You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@hive.apache.org by Jimmy Xiang <jx...@cloudera.com> on 2015/02/26 02:42:13 UTC

Review Request 31452: HIVE-9781 Utilize spark.kryo.classesToRegister [Spark Branch]

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/31452/
-----------------------------------------------------------

Review request for hive, Brock Noland and Xuefu Zhang.


Bugs: HIVE-9781
    https://issues.apache.org/jira/browse/HIVE-9781


Repository: hive-git


Description
-------

Now, Spark RSC can use setting spark.kryo.classesToRegister. Reference checking is still needed since there are loops.


Diffs
-----

  ql/src/java/org/apache/hadoop/hive/ql/exec/spark/KryoSerializer.java 76a4fd1 
  ql/src/java/org/apache/hadoop/hive/ql/exec/spark/RemoteHiveSparkClient.java f1d65ea 

Diff: https://reviews.apache.org/r/31452/diff/


Testing
-------

Tested it on a yarn cluster with spark.kryo.classesToRegister set.


Thanks,

Jimmy Xiang


Re: Review Request 31452: HIVE-9781 Utilize spark.kryo.classesToRegister [Spark Branch]

Posted by Xuefu Zhang <xz...@cloudera.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/31452/#review74370
-----------------------------------------------------------

Ship it!


Ship It!

- Xuefu Zhang


On Feb. 26, 2015, 9:38 p.m., Jimmy Xiang wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/31452/
> -----------------------------------------------------------
> 
> (Updated Feb. 26, 2015, 9:38 p.m.)
> 
> 
> Review request for hive, Brock Noland and Xuefu Zhang.
> 
> 
> Bugs: HIVE-9781
>     https://issues.apache.org/jira/browse/HIVE-9781
> 
> 
> Repository: hive-git
> 
> 
> Description
> -------
> 
> Now, Spark RSC can use setting spark.kryo.classesToRegister. Reference checking is still needed since there are loops.
> 
> 
> Diffs
> -----
> 
>   ql/src/java/org/apache/hadoop/hive/ql/exec/spark/KryoSerializer.java 76a4fd1 
>   ql/src/java/org/apache/hadoop/hive/ql/exec/spark/RemoteHiveSparkClient.java f1d65ea 
> 
> Diff: https://reviews.apache.org/r/31452/diff/
> 
> 
> Testing
> -------
> 
> Tested it on a yarn cluster with spark.kryo.classesToRegister set.
> 
> 
> Thanks,
> 
> Jimmy Xiang
> 
>


Re: Review Request 31452: HIVE-9781 Utilize spark.kryo.classesToRegister [Spark Branch]

Posted by Jimmy Xiang <jx...@cloudera.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/31452/
-----------------------------------------------------------

(Updated Feb. 26, 2015, 9:38 p.m.)


Review request for hive, Brock Noland and Xuefu Zhang.


Changes
-------

Instead of using spark.kryo.classesToRegister, we register those used classes, per Xuefu's suggestion.


Bugs: HIVE-9781
    https://issues.apache.org/jira/browse/HIVE-9781


Repository: hive-git


Description
-------

Now, Spark RSC can use setting spark.kryo.classesToRegister. Reference checking is still needed since there are loops.


Diffs (updated)
-----

  ql/src/java/org/apache/hadoop/hive/ql/exec/spark/KryoSerializer.java 76a4fd1 
  ql/src/java/org/apache/hadoop/hive/ql/exec/spark/RemoteHiveSparkClient.java f1d65ea 

Diff: https://reviews.apache.org/r/31452/diff/


Testing
-------

Tested it on a yarn cluster with spark.kryo.classesToRegister set.


Thanks,

Jimmy Xiang


Re: Review Request 31452: HIVE-9781 Utilize spark.kryo.classesToRegister [Spark Branch]

Posted by Jimmy Xiang <jx...@cloudera.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/31452/
-----------------------------------------------------------

(Updated Feb. 26, 2015, 5:39 p.m.)


Review request for hive, Brock Noland and Xuefu Zhang.


Bugs: HIVE-9781
    https://issues.apache.org/jira/browse/HIVE-9781


Repository: hive-git


Description
-------

Now, Spark RSC can use setting spark.kryo.classesToRegister. Reference checking is still needed since there are loops.


Diffs (updated)
-----

  ql/src/java/org/apache/hadoop/hive/ql/exec/spark/KryoSerializer.java 76a4fd1 
  ql/src/java/org/apache/hadoop/hive/ql/exec/spark/RemoteHiveSparkClient.java f1d65ea 

Diff: https://reviews.apache.org/r/31452/diff/


Testing
-------

Tested it on a yarn cluster with spark.kryo.classesToRegister set.


Thanks,

Jimmy Xiang


Re: Review Request 31452: HIVE-9781 Utilize spark.kryo.classesToRegister [Spark Branch]

Posted by Jimmy Xiang <jx...@cloudera.com>.

> On Feb. 26, 2015, 2:53 p.m., Xuefu Zhang wrote:
> > ql/src/java/org/apache/hadoop/hive/ql/exec/spark/RemoteHiveSparkClient.java, line 152
> > <https://reviews.apache.org/r/31452/diff/1/?file=877054#file877054line152>
> >
> >     Could we rename this variable for better readability?

Sure.


- Jimmy


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/31452/#review74273
-----------------------------------------------------------


On Feb. 26, 2015, 1:42 a.m., Jimmy Xiang wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/31452/
> -----------------------------------------------------------
> 
> (Updated Feb. 26, 2015, 1:42 a.m.)
> 
> 
> Review request for hive, Brock Noland and Xuefu Zhang.
> 
> 
> Bugs: HIVE-9781
>     https://issues.apache.org/jira/browse/HIVE-9781
> 
> 
> Repository: hive-git
> 
> 
> Description
> -------
> 
> Now, Spark RSC can use setting spark.kryo.classesToRegister. Reference checking is still needed since there are loops.
> 
> 
> Diffs
> -----
> 
>   ql/src/java/org/apache/hadoop/hive/ql/exec/spark/KryoSerializer.java 76a4fd1 
>   ql/src/java/org/apache/hadoop/hive/ql/exec/spark/RemoteHiveSparkClient.java f1d65ea 
> 
> Diff: https://reviews.apache.org/r/31452/diff/
> 
> 
> Testing
> -------
> 
> Tested it on a yarn cluster with spark.kryo.classesToRegister set.
> 
> 
> Thanks,
> 
> Jimmy Xiang
> 
>


Re: Review Request 31452: HIVE-9781 Utilize spark.kryo.classesToRegister [Spark Branch]

Posted by Xuefu Zhang <xz...@cloudera.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/31452/#review74273
-----------------------------------------------------------



ql/src/java/org/apache/hadoop/hive/ql/exec/spark/RemoteHiveSparkClient.java
<https://reviews.apache.org/r/31452/#comment120817>

    Could we rename this variable for better readability?


- Xuefu Zhang


On Feb. 26, 2015, 1:42 a.m., Jimmy Xiang wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/31452/
> -----------------------------------------------------------
> 
> (Updated Feb. 26, 2015, 1:42 a.m.)
> 
> 
> Review request for hive, Brock Noland and Xuefu Zhang.
> 
> 
> Bugs: HIVE-9781
>     https://issues.apache.org/jira/browse/HIVE-9781
> 
> 
> Repository: hive-git
> 
> 
> Description
> -------
> 
> Now, Spark RSC can use setting spark.kryo.classesToRegister. Reference checking is still needed since there are loops.
> 
> 
> Diffs
> -----
> 
>   ql/src/java/org/apache/hadoop/hive/ql/exec/spark/KryoSerializer.java 76a4fd1 
>   ql/src/java/org/apache/hadoop/hive/ql/exec/spark/RemoteHiveSparkClient.java f1d65ea 
> 
> Diff: https://reviews.apache.org/r/31452/diff/
> 
> 
> Testing
> -------
> 
> Tested it on a yarn cluster with spark.kryo.classesToRegister set.
> 
> 
> Thanks,
> 
> Jimmy Xiang
> 
>