TestActionScheduler failure stabilization.

[jun aoki <ja...@apache.org>]

Hi Yusaku/Sid,

https://issues.apache.org/jira/browse/AMBARI-7622

Are you OK with the first patch to stabilize it?
I prefer this way so that we can get the trunk build going.
If you guys think it is necessary to fix it comprehensively and no async
execution on unit test, let me know

- jun


On Thu, Oct 9, 2014 at 3:59 PM, jun aoki <ja...@apache.org> wrote:

> Hi Sid,
>
> I know my fix is not comprehensive but at least this should stabilize
> https://builds.apache.org/job/Ambari-trunk-Commit
> with false alert failures Yusaku pointed out.
>
> Could I just keep the fix as is and commit it, and we come back for the
> ideal solution later?
>
>
> On Thu, Oct 9, 2014 at 3:52 PM, Sid Wagle <sw...@hortonworks.com> wrote:
>
>>    This is an automatically generated e-mail. To reply, visit:
>> https://reviews.apache.org/r/26510/
>>
>> On October 9th, 2014, 10:28 p.m. UTC, *Sid Wagle* wrote:
>>
>>
>> ambari-server/src/test/java/org/apache/ambari/server/actionmanager/TestActionScheduler.java
>> <https://reviews.apache.org/r/26510/diff/2/?file=717035#file717035line431> (Diff
>> revision 2)
>>
>> public Object answer(InvocationOnMock invocation) throws Throwable {
>>
>>   431
>>
>>       Thread.sleep(100L);
>>
>> 431
>>
>>       }catch(NullPointerException e){
>>
>>   The NPE seems to be from,
>>
>> line 782: commandParamsStageCache.put(stagePk, commandParams)
>>
>> Seems like we have command params as null so guava cache complains about it. Maybe, "StageUtils.getGson().fromJson(s.getCommandParamsStage(), type)" returns null for empty string, note: we are passing empty string for comandParamsStage.
>>
>> Not sure how this would be resolved after some time?
>>
>> A null check in ActionScheduler might not be a bad idea, your call.
>>
>>  On October 9th, 2014, 10:38 p.m. UTC, *jun aoki* wrote:
>>
>> Sid, yes it is from StageUtils.getGso()... call returning a null.
>>
>> I'm not too sure either how it is resolved but it has been working this way and we solved asychronous behaviour. I'd like to say it is good to go. Please give a Ship it if you are OK.
>>
>>  Could you try putting a Null check after, line 782: commandParamsStageCache.put(stagePk, commandParams) and removing the while condition?
>>
>> If that works there is no need to catch the NPE, I am not convinced NPE is expected behavior.
>>
>>
>> - Sid
>>
>> On October 9th, 2014, 10:05 p.m. UTC, jun aoki wrote:
>>   Review request for Ambari and Yusaku Sako.
>> By jun aoki.
>>
>> *Updated Oct. 9, 2014, 10:05 p.m.*
>>  *Bugs: * AMBARI-7622 <https://issues.apache.org/jira/browse/AMBARI-7622>
>>  *Repository: * ambari
>> Description
>>
>> Tweaked the waiting condition upon ActionScheduler
>>
>>   Diffs
>>
>>    - ambari-server/src/test/java/org/apache/ambari/server/actionmanager/TestActionScheduler.java
>>    (a20f252)
>>
>> View Diff <https://reviews.apache.org/r/26510/diff/>
>>
>
>
>
> --
> -jun
>



-- 
-jun

Re: TestActionScheduler failure stabilization.

[Siddharth Wagle <sw...@hortonworks.com>]

+1, Yusaku.

Why not suppress the UT (@Ignore) and open a new Jira to fix the UT class ?

Regards,
Sid

On Fri, Oct 10, 2014 at 10:46 AM, Yusaku Sako <yu...@hortonworks.com>
wrote:

> Let's first make sure that this UT does not result in false reports.
> We can open another ticket to properly fix the UT itself.
> I'd rather even suppress this particular flaky UT than keep getting false
> reports from Jenkins.
>
> Yusaku
> ​
>

-- 
CONFIDENTIALITY NOTICE
NOTICE: This message is intended for the use of the individual or entity to 
which it is addressed and may contain information that is confidential, 
privileged and exempt from disclosure under applicable law. If the reader 
of this message is not the intended recipient, you are hereby notified that 
any printing, copying, dissemination, distribution, disclosure or 
forwarding of this communication is strictly prohibited. If you have 
received this communication in error, please contact the sender immediately 
and delete it from your system. Thank You.

Re: TestActionScheduler failure stabilization.

[Yusaku Sako <yu...@hortonworks.com>]

Let's first make sure that this UT does not result in false reports.
We can open another ticket to properly fix the UT itself.
I'd rather even suppress this particular flaky UT than keep getting false
reports from Jenkins.

Yusaku
​

-- 
CONFIDENTIALITY NOTICE
NOTICE: This message is intended for the use of the individual or entity to 
which it is addressed and may contain information that is confidential, 
privileged and exempt from disclosure under applicable law. If the reader 
of this message is not the intended recipient, you are hereby notified that 
any printing, copying, dissemination, distribution, disclosure or 
forwarding of this communication is strictly prohibited. If you have 
received this communication in error, please contact the sender immediately 
and delete it from your system. Thank You.