-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/50512/#review143771
-----------------------------------------------------------



I think the Spark and Spark2 design assumes the same keytab and user for Spark 
and Spark2 so that the transition between the two and concurrent use of both is 
seamless. So HDFS files, e.g for ATS data from both apps should be owned by the 
same user.

- bikassaha


On July 27, 2016, 10:10 a.m., Robert Levas wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/50512/
> -----------------------------------------------------------
> 
> (Updated July 27, 2016, 10:10 a.m.)
> 
> 
> Review request for Ambari, bikassaha, Saisai Shao, and Sumit Mohanty.
> 
> 
> Bugs: AMBARI-17921
>     https://issues.apache.org/jira/browse/AMBARI-17921
> 
> 
> Repository: ambari
> 
> 
> Description
> -------
> 
> If both Spark and Spark2 is installed and each run as a different user, then 
> the ACLs on the _shared_ keytab files may block access by components in 
> either service to needed keytab files. 
> 
> For example if Spark is set to run as the user with username `spark` and 
> Spark2 is set to run as the user with username `spark2`:
> ```
> spark-env/spark_user = spark
> spark2-env/spark_user = spark2
> ```
> 
> Then the keytab file for the shared headless principal - 
> spark.headless.keytab - will have an ACL set that either the spark or the 
> spark2 user can read it (depending on the order the keytab file is written). 
> 
> In this case, the following error will be encountered.... 
> 
> ```
> Traceback (most recent call last):
>   File 
> "/var/lib/ambari-agent/cache/common-services/SPARK/1.2.1/package/scripts/spark_thrift_server.py",
>  line 87, in <module>
>     SparkThriftServer().execute()
>   File 
> "/usr/lib/python2.6/site-packages/resource_management/libraries/script/script.py",
>  line 280, in execute
>     method(env)
>   File 
> "/var/lib/ambari-agent/cache/common-services/SPARK/1.2.1/package/scripts/spark_thrift_server.py",
>  line 54, in start
>     spark_service('sparkthriftserver', upgrade_type=upgrade_type, 
> action='start')
>   File 
> "/var/lib/ambari-agent/cache/common-services/SPARK/1.2.1/package/scripts/spark_service.py",
>  line 57, in spark_service
>     Execute(spark_kinit_cmd, user=params.spark_user)
>   File "/usr/lib/python2.6/site-packages/resource_management/core/base.py", 
> line 155, in __init__
>     self.env.run()
>   File 
> "/usr/lib/python2.6/site-packages/resource_management/core/environment.py", 
> line 160, in run
>     self.run_action(resource, action)
>   File 
> "/usr/lib/python2.6/site-packages/resource_management/core/environment.py", 
> line 124, in run_action
>     provider_action()
>   File 
> "/usr/lib/python2.6/site-packages/resource_management/core/providers/system.py",
>  line 273, in action_run
>     tries=self.resource.tries, try_sleep=self.resource.try_sleep)
>   File "/usr/lib/python2.6/site-packages/resource_management/core/shell.py", 
> line 71, in inner
>     result = function(command, **kwargs)
>   File "/usr/lib/python2.6/site-packages/resource_management/core/shell.py", 
> line 93, in checked_call
>     tries=tries, try_sleep=try_sleep)
>   File "/usr/lib/python2.6/site-packages/resource_management/core/shell.py", 
> line 141, in _call_wrapper
>     result = _call(command, **kwargs_copy)
>   File "/usr/lib/python2.6/site-packages/resource_management/core/shell.py", 
> line 294, in _call
>     raise Fail(err_msg)
> resource_management.core.exceptions.Fail: Execution of '/usr/bin/kinit -kt 
> /etc/security/keytabs/spark.headless.keytab 
> spark2rndygi0zfoo3ftqildwn5...@hwqe.hortonworks.com; ' returned 1. ######## 
> Hortonworks #############
> This is MOTD message, added for testing in qe infra
> kinit: Generic preauthentication failure while getting initial credentials
> ```
> 
> "kinit: Generic preauthentication failure while getting initial credentials" 
> indicates, in this case, the the user running the Spark service does not have 
> access to the specified keytab file.
> 
> To ensure this does not happen, keytab files for both services should have 
> different file names.
> 
> 
> Diffs
> -----
> 
>   ambari-server/src/main/resources/common-services/SPARK2/2.0.0/kerberos.json 
> 967adb0 
> 
> Diff: https://reviews.apache.org/r/50512/diff/
> 
> 
> Testing
> -------
> 
> Manualy tested
> 
> 
> Thanks,
> 
> Robert Levas
> 
>

Reply via email to