'connector' actions.
> 3. Fixed failing unit tests.
> 4. TestSLAEventGeneration unit test is failing - WIP.
> 
> 
> Diffs
> -----
> 
>   
> trunk/core/src/main/java/org/apache/oozie/command/coord/CoordActionReadyXCommand.java
>  1538746 
>   
> trunk/core/src/main/java/org/apache/oozie/command/coord/CoordActionUpdateXCommand.java
>  1538746 
>   trunk/core/src/main/java/org/apache/oozie/command/wf/ActionEndXCommand.java 
> 1538746 
>   
> trunk/core/src/main/java/org/apache/oozie/command/wf/ActionStartXCommand.java 
> 1538746 
>   trunk/core/src/main/java/org/apache/oozie/command/wf/SignalXCommand.java 
> 1538746 
>   
> trunk/core/src/main/java/org/apache/oozie/executor/jpa/BatchQueryExecutor.java
>  1538746 
>   trunk/core/src/main/java/org/apache/oozie/service/WorkflowStoreService.java 
> 1538746 
>   
> trunk/core/src/test/java/org/apache/oozie/TestCoordinatorEngineStreamLog.java 
> 1538746 
>   trunk/core/src/test/java/org/apache/oozie/action/TestActionFailover.java 
> 1538746 
>   
> trunk/core/src/test/java/org/apache/oozie/command/wf/TestSignalXCommand.java 
> 1538746 
>   trunk/core/src/test/java/org/apache/oozie/event/TestEventGeneration.java 
> 1538746 
>   trunk/core/src/test/java/org/apache/oozie/service/TestRecoveryService.java 
> 1538746 
>   trunk/core/src/test/java/org/apache/oozie/sla/TestSLAEventGeneration.java 
> 1538746 
>   trunk/core/src/test/java/org/apache/oozie/test/XDataTestCase.java 1538746 
>   trunk/core/src/test/resources/wf-fork.xml 1538746 
> 
> Diff: https://reviews.apache.org/r/15225/diff/
> 
> 
> Testing
> -------
> 
> unit tests fixed. stress testing performed to validate the fast start case.
> 
> 
> Thanks,
> 
> Mona Chitnis
> 
>

Reply via email to