[jira] [Commented] (YARN-3437) convert load test driver to timeline service v.2

2015-04-23 Thread Sangjin Lee (JIRA)

[ 
https://issues.apache.org/jira/browse/YARN-3437?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14509237#comment-14509237
 ] 

Sangjin Lee commented on YARN-3437:
---

Thanks Junping. I initially went with ConcurrentHashMap when I first created 
this as that is my preference as well. But it was really preventing multiple 
threads from starting their collector (should that situation arise) that made 
ConcurrentHashMap not an option. Again, if we want both, we would need to look 
at the LoadingCache. But since this is really a low contention situation, it 
would be an overkill. The chances of this code running into a lock contention 
should be low.

 convert load test driver to timeline service v.2
 

 Key: YARN-3437
 URL: https://issues.apache.org/jira/browse/YARN-3437
 Project: Hadoop YARN
  Issue Type: Sub-task
  Components: timelineserver
Reporter: Sangjin Lee
Assignee: Sangjin Lee
 Attachments: YARN-3437.001.patch, YARN-3437.002.patch, 
 YARN-3437.003.patch, YARN-3437.004.patch


 This subtask covers the work for converting the proposed patch for the load 
 test driver (YARN-2556) to work with the timeline service v.2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (YARN-3437) convert load test driver to timeline service v.2

2015-04-23 Thread Zhijie Shen (JIRA)

[ 
https://issues.apache.org/jira/browse/YARN-3437?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14509717#comment-14509717
 ] 

Zhijie Shen commented on YARN-3437:
---

Sorry for the late comments. This patch has half MR code and half YARN code. 
It's not good to commit it as one patch. I have one thought of managing the 
commits:

1. The YARN code is nearly duplicate with YARN-3390. As YARN-3390 is almost 
ready, we can get that patch in first.

2. Move this jira to MR project, only retain the MR code in the patch and do 
some minor rebase according to YARN-3390.

3. TimelineServicePerformanceTest is in different package and has the different 
name. Hopefully it won't conflict with YARN-2556. So once YARN-2556 gets 
committed, we just need to refactor TimelineServicePerformanceTest to reuse 
YARN-2556 code. BTW, can we put TimelineServicePerformanceTest into the same 
package of TimelineServicePerformance in YARN-2556, and rename it to 
TimelineServicePerformanceTestv2?

How do you think about the plan for the commits?

W.R.T to the patch, I'm a bit concerned that the write which contains one event 
per entity is not so typical to represent real use case. And configuration and 
metrics are even not covered. Is it more realistic to write an entity with 10 
events and 10 metrics, which have 100 points in the time series?

And one nit in the patch: {{entity.setEntityType(TEZ_DAG_ID);}}. How about 
not mentioning TEZ in the MR code?

 convert load test driver to timeline service v.2
 

 Key: YARN-3437
 URL: https://issues.apache.org/jira/browse/YARN-3437
 Project: Hadoop YARN
  Issue Type: Sub-task
  Components: timelineserver
Reporter: Sangjin Lee
Assignee: Sangjin Lee
 Attachments: YARN-3437.001.patch, YARN-3437.002.patch, 
 YARN-3437.003.patch, YARN-3437.004.patch


 This subtask covers the work for converting the proposed patch for the load 
 test driver (YARN-2556) to work with the timeline service v.2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (YARN-3437) convert load test driver to timeline service v.2

2015-04-23 Thread Sangjin Lee (JIRA)

[ 
https://issues.apache.org/jira/browse/YARN-3437?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14509905#comment-14509905
 ] 

Sangjin Lee commented on YARN-3437:
---

Well it's not entirely true. It seems I still need to change 
TimelineCollector.getTimelineEntityContext() from protected to public. But 
creating another YARN JIRA just to make those several lines of changes seems 
too much. Thoughts folks?

 convert load test driver to timeline service v.2
 

 Key: YARN-3437
 URL: https://issues.apache.org/jira/browse/YARN-3437
 Project: Hadoop YARN
  Issue Type: Sub-task
  Components: timelineserver
Reporter: Sangjin Lee
Assignee: Sangjin Lee
 Attachments: YARN-3437.001.patch, YARN-3437.002.patch, 
 YARN-3437.003.patch, YARN-3437.004.patch


 This subtask covers the work for converting the proposed patch for the load 
 test driver (YARN-2556) to work with the timeline service v.2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (YARN-3437) convert load test driver to timeline service v.2

2015-04-23 Thread Zhijie Shen (JIRA)

[ 
https://issues.apache.org/jira/browse/YARN-3437?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14509932#comment-14509932
 ] 

Zhijie Shen commented on YARN-3437:
---

bq.  How's that sound?

It's also good to me.

bq. We would use that one for more realistic load whereas we could keep this 
mode as a simpler test. Thoughts?

It's okay to make it a simpler case, but could we at least cover one config, 
and one metric, hence we can verify the db that storing this info also works?

bq.  But creating another YARN JIRA just to make those several lines of changes 
seems too much

A couple of lines change in YARN for MR patch is okay.

 convert load test driver to timeline service v.2
 

 Key: YARN-3437
 URL: https://issues.apache.org/jira/browse/YARN-3437
 Project: Hadoop YARN
  Issue Type: Sub-task
  Components: timelineserver
Reporter: Sangjin Lee
Assignee: Sangjin Lee
 Attachments: YARN-3437.001.patch, YARN-3437.002.patch, 
 YARN-3437.003.patch, YARN-3437.004.patch


 This subtask covers the work for converting the proposed patch for the load 
 test driver (YARN-2556) to work with the timeline service v.2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (YARN-3437) convert load test driver to timeline service v.2

2015-04-23 Thread Zhijie Shen (JIRA)

[ 
https://issues.apache.org/jira/browse/YARN-3437?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14509934#comment-14509934
 ] 

Zhijie Shen commented on YARN-3437:
---

Oh, previously I said TimelineServicePerformanceTestv2, but actually I meant 
TimelineServicePerformanceV2. Just a minor suggestion, and it's up to you to 
find the suitable class name.

 convert load test driver to timeline service v.2
 

 Key: YARN-3437
 URL: https://issues.apache.org/jira/browse/YARN-3437
 Project: Hadoop YARN
  Issue Type: Sub-task
  Components: timelineserver
Reporter: Sangjin Lee
Assignee: Sangjin Lee
 Attachments: YARN-3437.001.patch, YARN-3437.002.patch, 
 YARN-3437.003.patch, YARN-3437.004.patch


 This subtask covers the work for converting the proposed patch for the load 
 test driver (YARN-2556) to work with the timeline service v.2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (YARN-3437) convert load test driver to timeline service v.2

2015-04-23 Thread Sangjin Lee (JIRA)

[ 
https://issues.apache.org/jira/browse/YARN-3437?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14509862#comment-14509862
 ] 

Sangjin Lee commented on YARN-3437:
---

Thanks for your comments [~zjshen].

{quote}
1. The YARN code is nearly duplicate with YARN-3390. As YARN-3390 is almost 
ready, we can get that patch in first.
2. Move this jira to MR project, only retain the MR code in the patch and do 
some minor rebase according to YARN-3390.
{quote}

Let's do this. While I was working on YARN-3438, I'm realizing that for the 
performance tests it is probably OK to use the TimelineCollectors directly and 
bypass the TimelineCollectorManager altogether. If we do that, then this could 
become purely a MR patch. I'll update this patch to remove the use of 
TimelineCollectorManager and move this JIRA to MAPREDUCE. How's that sound?

{quote}
3. TimelineServicePerformanceTest is in different package and has the different 
name. Hopefully it won't conflict with YARN-2556. So once YARN-2556 gets 
committed, we just need to refactor TimelineServicePerformanceTest to reuse 
YARN-2556 code. BTW, can we put TimelineServicePerformanceTest into the same 
package of TimelineServicePerformance in YARN-2556, and rename it to 
TimelineServicePerformanceTestv2?
{quote}
That's fine. I'll move it back to the same package.

{quote}
W.R.T to the patch, I'm a bit concerned that the write which contains one event 
per entity is not so typical to represent real use case. And configuration and 
metrics are even not covered. Is it more realistic to write an entity with 10 
events and 10 metrics, which have 100 points in the time series?
And one nit in the patch: entity.setEntityType(TEZ_DAG_ID);. How about not 
mentioning TEZ in the MR code?
{quote}
Note that this is adding simple entity writes. The more realistic part of the 
test is coming in YARN-3438 (I'm nearly finished with that), and it will have 
multiple levels of entities as well as metrics and configuration. We would use 
that one for more realistic load whereas we could keep this mode as a simpler 
test. Thoughts?

I'll change the name of the entity to be something else.

 convert load test driver to timeline service v.2
 

 Key: YARN-3437
 URL: https://issues.apache.org/jira/browse/YARN-3437
 Project: Hadoop YARN
  Issue Type: Sub-task
  Components: timelineserver
Reporter: Sangjin Lee
Assignee: Sangjin Lee
 Attachments: YARN-3437.001.patch, YARN-3437.002.patch, 
 YARN-3437.003.patch, YARN-3437.004.patch


 This subtask covers the work for converting the proposed patch for the load 
 test driver (YARN-2556) to work with the timeline service v.2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (YARN-3437) convert load test driver to timeline service v.2

2015-04-23 Thread Junping Du (JIRA)

[ 
https://issues.apache.org/jira/browse/YARN-3437?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14508742#comment-14508742
 ] 

Junping Du commented on YARN-3437:
--

bq. Also, when you said may be move out of collector.start(), did you mean 
moving the collector.start() call outside the synchronization block? If so, I'd 
be hesitant to do that. We just had a discussion on this in another JIRA.
I see. We don't want the other concurrent thread get a non-start collector. We 
can improve this later (in some other JIRA), e.g. attach an additional monitor 
object to specific collector (rather than sync on all collectors), and do 
something like: startIfNotStarted(). 

Another NIT is to replace Collections.synchronizedMap with ConcurrentHashMap, 
the later one has better performance and better safety (never throw 
ConcurrentModificationException) but just not keep order of element which we 
don't need in our case. Given we have YARN-3390 to continue refactor work on 
the same piece of code, we can continue the discussion there.

Latest patch LGTM. +1. Will go ahead to commit it within 24h if no 
object/further comments from others.

 convert load test driver to timeline service v.2
 

 Key: YARN-3437
 URL: https://issues.apache.org/jira/browse/YARN-3437
 Project: Hadoop YARN
  Issue Type: Sub-task
  Components: timelineserver
Reporter: Sangjin Lee
Assignee: Sangjin Lee
 Attachments: YARN-3437.001.patch, YARN-3437.002.patch, 
 YARN-3437.003.patch, YARN-3437.004.patch


 This subtask covers the work for converting the proposed patch for the load 
 test driver (YARN-2556) to work with the timeline service v.2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (YARN-3437) convert load test driver to timeline service v.2

2015-04-22 Thread Sangjin Lee (JIRA)

[ 
https://issues.apache.org/jira/browse/YARN-3437?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14508151#comment-14508151
 ] 

Sangjin Lee commented on YARN-3437:
---

Could you kindly take a look at the latest patch? Thanks!

 convert load test driver to timeline service v.2
 

 Key: YARN-3437
 URL: https://issues.apache.org/jira/browse/YARN-3437
 Project: Hadoop YARN
  Issue Type: Sub-task
  Components: timelineserver
Reporter: Sangjin Lee
Assignee: Sangjin Lee
 Attachments: YARN-3437.001.patch, YARN-3437.002.patch, 
 YARN-3437.003.patch


 This subtask covers the work for converting the proposed patch for the load 
 test driver (YARN-2556) to work with the timeline service v.2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (YARN-3437) convert load test driver to timeline service v.2

2015-04-22 Thread Sangjin Lee (JIRA)

[ 
https://issues.apache.org/jira/browse/YARN-3437?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14508378#comment-14508378
 ] 

Sangjin Lee commented on YARN-3437:
---

Thanks for the review [~djp]!

bq. For performance perspective, we should move LOG.info() out of synchronized 
block (may be move out of collector.start()?).

I can move the LOG.info() call outside the synchronized block. That said, I 
don't think this would have a meaningful performance impact. Aside from the 
fact that logging calls are usually synchronized themselves, it is reasonable 
to expect that the contention for this lock (collectors) would be quite low. 
We're talking about contention when multiple AMs are competing to create 
collectors on the same node, and the chances that there is any contention on 
this lock would be very low.

Also, when you said may be move out of collector.start(), did you mean moving 
the collector.start() call outside the synchronization block? If so, I'd be 
hesitant to do that. We just had a discussion on this in another JIRA (see 
https://issues.apache.org/jira/browse/YARN-3390?focusedCommentId=14508121page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-14508121).

bq. we don't need to LOG.ERROR (replace with INFO?)

That is a good suggestion. I'll update this (and remove()) to lower the logging 
level for this.

bq. For remove(), similar that we should move collector.stop() and LOG.info() 
out of synchronized block.
This we can do safely. I'll update the patch.

 convert load test driver to timeline service v.2
 

 Key: YARN-3437
 URL: https://issues.apache.org/jira/browse/YARN-3437
 Project: Hadoop YARN
  Issue Type: Sub-task
  Components: timelineserver
Reporter: Sangjin Lee
Assignee: Sangjin Lee
 Attachments: YARN-3437.001.patch, YARN-3437.002.patch, 
 YARN-3437.003.patch


 This subtask covers the work for converting the proposed patch for the load 
 test driver (YARN-2556) to work with the timeline service v.2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (YARN-3437) convert load test driver to timeline service v.2

2015-04-22 Thread Hadoop QA (JIRA)

[ 
https://issues.apache.org/jira/browse/YARN-3437?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14508400#comment-14508400
 ] 

Hadoop QA commented on YARN-3437:
-

\\
\\
| (x) *{color:red}-1 overall{color}* |
\\
\\
|| Vote || Subsystem || Runtime || Comment ||
| {color:red}-1{color} | patch |   0m  0s | The patch command could not apply 
the patch during dryrun. |
\\
\\
|| Subsystem || Report/Notes ||
| Patch URL | 
http://issues.apache.org/jira/secure/attachment/12727521/YARN-3437.004.patch |
| Optional Tests | javac unit findbugs checkstyle javadoc |
| git revision | trunk / a100be6 |
| Console output | 
https://builds.apache.org/job/PreCommit-YARN-Build/7466//console |


This message was automatically generated.

 convert load test driver to timeline service v.2
 

 Key: YARN-3437
 URL: https://issues.apache.org/jira/browse/YARN-3437
 Project: Hadoop YARN
  Issue Type: Sub-task
  Components: timelineserver
Reporter: Sangjin Lee
Assignee: Sangjin Lee
 Attachments: YARN-3437.001.patch, YARN-3437.002.patch, 
 YARN-3437.003.patch, YARN-3437.004.patch


 This subtask covers the work for converting the proposed patch for the load 
 test driver (YARN-2556) to work with the timeline service v.2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (YARN-3437) convert load test driver to timeline service v.2

2015-04-22 Thread Hadoop QA (JIRA)

[ 
https://issues.apache.org/jira/browse/YARN-3437?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14507948#comment-14507948
 ] 

Hadoop QA commented on YARN-3437:
-

\\
\\
| (x) *{color:red}-1 overall{color}* |
\\
\\
|| Vote || Subsystem || Runtime || Comment ||
| {color:red}-1{color} | patch |   0m  0s | The patch command could not apply 
the patch during dryrun. |
\\
\\
|| Subsystem || Report/Notes ||
| Patch URL | 
http://issues.apache.org/jira/secure/attachment/12727346/YARN-3437.003.patch |
| Optional Tests | javac unit findbugs checkstyle javadoc |
| git revision | trunk / 0ebe84d |
| Console output | 
https://builds.apache.org/job/PreCommit-YARN-Build/7459//console |


This message was automatically generated.

 convert load test driver to timeline service v.2
 

 Key: YARN-3437
 URL: https://issues.apache.org/jira/browse/YARN-3437
 Project: Hadoop YARN
  Issue Type: Sub-task
  Components: timelineserver
Reporter: Sangjin Lee
Assignee: Sangjin Lee
 Attachments: YARN-3437.001.patch, YARN-3437.002.patch, 
 YARN-3437.003.patch


 This subtask covers the work for converting the proposed patch for the load 
 test driver (YARN-2556) to work with the timeline service v.2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (YARN-3437) convert load test driver to timeline service v.2

2015-04-22 Thread Junping Du (JIRA)

[ 
https://issues.apache.org/jira/browse/YARN-3437?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14506792#comment-14506792
 ] 

Junping Du commented on YARN-3437:
--

bq. The downside is code duplication.
Agree. However, I think this is the least price we could pay when developing on 
different branches in parallel.

bq. If others are comfortable with that, I could quickly move it to a different 
package.
I am OK with this in case we have additional JIRA to track removing duplicated 
code after YARN-2556 land. We can also add some credit to contributors of 
YARN-2556 in commit/CHANGE messages which should be very common.

[~zjshen] and [~jeagles], any ideas on this?

 convert load test driver to timeline service v.2
 

 Key: YARN-3437
 URL: https://issues.apache.org/jira/browse/YARN-3437
 Project: Hadoop YARN
  Issue Type: Sub-task
  Components: timelineserver
Reporter: Sangjin Lee
Assignee: Sangjin Lee
 Attachments: YARN-3437.001.patch, YARN-3437.002.patch


 This subtask covers the work for converting the proposed patch for the load 
 test driver (YARN-2556) to work with the timeline service v.2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (YARN-3437) convert load test driver to timeline service v.2

2015-04-21 Thread Junping Du (JIRA)

[ 
https://issues.apache.org/jira/browse/YARN-3437?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14505593#comment-14505593
 ] 

Junping Du commented on YARN-3437:
--

I agree too if we don't have clear plan for YARN-2556 so far. There should be 
no reason to block other going efforts.
An suggestion (optional only) is: can we adjust name (or package path) slightly 
for duplicated file (TimelineServerPerformance.java) with YARN-2556? We can 
have an additional patch to remove duplicated file when YARN-2556 get in trunk.
I assume this could be easier for YARN-2928 rebase back to trunk/branch-2 as 
less conflict. Thoughts?

 convert load test driver to timeline service v.2
 

 Key: YARN-3437
 URL: https://issues.apache.org/jira/browse/YARN-3437
 Project: Hadoop YARN
  Issue Type: Sub-task
  Components: timelineserver
Reporter: Sangjin Lee
Assignee: Sangjin Lee
 Attachments: YARN-3437.001.patch, YARN-3437.002.patch


 This subtask covers the work for converting the proposed patch for the load 
 test driver (YARN-2556) to work with the timeline service v.2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (YARN-3437) convert load test driver to timeline service v.2

2015-04-21 Thread Sangjin Lee (JIRA)

[ 
https://issues.apache.org/jira/browse/YARN-3437?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14505382#comment-14505382
 ] 

Sangjin Lee commented on YARN-3437:
---

I think we need to make progress on this as this is blocking other JIRAs and 
also it's tied to the schema evaluation.

My vote is to get this committed, and adjust this once YARN-2556 lands and we 
rebase. Thoughts?

 convert load test driver to timeline service v.2
 

 Key: YARN-3437
 URL: https://issues.apache.org/jira/browse/YARN-3437
 Project: Hadoop YARN
  Issue Type: Sub-task
  Components: timelineserver
Reporter: Sangjin Lee
Assignee: Sangjin Lee
 Attachments: YARN-3437.001.patch, YARN-3437.002.patch


 This subtask covers the work for converting the proposed patch for the load 
 test driver (YARN-2556) to work with the timeline service v.2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (YARN-3437) convert load test driver to timeline service v.2

2015-04-21 Thread Junping Du (JIRA)

[ 
https://issues.apache.org/jira/browse/YARN-3437?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14504842#comment-14504842
 ] 

Junping Du commented on YARN-3437:
--

Filed a new JIRA to track fine-grained performance data sounds good to me. 
Given the patch here have duplicated code with YARN-2556, I would like to 
understand what's our plan for YARN-2556. [~jeagles], can you share your vision 
on this? 
Looks like this JIRA block YARN-3390 (a refactor JIRA) which block YARN-3044 
(RM writing events to v2 ATS service). I would like to have a clear path to 
make all patches goes in as a pipeline with getting ride of any potential 
deadlock. :) 
May be the first step is to make YARN-2556 get committed it, and get patch here 
rebased? [~jeagles], [~sjlee0] and [~zjshen], what's your opinion on this? 

 convert load test driver to timeline service v.2
 

 Key: YARN-3437
 URL: https://issues.apache.org/jira/browse/YARN-3437
 Project: Hadoop YARN
  Issue Type: Sub-task
  Components: timelineserver
Reporter: Sangjin Lee
Assignee: Sangjin Lee
 Attachments: YARN-3437.001.patch, YARN-3437.002.patch


 This subtask covers the work for converting the proposed patch for the load 
 test driver (YARN-2556) to work with the timeline service v.2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (YARN-3437) convert load test driver to timeline service v.2

2015-04-21 Thread Sangjin Lee (JIRA)

[ 
https://issues.apache.org/jira/browse/YARN-3437?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14506157#comment-14506157
 ] 

Sangjin Lee commented on YARN-3437:
---

The downside is code duplication, but as you said we would not be impacted 
when/if YARN-2556 lands. If others are comfortable with that, I could quickly 
move it to a different package. Let me know soon, and I'll update the patch.

 convert load test driver to timeline service v.2
 

 Key: YARN-3437
 URL: https://issues.apache.org/jira/browse/YARN-3437
 Project: Hadoop YARN
  Issue Type: Sub-task
  Components: timelineserver
Reporter: Sangjin Lee
Assignee: Sangjin Lee
 Attachments: YARN-3437.001.patch, YARN-3437.002.patch


 This subtask covers the work for converting the proposed patch for the load 
 test driver (YARN-2556) to work with the timeline service v.2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (YARN-3437) convert load test driver to timeline service v.2

2015-04-20 Thread Sangjin Lee (JIRA)

[ 
https://issues.apache.org/jira/browse/YARN-3437?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14503383#comment-14503383
 ] 

Sangjin Lee commented on YARN-3437:
---

[~jeagles], thanks for your comment.

Yes, I agree the level of reporting in this patch is mostly basic and very high 
level. I definitely agree that we need to add more reporting in terms of the 
actual number of writes/puts done on the storage, etc. The first use of this 
tool is to compare the overall throughput/performance of the storage 
implementations and identify high level issues. I suspect for that the 
coarse-grained reporting might be enough.

As a follow-up JIRA, I'll add a new JIRA that reports more fine-grained write 
performance data.

 convert load test driver to timeline service v.2
 

 Key: YARN-3437
 URL: https://issues.apache.org/jira/browse/YARN-3437
 Project: Hadoop YARN
  Issue Type: Sub-task
  Components: timelineserver
Reporter: Sangjin Lee
Assignee: Sangjin Lee
 Attachments: YARN-3437.001.patch, YARN-3437.002.patch


 This subtask covers the work for converting the proposed patch for the load 
 test driver (YARN-2556) to work with the timeline service v.2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (YARN-3437) convert load test driver to timeline service v.2

2015-04-20 Thread Sangjin Lee (JIRA)

[ 
https://issues.apache.org/jira/browse/YARN-3437?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14503391#comment-14503391
 ] 

Sangjin Lee commented on YARN-3437:
---

Added YARN-3512.

 convert load test driver to timeline service v.2
 

 Key: YARN-3437
 URL: https://issues.apache.org/jira/browse/YARN-3437
 Project: Hadoop YARN
  Issue Type: Sub-task
  Components: timelineserver
Reporter: Sangjin Lee
Assignee: Sangjin Lee
 Attachments: YARN-3437.001.patch, YARN-3437.002.patch


 This subtask covers the work for converting the proposed patch for the load 
 test driver (YARN-2556) to work with the timeline service v.2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (YARN-3437) convert load test driver to timeline service v.2

2015-04-17 Thread Zhijie Shen (JIRA)

[ 
https://issues.apache.org/jira/browse/YARN-3437?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14499247#comment-14499247
 ] 

Zhijie Shen commented on YARN-3437:
---

Per my comment on [YARN-3390 | 
https://issues.apache.org/jira/browse/YARN-3390?focusedCommentId=14499245page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-14499245].
 Please feel free to move forward.

 convert load test driver to timeline service v.2
 

 Key: YARN-3437
 URL: https://issues.apache.org/jira/browse/YARN-3437
 Project: Hadoop YARN
  Issue Type: Sub-task
  Components: timelineserver
Reporter: Sangjin Lee
Assignee: Sangjin Lee
 Attachments: YARN-3437.001.patch, YARN-3437.002.patch


 This subtask covers the work for converting the proposed patch for the load 
 test driver (YARN-2556) to work with the timeline service v.2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (YARN-3437) convert load test driver to timeline service v.2

2015-04-17 Thread Jonathan Eagles (JIRA)

[ 
https://issues.apache.org/jira/browse/YARN-3437?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14500187#comment-14500187
 ] 

Jonathan Eagles commented on YARN-3437:
---

Now that I have dug into timeline server performance (YARN-3448). I have a 
better understanding of what type of writes are costly. For example, a single 
entity will generate dozens or writes to the database. The number of primary 
keys, the number of related entities, and the write batch size (entities per 
put) greatly affect the time an entity put takes. While this is a good start, I 
think there should at least be a follow up that addresses these issues to 
better measure the write performance.

 convert load test driver to timeline service v.2
 

 Key: YARN-3437
 URL: https://issues.apache.org/jira/browse/YARN-3437
 Project: Hadoop YARN
  Issue Type: Sub-task
  Components: timelineserver
Reporter: Sangjin Lee
Assignee: Sangjin Lee
 Attachments: YARN-3437.001.patch, YARN-3437.002.patch


 This subtask covers the work for converting the proposed patch for the load 
 test driver (YARN-2556) to work with the timeline service v.2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (YARN-3437) convert load test driver to timeline service v.2

2015-04-17 Thread Zhijie Shen (JIRA)

[ 
https://issues.apache.org/jira/browse/YARN-3437?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14500112#comment-14500112
 ] 

Zhijie Shen commented on YARN-3437:
---

Will take a look today.

 convert load test driver to timeline service v.2
 

 Key: YARN-3437
 URL: https://issues.apache.org/jira/browse/YARN-3437
 Project: Hadoop YARN
  Issue Type: Sub-task
  Components: timelineserver
Reporter: Sangjin Lee
Assignee: Sangjin Lee
 Attachments: YARN-3437.001.patch, YARN-3437.002.patch


 This subtask covers the work for converting the proposed patch for the load 
 test driver (YARN-2556) to work with the timeline service v.2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (YARN-3437) convert load test driver to timeline service v.2

2015-04-17 Thread Sangjin Lee (JIRA)

[ 
https://issues.apache.org/jira/browse/YARN-3437?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14500088#comment-14500088
 ] 

Sangjin Lee commented on YARN-3437:
---

Thanks [~zjshen]! Could someone please commit this patch unless there are 
further comments?

 convert load test driver to timeline service v.2
 

 Key: YARN-3437
 URL: https://issues.apache.org/jira/browse/YARN-3437
 Project: Hadoop YARN
  Issue Type: Sub-task
  Components: timelineserver
Reporter: Sangjin Lee
Assignee: Sangjin Lee
 Attachments: YARN-3437.001.patch, YARN-3437.002.patch


 This subtask covers the work for converting the proposed patch for the load 
 test driver (YARN-2556) to work with the timeline service v.2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (YARN-3437) convert load test driver to timeline service v.2

2015-04-15 Thread Hadoop QA (JIRA)

[ 
https://issues.apache.org/jira/browse/YARN-3437?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14497432#comment-14497432
 ] 

Hadoop QA commented on YARN-3437:
-

{color:red}-1 overall{color}.  Here are the results of testing the latest 
attachment 
  http://issues.apache.org/jira/secure/attachment/12725758/YARN-3437.002.patch
  against trunk revision 1b89a3e.

{color:red}-1 patch{color}.  The patch command could not apply the patch.

Console output: https://builds.apache.org/job/PreCommit-YARN-Build/7354//console

This message is automatically generated.

 convert load test driver to timeline service v.2
 

 Key: YARN-3437
 URL: https://issues.apache.org/jira/browse/YARN-3437
 Project: Hadoop YARN
  Issue Type: Sub-task
  Components: timelineserver
Reporter: Sangjin Lee
Assignee: Sangjin Lee
 Attachments: YARN-3437.001.patch, YARN-3437.002.patch


 This subtask covers the work for converting the proposed patch for the load 
 test driver (YARN-2556) to work with the timeline service v.2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (YARN-3437) convert load test driver to timeline service v.2

2015-04-14 Thread Sangjin Lee (JIRA)

[ 
https://issues.apache.org/jira/browse/YARN-3437?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14494524#comment-14494524
 ] 

Sangjin Lee commented on YARN-3437:
---

Forgot to add that I copied it to get the load testing going. Once YARN-2556 
gets committed on trunk, I'd need to update this to handle both the old 
timeline service and the new timeline service.

 convert load test driver to timeline service v.2
 

 Key: YARN-3437
 URL: https://issues.apache.org/jira/browse/YARN-3437
 Project: Hadoop YARN
  Issue Type: Sub-task
  Components: timelineserver
Reporter: Sangjin Lee
Assignee: Sangjin Lee
 Attachments: YARN-3437.001.patch


 This subtask covers the work for converting the proposed patch for the load 
 test driver (YARN-2556) to work with the timeline service v.2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (YARN-3437) convert load test driver to timeline service v.2

2015-04-14 Thread Sangjin Lee (JIRA)

[ 
https://issues.apache.org/jira/browse/YARN-3437?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14494521#comment-14494521
 ] 

Sangjin Lee commented on YARN-3437:
---

Thanks [~djp] for your comments!

I agree with most of your comments. This particular subtask generates a pretty 
simple model (just a straightforward conversion of YARN-2556) where each map 
task creates one app collector and simple entities. But we will add more test 
modes that will have different characteristics. For one, we will add a mode 
where it iterates over MR job history files and pump them into the timeline 
storage.

The number of mappers is a pretty useful way of controlling the parallelism of 
tests, and we can add more parameters to have finer-grained parallelism 
controls (one app collector manager having multiple app collectors, etc.).

bq. BTW, it has some duplicated code with YARN-2556 (like 
TimelineServerPerformance.java).
That's because I copied it (as mentioned in the description). :) I had to fix a 
few issues with the original patch there to get it going. Some appear to be 
unrelated to porting it to the timeline service v.2. I'll add some comments 
there to help it get committed.

 convert load test driver to timeline service v.2
 

 Key: YARN-3437
 URL: https://issues.apache.org/jira/browse/YARN-3437
 Project: Hadoop YARN
  Issue Type: Sub-task
  Components: timelineserver
Reporter: Sangjin Lee
Assignee: Sangjin Lee
 Attachments: YARN-3437.001.patch


 This subtask covers the work for converting the proposed patch for the load 
 test driver (YARN-2556) to work with the timeline service v.2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (YARN-3437) convert load test driver to timeline service v.2

2015-04-03 Thread Junping Du (JIRA)

[ 
https://issues.apache.org/jira/browse/YARN-3437?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14394437#comment-14394437
 ] 

Junping Du commented on YARN-3437:
--

Thanks [~sjlee0] for delivering a patch here! 
Just quickly go through the patch, looks like we are generating one app 
collector per map task. I think this is good for scalability test on backend 
storage which can be a bottleneck in mainstream cases. In addition, do we want 
to address some extreme cases, e.g. a huge applications will have hundreds of 
thousands or even millions tasks? If so, then may be we want to know a single 
app collector's bottleneck as well for accepting/forwarding messages from 
hundreds of thousands maps. Also, in a real cluster, the mapping from cluster 
to app, and app to tasks are all 1-N mapping. May be making app aggregator 
number configurable (just like map task number, and byte per map, etc.) is 
something we can do for next step?
BTW, it has some duplicated code with YARN-2556 (like 
TimelineServerPerformance.java). Looks like YARN-2556 is in pretty good shape 
and possible to go to trunk and branch-2 quickly. I would remind to keep 
watching that JIRA status and do necessary rebase work if that patch go in and 
we may want to merge it into YARN-2928 branch soon.

 convert load test driver to timeline service v.2
 

 Key: YARN-3437
 URL: https://issues.apache.org/jira/browse/YARN-3437
 Project: Hadoop YARN
  Issue Type: Sub-task
  Components: timelineserver
Reporter: Sangjin Lee
Assignee: Sangjin Lee
 Attachments: YARN-3437.001.patch


 This subtask covers the work for converting the proposed patch for the load 
 test driver (YARN-2556) to work with the timeline service v.2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (YARN-3437) convert load test driver to timeline service v.2

2015-04-02 Thread Hadoop QA (JIRA)

[ 
https://issues.apache.org/jira/browse/YARN-3437?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14393554#comment-14393554
 ] 

Hadoop QA commented on YARN-3437:
-

{color:red}-1 overall{color}.  Here are the results of testing the latest 
attachment 
  http://issues.apache.org/jira/secure/attachment/12709078/YARN-3437.001.patch
  against trunk revision 6a6a59d.

{color:red}-1 patch{color}.  The patch command could not apply the patch.

Console output: https://builds.apache.org/job/PreCommit-YARN-Build/7206//console

This message is automatically generated.

 convert load test driver to timeline service v.2
 

 Key: YARN-3437
 URL: https://issues.apache.org/jira/browse/YARN-3437
 Project: Hadoop YARN
  Issue Type: Sub-task
  Components: timelineserver
Reporter: Sangjin Lee
Assignee: Sangjin Lee
 Attachments: YARN-3437.001.patch


 This subtask covers the work for converting the proposed patch for the load 
 test driver (YARN-2556) to work with the timeline service v.2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)


[jira] [Commented] (YARN-3437) convert load test driver to timeline service v.2

2015-04-02 Thread Sangjin Lee (JIRA)

[ 
https://issues.apache.org/jira/browse/YARN-3437?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14393530#comment-14393530
 ] 

Sangjin Lee commented on YARN-3437:
---

Added a few folks for review.

 convert load test driver to timeline service v.2
 

 Key: YARN-3437
 URL: https://issues.apache.org/jira/browse/YARN-3437
 Project: Hadoop YARN
  Issue Type: Sub-task
  Components: timelineserver
Reporter: Sangjin Lee
Assignee: Sangjin Lee
 Attachments: YARN-3437.001.patch


 This subtask covers the work for converting the proposed patch for the load 
 test driver (YARN-2556) to work with the timeline service v.2.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)