[ 
https://issues.apache.org/jira/browse/SPARK-33738?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

chenliang updated SPARK-33738:
------------------------------
    Description: 
In order to facilitate the troubleshooting of hotspots or data skew, it's 
better to add more information on  ShuffleBlockFetcherIterator.

The new log are shown as below:
{code:java}
2020-07-27 21:00:49,910 TRACE 
org.apache.spark.storage.ShuffleBlockFetcherIterator: Got remote block 
ArrayBuffer(shuffle_1_39_5, shuffle_1_42_5, shuffle_1_45_5, shuffle_1_48_5, 
shuffle_1_51_5, shuffle_1_434_5, shuffle_1_627_5, shuffle_1_689_5, 
shuffle_1_791_5, shuffle_1_856_5, shuffle_1_859_5, shuffle_1_867_5, 
shuffle_1_875_5, shuffle_1_915_5) from hostname:7337 cost  3511 ms
{code}

We can easily get the point that  the current container cost 3511 ms for 
getting remote block list from  node hostname:7337.

  was:
In order to facilitate the troubleshooting of hotspots or data skew, it's 
better to add more information on  ShuffleBlockFetcherIterator.

The new log are shown as below:
{code:java}
2020-07-27 21:00:49,910 TRACE 
org.apache.spark.storage.ShuffleBlockFetcherIterator: Got remote block 
ArrayBuffer(shuffle_1_39_5, shuffle_1_42_5, shuffle_1_45_5, shuffle_1_48_5, 
shuffle_1_51_5, shuffle_1_434_5, shuffle_1_627_5, shuffle_1_689_5, 
shuffle_1_791_5, shuffle_1_856_5, shuffle_1_859_5, shuffle_1_867_5, 
shuffle_1_875_5, shuffle_1_915_5) from hostname:7337 cost  3511 ms
{code}

We can easily get the point that  the current container cost 3511 ms for 
getting remote block list from  node hostname.diditaxi.com:7337.


> Add timeout info for ShuffleBlockFetcherIterator when get remote block
> ----------------------------------------------------------------------
>
>                 Key: SPARK-33738
>                 URL: https://issues.apache.org/jira/browse/SPARK-33738
>             Project: Spark
>          Issue Type: Improvement
>          Components: Spark Core
>    Affects Versions: 2.4.3, 3.0.0
>            Reporter: chenliang
>            Priority: Minor
>
> In order to facilitate the troubleshooting of hotspots or data skew, it's 
> better to add more information on  ShuffleBlockFetcherIterator.
> The new log are shown as below:
> {code:java}
> 2020-07-27 21:00:49,910 TRACE 
> org.apache.spark.storage.ShuffleBlockFetcherIterator: Got remote block 
> ArrayBuffer(shuffle_1_39_5, shuffle_1_42_5, shuffle_1_45_5, shuffle_1_48_5, 
> shuffle_1_51_5, shuffle_1_434_5, shuffle_1_627_5, shuffle_1_689_5, 
> shuffle_1_791_5, shuffle_1_856_5, shuffle_1_859_5, shuffle_1_867_5, 
> shuffle_1_875_5, shuffle_1_915_5) from hostname:7337 cost  3511 ms
> {code}
> We can easily get the point that  the current container cost 3511 ms for 
> getting remote block list from  node hostname:7337.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to