ambari-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Alejandro Fernandez" <afernan...@hortonworks.com>
Subject Re: Review Request 34618: Spark History Server in HDP 2.3 no longer needs tez.tar.gz, which now needs to be uploaded to HDFS during Tez Service Check
Date Sat, 23 May 2015 02:04:59 GMT

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/34618/#review85051
-----------------------------------------------------------



ambari-server/src/main/resources/common-services/TEZ/0.4.0.2.1/package/scripts/service_check.py
<https://reviews.apache.org/r/34618/#comment136522>

    In HDP 2.2, it would rely on other services to copy the tarball to HDFS, primarily Hive,
Pig, and then Spark in 2.2.4.2. Technically, it should have been responsible too, so I'll
fix this.



ambari-server/src/main/resources/stacks/HDP/2.3/services/SPARK/metainfo.xml
<https://reviews.apache.org/r/34618/#comment136523>

    I can create a SPARK folder in common-services with the newer version of Spark, and remove
TEZ from that metainfo.xml file, so that this file in HDP 2.3 would not need anything.


- Alejandro Fernandez


On May 23, 2015, 1:52 a.m., Alejandro Fernandez wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/34618/
> -----------------------------------------------------------
> 
> (Updated May 23, 2015, 1:52 a.m.)
> 
> 
> Review request for Ambari, Dmitro Lisnichenko, Jonathan Hurley, and Nate Cole.
> 
> 
> Bugs: AMBARI-11357
>     https://issues.apache.org/jira/browse/AMBARI-11357
> 
> 
> Repository: ambari
> 
> 
> Description
> -------
> 
> In HDP 2.3, Spark no longer needs Tez as a dependency.
> This means that
> 1. Spark metainfo.xml does not need to co-host Tez-client
> 2. Spark History Server restart or RU restart does not need to copy tez tarball to HDFS.
Note, this is still needed in HDP 2.2
> 3. Tez needs to be responsible for copying its own tarball to HDFS because its service
check depends on it.
> 
> 
> Diffs
> -----
> 
>   ambari-server/src/main/resources/common-services/SPARK/1.2.0.2.2/package/scripts/job_history_server.py
106b8cc 
>   ambari-server/src/main/resources/common-services/SPARK/1.2.0.2.2/package/scripts/spark_service.py
834e208 
>   ambari-server/src/main/resources/common-services/TEZ/0.4.0.2.1/package/scripts/service_check.py
98973ae 
>   ambari-server/src/main/resources/stacks/HDP/2.3/services/SPARK/metainfo.xml 205ccc9

>   ambari-server/src/test/python/stacks/2.2/SPARK/test_job_history_server.py 489037c 
> 
> Diff: https://reviews.apache.org/r/34618/diff/
> 
> 
> Testing
> -------
> 
> ----------------------------------------------------------------------
> Total run:742
> Total errors:0
> Total failures:0
> OK
> 
> Copied changes files to a cluster with HDP 2.3.0.0 and HDFS, YARN, MR, ZK.
> Then installed Spark, which did not require Tez, as expected.
> Then installed Tez, whose service checked passed.
> Then ran an RU and everything passed
> 
> 
> Thanks,
> 
> Alejandro Fernandez
> 
>


Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message