Re: TimeoutLifecycleManager not doing anything

2018-04-06 Thread Paul Brenner 
These are Spark interpreters that aren’t getting killed. Does 
TimeoutLifecycleManager kill spark interpreters when the last submitted job 
finished was more than zeppelin.interpreter.lifecyclemanager.timeout.threshold 
milliseconds ago (+/- 
zeppelin.interpreter.lifecyclemanager.timeout.checkinterval)? Or is it based on 
something other than when the last job finished (leaving a notebook open in a 
browser tab maybe)? What happens if you have an interpreter set to be 
instantiated per note and isolated. If I run that interpreter on two different 
notebooks (creating two different spark jobs) and leave one idle but use the 
other, should the idle one still get killed? That would be ideal.

I’m not sure what is causing these interpreters not to get killed so I don’t 
know how to explain how to reproduce the problem. 

( http://www.placeiq.com/ ) ( http://www.placeiq.com/ ) ( 
http://www.placeiq.com/ ) *Paul Brenner* ( https://twitter.com/placeiq ) ( 
https://twitter.com/placeiq ) ( https://twitter.com/placeiq ) ( 
https://www.facebook.com/PlaceIQ ) ( https://www.facebook.com/PlaceIQ ) ( 
https://www.linkedin.com/company/placeiq ) ( 
https://www.linkedin.com/company/placeiq ) SR. DATA SCIENTIST (217) 390-3033 

( 
http://www.placeiq.com/2015/05/26/placeiq-named-winner-of-prestigious-2015-oracle-data-cloud-activate-award/
 ) ( 
http://placeiq.com/2015/12/18/accuracy-vs-precision-in-location-data-mma-webinar/
 ) ( 
http://placeiq.com/2015/12/18/accuracy-vs-precision-in-location-data-mma-webinar/
 ) ( 
http://placeiq.com/2015/12/18/accuracy-vs-precision-in-location-data-mma-webinar/
 ) ( 
http://placeiq.com/2015/12/18/accuracy-vs-precision-in-location-data-mma-webinar/
 ) ( 
http://placeiq.com/2016/03/08/measuring-addressable-tv-campaigns-is-now-possible/
 ) ( 
http://placeiq.com/2016/04/13/placeiq-joins-the-network-advertising-initiative-nai-as-100th-member/
 ) ( 
http://placeiq.com/2016/04/13/placeiq-joins-the-network-advertising-initiative-nai-as-100th-member/
 ) ( 
http://placeiq.com/2016/04/13/placeiq-joins-the-network-advertising-initiative-nai-as-100th-member/
 ) ( 
http://placeiq.com/2016/04/13/placeiq-joins-the-network-advertising-initiative-nai-as-100th-member/
 ) ( 
http://placeiq.com/2016/04/13/placeiq-joins-the-network-advertising-initiative-nai-as-100th-member/
 ) ( 
http://pages.placeiq.com/Location-Data-Accuracy-Whitepaper-Download.html?utm_source=Signature&utm_medium=Email&utm_campaign=AccuracyWP
 ) ( 
http://placeiq.com/2016/08/03/placeiq-bolsters-location-intelligence-platform-with-mastercard-insights/
 ) ( 
http://placeiq.com/2016/10/26/the-making-of-a-location-data-industry-milestone/ 
) ( 
http://placeiq.com/2016/12/07/placeiq-introduces-landmark-a-groundbreaking-offering-that-delivers-access-to-the-highest-quality-location-data-for-insights-that-fuel-limitless-business-decisions/
 ) ( 
http://placeiq.com/2016/12/07/placeiq-introduces-landmark-a-groundbreaking-offering-that-delivers-access-to-the-highest-quality-location-data-for-insights-that-fuel-limitless-business-decisions/
 ) ( 
https://www.placeiq.com/2017/05/placeiqs-landmark-powers-location-based-insight-innovation-for-ansible-gstv-havas-media-the-media-kitchen-and-more/
 ) ( http://pages.placeiq.com/CES2018_MeetingRequest.html ) PlaceIQ:CES 2018 ( 
http://pages.placeiq.com/2017-Integrated-Marketing-Whitepaper_LP_Download.html )

On Fri, Apr 06, 2018 at 1:28 AM Jeff Zhang < Jeff Zhang ( Jeff Zhang 
 ) > wrote:

> 
> 
> Hi Paul,
> 
> 
> What interpreter do you find that is not killed ? Unfortunately there's no
> log for the lasttimeusing (I should add that for debugging).
> 
> 
> And let me know if it is easy to reproduce.
> 
> 
> 
> 
> 
> Paul Brenner < pbren...@placeiq.com >于2018年4月5日周四 下午11:21写道：
> 
> 
>> 
>> Wow, looks like I was wrong. TimeoutLifecycleManager is definitely killing
>> some idle interpreters, just not all idle interpreters. I have no idea
>> what could be causing the difference. It looks like the manager is pulling
>> some value for lasttimeusing here 
>> https://github.com/apache/zeppelin/blob/86f387e940eacb6dac7531a4ed2cadadaa381bbe/zeppelin-zengine/src/main/java/org/apache/zeppelin/interpreter/lifecycle/TimeoutLifecycleManager.java#L47
>> (
>> https://share.polymail.io/v1/z/b/NWFjNjNjMDc0ZWFi/6Zff_ZkywYC3DLCbIS8X_WEmw1V5BaikknznZ6LCyccNZKg9pLOcuXn-dfsZesa53zGcAv6M3lEqmXb3IGo21ejezflaXMGokN93jD1gpMd8ga0G2qZtbGCqRC3iyCA3sP6dnGVwSuyIpJOz_ZIrnMDJqQy8jv77XulymBRkRw6KViRLB-skqfm4plocl69EZ4kINl1cymifTZA27_f1MdHiCBWqXh5OpuesbRHHTMfR71SBpq8dWD3OF4Irlfdk5SFEKguu8BG4JXh_LT9SRBYaGRIJm2uP6s_THac9GNV_ljM_qgKTIOtCQfPRJALlgKpbSyD8evmeQ0WWxdPPd5qRY2_-_nhNdTlg6SzqRgNBdE9gHQvlpHFwDerhUZiYHC7TIL90
>> )
>> 
>> 
>> 
>> but I assume there is no way I can figure out what value that is for the
>> idle interpreters I’m seeing that haven’t been killed. Is there anything
>> else I can look into or should I just accept that this feature is going to
>> be partly broken for us going forward?
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>> ( http://www.p

Re: TimeoutLifecycleManager not doing anything

2018-04-05 Thread Jeff Zhang 
Hi Paul,

What interpreter do you find that is not killed ? Unfortunately there's no
log for the lasttimeusing (I should add that for debugging).

And let me know if it is easy to reproduce.



Paul Brenner 于2018年4月5日周四 下午11:21写道：

> Wow, looks like I was wrong. TimeoutLifecycleManager is definitely killing
> some idle interpreters, just not all idle interpreters. I have no idea what
> could be causing the difference. It looks like the manager is pulling some
> value for lasttimeusing here
> https://github.com/apache/zeppelin/blob/86f387e940eacb6dac7531a4ed2cadadaa381bbe/zeppelin-zengine/src/main/java/org/apache/zeppelin/interpreter/lifecycle/TimeoutLifecycleManager.java#L47
> 
>
> but I assume there is no way I can figure out what value that is for the
> idle interpreters I’m seeing that haven’t been killed. Is there anything
> else I can look into or should I just accept that this feature is going to
> be partly broken for us going forward?
>
>  
>  *Paul Brenner* 
>  
>  
> 
> 
> SR. DATA SCIENTIST
> *(217) 390-3033 <(217)%20390-3033> *
>
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> [image: PlaceIQ:CES
> 2018]
> 
>
> On Thu, Apr 05, 2018 at 10:43 AM Jeff Zhang  > wrote:
>
>>
>> Try to look for logs like following in zeppelin server log file. Make
>> sure you are using TimeoutLifecycleManager instead of
>> NullLifecycleManager
>>
>> INFO [2018-04-04 18:12:26,209] ({main}
>> InterpreterSettingManager.java[]:170) - Using LifecycleManager:
>> org.apache.zeppelin.interpreter.lifecycle.NullLifecycleManager
>>
>> Paul Brenner 于2018年4月5日周四 下午10:32写道：
>>
>>> We finally got our team up and running on a version of Zeppelin 0.8
>>> built from source. However, we can’t seem to get TimeoutLifecycleManager
>>> manager to kill idle interpreters. When I look at the config page I see:
>>>
>>> zeppelin.interpreter.lifecyclemanager.class
>>> org.apache.zeppelin.interpreter.lifecycle.TimeoutLifecycleManager
>>> zeppelin.interpreter.lifecyclemanager.timeout.checkinterval
>>> 6
>>> zeppelin.interpreter.lifecyclemanager.timeout.threshold
>>> 360
>>>
>>> So it seems like everything should be configured correctly. However, if
>>> we leave interpreters idling overnight they are still

Re: TimeoutLifecycleManager not doing anything

2018-04-05 Thread Paul Brenner 
Wow, looks like I was wrong. TimeoutLifecycleManager is definitely killing some 
idle interpreters, just not all idle interpreters. I have no idea what could be 
causing the difference. It looks like the manager is pulling some value for 
lasttimeusing here 
https://github.com/apache/zeppelin/blob/86f387e940eacb6dac7531a4ed2cadadaa381bbe/zeppelin-zengine/src/main/java/org/apache/zeppelin/interpreter/lifecycle/TimeoutLifecycleManager.java#L47
 ( 
https://share.polymail.io/v1/z/b/NWFjNjNjMDc0ZWFi/6Zff_ZkywYC3DLCbIS8X_WEmw1V5BaikknznZ6LCyccNZKg9pLOcuXn-dfsZesa53zGcAv6M3lEqmXb3IGo21ejezflaXMGokN93gyRjsNBJu68by6ptK23rSFzr3zwi7avA1ANxSvCNtYf_vZhtidjdqEq6zumxT-l2lxVvRxaDSTYBU7QstaPjrAEHn_xDN9MJNwULzTjNH5M07vKkZYW1WBD_WklE9OSqaVrcSISZu1qKra0dWCPZG4An1Kls42BOIgyl8AirMDY9IyRbRB0LDhBOkTaSoN7TGag8E9VsnTcqvR6IK7BOXajYPxTyhqddVWq_R_aWRVOA3fqJRZaff2Pn73lFfT1k9T2oSS9Wew1LTE78x00hKWkh_1r9w3i_sM4V61_eQlhLDxnO
 )

but I assume there is no way I can figure out what value that is for the idle 
interpreters I’m seeing that haven’t been killed. Is there anything else I can 
look into or should I just accept that this feature is going to be partly 
broken for us going forward?

( http://www.placeiq.com/ ) ( http://www.placeiq.com/ ) ( 
http://www.placeiq.com/ ) *Paul Brenner* ( https://twitter.com/placeiq ) ( 
https://twitter.com/placeiq ) ( https://twitter.com/placeiq ) ( 
https://www.facebook.com/PlaceIQ ) ( https://www.facebook.com/PlaceIQ ) ( 
https://www.linkedin.com/company/placeiq ) ( 
https://www.linkedin.com/company/placeiq ) SR. DATA SCIENTIST (217) 390-3033 

( 
http://www.placeiq.com/2015/05/26/placeiq-named-winner-of-prestigious-2015-oracle-data-cloud-activate-award/
 ) ( 
http://placeiq.com/2015/12/18/accuracy-vs-precision-in-location-data-mma-webinar/
 ) ( 
http://placeiq.com/2015/12/18/accuracy-vs-precision-in-location-data-mma-webinar/
 ) ( 
http://placeiq.com/2015/12/18/accuracy-vs-precision-in-location-data-mma-webinar/
 ) ( 
http://placeiq.com/2015/12/18/accuracy-vs-precision-in-location-data-mma-webinar/
 ) ( 
http://placeiq.com/2016/03/08/measuring-addressable-tv-campaigns-is-now-possible/
 ) ( 
http://placeiq.com/2016/04/13/placeiq-joins-the-network-advertising-initiative-nai-as-100th-member/
 ) ( 
http://placeiq.com/2016/04/13/placeiq-joins-the-network-advertising-initiative-nai-as-100th-member/
 ) ( 
http://placeiq.com/2016/04/13/placeiq-joins-the-network-advertising-initiative-nai-as-100th-member/
 ) ( 
http://placeiq.com/2016/04/13/placeiq-joins-the-network-advertising-initiative-nai-as-100th-member/
 ) ( 
http://placeiq.com/2016/04/13/placeiq-joins-the-network-advertising-initiative-nai-as-100th-member/
 ) ( 
http://pages.placeiq.com/Location-Data-Accuracy-Whitepaper-Download.html?utm_source=Signature&utm_medium=Email&utm_campaign=AccuracyWP
 ) ( 
http://placeiq.com/2016/08/03/placeiq-bolsters-location-intelligence-platform-with-mastercard-insights/
 ) ( 
http://placeiq.com/2016/10/26/the-making-of-a-location-data-industry-milestone/ 
) ( 
http://placeiq.com/2016/12/07/placeiq-introduces-landmark-a-groundbreaking-offering-that-delivers-access-to-the-highest-quality-location-data-for-insights-that-fuel-limitless-business-decisions/
 ) ( 
http://placeiq.com/2016/12/07/placeiq-introduces-landmark-a-groundbreaking-offering-that-delivers-access-to-the-highest-quality-location-data-for-insights-that-fuel-limitless-business-decisions/
 ) ( 
https://www.placeiq.com/2017/05/placeiqs-landmark-powers-location-based-insight-innovation-for-ansible-gstv-havas-media-the-media-kitchen-and-more/
 ) ( http://pages.placeiq.com/CES2018_MeetingRequest.html ) PlaceIQ:CES 2018 ( 
http://pages.placeiq.com/2017-Integrated-Marketing-Whitepaper_LP_Download.html )

On Thu, Apr 05, 2018 at 10:43 AM Jeff Zhang < Jeff Zhang ( Jeff Zhang 
 ) > wrote:

> 
> 
> 
> Try to look for logs like following in zeppelin server log file. Make sure
> you are using TimeoutLifecycleManager instead of NullLifecycleManager
> 
> 
> INFO [2018-04-04 18:12:26,209] ({main}
> InterpreterSettingManager.java[]:170) - Using LifecycleManager:
> org.apache.zeppelin.interpreter.lifecycle.NullLifecycleManager
> 
> 
> Paul Brenner < pbren...@placeiq.com >于2018年4月5日周四 下午10:32写道：
> 
> 
>> 
>> We finally got our team up and running on a version of Zeppelin 0.8 built
>> from source. However, we can’t seem to get TimeoutLifecycleManager manager
>> to kill idle interpreters. When I look at the config page I see:
>> 
>> 
>> 
>> 
>> zeppelin.interpreter.lifecyclemanager.class
>> org.apache.zeppelin.interpreter.lifecycle.TimeoutLifecycleManager
>> zeppelin.interpreter.lifecyclemanager.timeout.checkinterval
>> 6
>> zeppelin.interpreter.lifecyclemanager.timeout.threshold
>> 360
>> 
>> 
>> 
>> 
>> 
>> So it seems like everything should be configured correctly. However, if we
>> leave interpreters idling overnight they are still there idling the next
>> morning.
>> 
>> 
>> Is there any way to diagnose what is going wrong? Can anyone confirm that
>> they have had succe

Re: TimeoutLifecycleManager not doing anything

2018-04-05 Thread Michael Bullock 
Hi Jeff,
If I am using zeppelin in an AWS environment what type of metric could I
gather using cloudformation scripts?

On Thu, Apr 5, 2018 at 10:43 AM, Jeff Zhang  wrote:

>
> Try to look for logs like following in zeppelin server log file. Make sure
> you are using TimeoutLifecycleManager instead of NullLifecycleManager
>
> INFO [2018-04-04 18:12:26,209] ({main} 
> InterpreterSettingManager.java[]:170)
> - Using LifecycleManager: org.apache.zeppelin.interpreter.lifecycle.
> NullLifecycleManager
>
> Paul Brenner 于2018年4月5日周四 下午10:32写道：
>
>> We finally got our team up and running on a version of Zeppelin 0.8 built
>> from source. However, we can’t seem to get TimeoutLifecycleManager
>> manager to kill idle interpreters. When I look at the config page I see:
>>
>> zeppelin.interpreter.lifecyclemanager.class
>> org.apache.zeppelin.interpreter.lifecycle.TimeoutLifecycleManager
>> zeppelin.interpreter.lifecyclemanager.timeout.checkinterval
>> 6
>> zeppelin.interpreter.lifecyclemanager.timeout.threshold
>> 360
>>
>> So it seems like everything should be configured correctly. However, if
>> we leave interpreters idling overnight they are still there idling the next
>> morning.
>>
>> Is there any way to diagnose what is going wrong? Can anyone confirm that
>> they have had success with TimeoutLifecycleManager? Should there
>> be anything in any of the log files?
>>  
>>  *Paul Brenner* 
>>  
>>  
>> 
>> 
>> SR. DATA SCIENTIST
>> *(217) 390-3033 <(217)%20390-3033> *
>>
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>> [image:
>> PlaceIQ:CES 2018]
>> 
>>
>


-- 
*Michael L. Bullock II*
*Email: michaelbullo...@gmail.com *
*Mobile: (240)504-7520*

Re: TimeoutLifecycleManager not doing anything

2018-04-05 Thread Jeff Zhang 
Try to look for logs like following in zeppelin server log file. Make sure
you are using TimeoutLifecycleManager instead of NullLifecycleManager

INFO [2018-04-04 18:12:26,209] ({main}
InterpreterSettingManager.java[]:170) - Using LifecycleManager:
org.apache.zeppelin.interpreter.lifecycle.NullLifecycleManager

Paul Brenner 于2018年4月5日周四 下午10:32写道：

> We finally got our team up and running on a version of Zeppelin 0.8 built
> from source. However, we can’t seem to get TimeoutLifecycleManager
> manager to kill idle interpreters. When I look at the config page I see:
>
> zeppelin.interpreter.lifecyclemanager.class
> org.apache.zeppelin.interpreter.lifecycle.TimeoutLifecycleManager
> zeppelin.interpreter.lifecyclemanager.timeout.checkinterval
> 6
> zeppelin.interpreter.lifecyclemanager.timeout.threshold
> 360
>
> So it seems like everything should be configured correctly. However, if we
> leave interpreters idling overnight they are still there idling the next
> morning.
>
> Is there any way to diagnose what is going wrong? Can anyone confirm that
> they have had success with TimeoutLifecycleManager? Should there
> be anything in any of the log files?
>  
>  *Paul Brenner* 
>  
>  
> 
> 
> SR. DATA SCIENTIST
> *(217) 390-3033 <(217)%20390-3033> *
>
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> [image: PlaceIQ:CES
> 2018]
> 
>