-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/50594/#review144425
-----------------------------------------------------------
It doesnt look like you have verified ambari-server upgrade scenarios with the
patch. This is very important to validate. You need to ensure backward
compatibility is maintained
Please test the following scenarios
Scenario 1:
1. Install Ambari 2.1.0
2. Deploy HDP-2.3.2
3. Upgrade to Ambari 2.4.0
4. Restart SPARK
5. Verify SPARK restarts
6. Upgrade HDP to 2.3.4 and verify upgrade succeeds
7. Also upgrade to HDP 2.4+ and verify upgrade succeeds.
Scenario 2:
1. Install Ambari 2.1.0
2. Deploy HDP-2.3.2
3. Upgrade to Ambari 2.2.0
4. Restart Spark
5. Confirm that the Spark configs are not getting written correctly due to the
backward compatibility issue. Restart spark will cause issues.
6. Upgrade to Hotfix build 2.2.0.13-1
7. Update java_opts using following command
./configs.sh set <AMBARI-SERVER-HOSTNAME> <CLUSTERNAME>
spark-javaopts-properties content "-Dhdp.version=hdp_full_version"
8. Restart SPARK service
9. Verify that the Spark configs are getting written correctly.
10. Now upgrade Ambari to Ambari 2.4.0
11. Restart Spark
12. Verify SPARK restarts
13. Upgrade to HDP 2.3.4
14. Also upgrade to HDP-2.4+
- Jayush Luniya
On Aug. 1, 2016, 1:22 a.m., Saisai Shao wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/50594/
> -----------------------------------------------------------
>
> (Updated Aug. 1, 2016, 1:22 a.m.)
>
>
> Review request for Ambari, Jayush Luniya and Sumit Mohanty.
>
>
> Bugs: AMBARI-17954
> https://issues.apache.org/jira/browse/AMBARI-17954
>
>
> Repository: ambari
>
>
> Description
> -------
>
> Fix Spark hdp.version issues in upgrading and fresh install
>
>
> Diffs
> -----
>
>
> ambari-common/src/main/python/resource_management/libraries/functions/constants.py
> 42ecf6a
>
> ambari-server/src/main/java/org/apache/ambari/server/upgrade/UpgradeCatalog240.java
> 5495655
>
> ambari-server/src/main/resources/common-services/SPARK/1.2.1/configuration/spark-defaults.xml
> 646239e
>
> ambari-server/src/main/resources/common-services/SPARK/1.2.1/package/scripts/params.py
> 86aeb70
>
> ambari-server/src/main/resources/common-services/SPARK/1.2.1/package/scripts/setup_spark.py
> 2b5d36b
>
> ambari-server/src/main/resources/common-services/SPARK/1.5.2/configuration/spark-thrift-sparkconf.xml
> e27c986
>
> ambari-server/src/main/resources/stacks/HDP/2.0.6/properties/stack_features.json
> 6d26052
> ambari-server/src/main/resources/stacks/HDP/2.3/upgrades/config-upgrade.xml
> b295cc9
>
> ambari-server/src/main/resources/stacks/HDP/2.3/upgrades/nonrolling-upgrade-2.4.xml
> 5b20a72
>
> ambari-server/src/main/resources/stacks/HDP/2.3/upgrades/nonrolling-upgrade-2.5.xml
> 5d2cc54
> ambari-server/src/main/resources/stacks/HDP/2.3/upgrades/upgrade-2.4.xml
> c9f9823
> ambari-server/src/main/resources/stacks/HDP/2.3/upgrades/upgrade-2.5.xml
> d0332ed
>
> ambari-server/src/main/resources/stacks/HDP/2.4/services/SPARK/configuration/spark-defaults.xml
> PRE-CREATION
>
> ambari-server/src/main/resources/stacks/HDP/2.4/services/SPARK/configuration/spark-thrift-sparkconf.xml
> PRE-CREATION
>
> ambari-server/src/main/resources/stacks/HDP/2.5/services/SPARK/configuration/spark-defaults.xml
> b3d05d8
>
> ambari-server/src/test/java/org/apache/ambari/server/upgrade/UpgradeCatalog240Test.java
> 34ca199
> ambari-server/src/test/python/stacks/2.2/SPARK/test_job_history_server.py
> b2ed5df
> ambari-server/src/test/python/stacks/2.2/SPARK/test_spark_client.py ed331ea
> ambari-server/src/test/python/stacks/2.3/SPARK/test_spark_thrift_server.py
> a1abdfa
>
> Diff: https://reviews.apache.org/r/50594/diff/
>
>
> Testing
> -------
>
> Manual test with different scenarios:
>
> 1. Fresh install of HDP 2.3.6, 2.4.3, 2.5.0
> 2. Upgrade for 2.3.6 to 2.5.0.
> 3. Downgrade from 2.5.0 to 2.3.6.
>
>
> Thanks,
>
> Saisai Shao
>
>
