[GitHub] spark issue #22849: [SPARK-25852][Core] we should filter the workOffers with...

2018-10-30 Thread jiangxb1987
Github user jiangxb1987 commented on the issue:

https://github.com/apache/spark/pull/22849
  
It may happen that a busy executor is marked as lost and later it 
re-register to the driver, in that case currently we call `makeOffers()` and 
that will add the executor into `TaskSchedulerImpl.hostToExecutors`. This is 
bad implementation here since it shall not have depend on the `makeOffers()` 
function to update a unrelated protected val, but that's what we have for now.


---

-
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org



[GitHub] spark issue #22849: [SPARK-25852][Core] we should filter the workOffers with...

2018-10-30 Thread jiangxb1987
Github user jiangxb1987 commented on the issue:

https://github.com/apache/spark/pull/22849
  
What do you mean by "better performance" ? If that means we can spend less 
time on `TaskSchedulerImpl.resourceOffers()` then I agree it's true, but AFAIK 
it's never reported this can be a bottleneck of the whole cluster, so maybe the 
perf gain is trivial here. If you expect better task distribution over existing 
executors then I don't see any case can be improved by this proposed change. 
Please correct me if I'm wrong.


---

-
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org



[GitHub] spark issue #22849: [SPARK-25852][Core] we should filter the workOffers with...

2018-10-30 Thread srowen
Github user srowen commented on the issue:

https://github.com/apache/spark/pull/22849
  
Yes, that is the comment I have been referring to. So it seems you can't 
filter, right? it's not scheduling work here. @jiangxb1987 do you know this 
part of the code?


---

-
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org



[GitHub] spark issue #22849: [SPARK-25852][Core] we should filter the workOffers of w...

2018-10-26 Thread AmplabJenkins
Github user AmplabJenkins commented on the issue:

https://github.com/apache/spark/pull/22849
  
Can one of the admins verify this patch?


---

-
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org



[GitHub] spark issue #22849: [SPARK-25852][Core] we should filter the workOffers of w...

2018-10-26 Thread AmplabJenkins
Github user AmplabJenkins commented on the issue:

https://github.com/apache/spark/pull/22849
  
Can one of the admins verify this patch?


---

-
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org



[GitHub] spark issue #22849: [SPARK-25852][Core] we should filter the workOffers of w...

2018-10-26 Thread AmplabJenkins
Github user AmplabJenkins commented on the issue:

https://github.com/apache/spark/pull/22849
  
Can one of the admins verify this patch?


---

-
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org



[GitHub] spark issue #22849: [SPARK-25852][Core] we should filter the workOffers of w...

2018-10-26 Thread zuotingbing
Github user zuotingbing commented on the issue:

https://github.com/apache/spark/pull/22849
  

![2018-10-26_162822](https://user-images.githubusercontent.com/24823338/47558814-9aa14b00-d946-11e8-8712-de05a93b25a9.png)



---

-
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org