You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@ambari.apache.org by Dmytro Sen <ds...@hortonworks.com> on 2015/08/18 15:13:31 UTC

Review Request 37579: Ambari should not reset UIDs for existing users in a cluster

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/37579/
-----------------------------------------------------------

Review request for Ambari, Myroslav Papirkovskyy and Vitalyi Brodetskyi.


Bugs: AMBARI-12808
    https://issues.apache.org/jira/browse/AMBARI-12808


Repository: ambari


Description
-------

If a user wishes to set the UIDs for the users in an HDP cluster, due to some internal security requirements or standards, problems with rolling upgrades can occur, since the upgrade process will create files and directories with a UID of "1001" for "ambari-qa", even if the user had already set the UID for this user to a different value.
The setup_users() method in the following stack script:
ambari-server/src/main/resources/stacks/HDP/2.0.6/hooks/before-ANY/scripts/shared_initialization.py
will always override the UID of a given cluster user, even if the customer has already customized this.
The shared initialization scripts should check to see if a user has already been created with a separate UID, and provide a way to skip the process of setting the new UID. The Ambari scripts set the UID in order to comply with Hadoop and Kerberos standards, so any attempt to override this behavior should provide some config logging


Diffs
-----

  ambari-server/src/main/resources/common-services/HBASE/0.96.0.2.0/configuration/hbase-env.xml 213b95f 
  ambari-server/src/main/resources/stacks/HDP/2.0.6/configuration/cluster-env.xml 41140ab 
  ambari-server/src/main/resources/stacks/HDP/2.0.6/hooks/before-ANY/scripts/params.py 17c748c 
  ambari-server/src/main/resources/stacks/HDP/2.0.6/hooks/before-ANY/scripts/shared_initialization.py 35db022 
  ambari-server/src/test/python/stacks/2.0.6/configs/default.json 12cc74d 
  ambari-web/app/data/HDP2/site_properties.js 16182b1 
  ambari-web/app/utils/config.js ff20568 

Diff: https://reviews.apache.org/r/37579/diff/


Testing
-------

Unit tests passed


Thanks,

Dmytro Sen


Re: Review Request 37579: Ambari should not reset UIDs for existing users in a cluster

Posted by Dmytro Grinenko <ha...@gmail.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/37579/#review96037
-----------------------------------------------------------



ambari-server/src/main/java/org/apache/ambari/server/upgrade/UpgradeCatalog212.java (line 180)
<https://reviews.apache.org/r/37579/#comment151227>

    hbase-end?



ambari-server/src/main/java/org/apache/ambari/server/upgrade/UpgradeCatalog212.java (line 183)
<https://reviews.apache.org/r/37579/#comment151225>

    Why we use singletonMap? in most cases we use "HashMap"



ambari-server/src/main/java/org/apache/ambari/server/upgrade/UpgradeCatalog212.java (line 184)
<https://reviews.apache.org/r/37579/#comment151226>

    Why we use singletonMap? in most cases we use "HashMap"


- Dmytro Grinenko


On Aug. 21, 2015, 9:29 a.m., Dmytro Sen wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/37579/
> -----------------------------------------------------------
> 
> (Updated Aug. 21, 2015, 9:29 a.m.)
> 
> 
> Review request for Ambari, Myroslav Papirkovskyy and Vitalyi Brodetskyi.
> 
> 
> Bugs: AMBARI-12808
>     https://issues.apache.org/jira/browse/AMBARI-12808
> 
> 
> Repository: ambari
> 
> 
> Description
> -------
> 
> If a user wishes to set the UIDs for the users in an HDP cluster, due to some internal security requirements or standards, problems with rolling upgrades can occur, since the upgrade process will create files and directories with a UID of "1001" for "ambari-qa", even if the user had already set the UID for this user to a different value.
> The setup_users() method in the following stack script:
> ambari-server/src/main/resources/stacks/HDP/2.0.6/hooks/before-ANY/scripts/shared_initialization.py
> will always override the UID of a given cluster user, even if the customer has already customized this.
> The shared initialization scripts should check to see if a user has already been created with a separate UID, and provide a way to skip the process of setting the new UID. The Ambari scripts set the UID in order to comply with Hadoop and Kerberos standards, so any attempt to override this behavior should provide some config logging
> 
> 
> Diffs
> -----
> 
>   ambari-server/src/main/java/org/apache/ambari/server/upgrade/UpgradeCatalog212.java 6f20423 
>   ambari-server/src/main/resources/common-services/HBASE/0.96.0.2.0/configuration/hbase-env.xml 213b95f 
>   ambari-server/src/main/resources/stacks/HDP/2.0.6/configuration/cluster-env.xml 41140ab 
>   ambari-server/src/main/resources/stacks/HDP/2.0.6/hooks/before-ANY/scripts/params.py 17c748c 
>   ambari-server/src/main/resources/stacks/HDP/2.0.6/hooks/before-ANY/scripts/shared_initialization.py 35db022 
>   ambari-server/src/test/java/org/apache/ambari/server/upgrade/UpgradeCatalog212Test.java fae8985 
>   ambari-server/src/test/python/stacks/2.0.6/configs/default.json 12cc74d 
>   ambari-web/app/data/HDP2/site_properties.js 060e3b5 
>   ambari-web/app/utils/config.js abd1e08 
> 
> Diff: https://reviews.apache.org/r/37579/diff/
> 
> 
> Testing
> -------
> 
> Unit tests passed
> 
> 
> Thanks,
> 
> Dmytro Sen
> 
>


Re: Review Request 37579: Ambari should not reset UIDs for existing users in a cluster

Posted by Dmytro Grinenko <ha...@gmail.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/37579/#review96039
-----------------------------------------------------------

Ship it!


Ship It!

- Dmytro Grinenko


On Aug. 21, 2015, 10:50 a.m., Dmytro Sen wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/37579/
> -----------------------------------------------------------
> 
> (Updated Aug. 21, 2015, 10:50 a.m.)
> 
> 
> Review request for Ambari, Myroslav Papirkovskyy and Vitalyi Brodetskyi.
> 
> 
> Bugs: AMBARI-12808
>     https://issues.apache.org/jira/browse/AMBARI-12808
> 
> 
> Repository: ambari
> 
> 
> Description
> -------
> 
> If a user wishes to set the UIDs for the users in an HDP cluster, due to some internal security requirements or standards, problems with rolling upgrades can occur, since the upgrade process will create files and directories with a UID of "1001" for "ambari-qa", even if the user had already set the UID for this user to a different value.
> The setup_users() method in the following stack script:
> ambari-server/src/main/resources/stacks/HDP/2.0.6/hooks/before-ANY/scripts/shared_initialization.py
> will always override the UID of a given cluster user, even if the customer has already customized this.
> The shared initialization scripts should check to see if a user has already been created with a separate UID, and provide a way to skip the process of setting the new UID. The Ambari scripts set the UID in order to comply with Hadoop and Kerberos standards, so any attempt to override this behavior should provide some config logging
> 
> 
> Diffs
> -----
> 
>   ambari-server/src/main/java/org/apache/ambari/server/upgrade/UpgradeCatalog212.java 6f20423 
>   ambari-server/src/main/resources/common-services/HBASE/0.96.0.2.0/configuration/hbase-env.xml 213b95f 
>   ambari-server/src/main/resources/stacks/HDP/2.0.6/configuration/cluster-env.xml 41140ab 
>   ambari-server/src/main/resources/stacks/HDP/2.0.6/hooks/before-ANY/scripts/params.py 17c748c 
>   ambari-server/src/main/resources/stacks/HDP/2.0.6/hooks/before-ANY/scripts/shared_initialization.py 35db022 
>   ambari-server/src/test/java/org/apache/ambari/server/upgrade/UpgradeCatalog212Test.java fae8985 
>   ambari-server/src/test/python/stacks/2.0.6/configs/default.json 12cc74d 
>   ambari-web/app/data/HDP2/site_properties.js 060e3b5 
>   ambari-web/app/utils/config.js abd1e08 
> 
> Diff: https://reviews.apache.org/r/37579/diff/
> 
> 
> Testing
> -------
> 
> Unit tests passed
> 
> 
> Thanks,
> 
> Dmytro Sen
> 
>


Re: Review Request 37579: Ambari should not reset UIDs for existing users in a cluster

Posted by Dmytro Sen <ds...@hortonworks.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/37579/
-----------------------------------------------------------

(Updated Авг. 21, 2015, 10:50 д.п.)


Review request for Ambari, Myroslav Papirkovskyy and Vitalyi Brodetskyi.


Bugs: AMBARI-12808
    https://issues.apache.org/jira/browse/AMBARI-12808


Repository: ambari


Description
-------

If a user wishes to set the UIDs for the users in an HDP cluster, due to some internal security requirements or standards, problems with rolling upgrades can occur, since the upgrade process will create files and directories with a UID of "1001" for "ambari-qa", even if the user had already set the UID for this user to a different value.
The setup_users() method in the following stack script:
ambari-server/src/main/resources/stacks/HDP/2.0.6/hooks/before-ANY/scripts/shared_initialization.py
will always override the UID of a given cluster user, even if the customer has already customized this.
The shared initialization scripts should check to see if a user has already been created with a separate UID, and provide a way to skip the process of setting the new UID. The Ambari scripts set the UID in order to comply with Hadoop and Kerberos standards, so any attempt to override this behavior should provide some config logging


Diffs (updated)
-----

  ambari-server/src/main/java/org/apache/ambari/server/upgrade/UpgradeCatalog212.java 6f20423 
  ambari-server/src/main/resources/common-services/HBASE/0.96.0.2.0/configuration/hbase-env.xml 213b95f 
  ambari-server/src/main/resources/stacks/HDP/2.0.6/configuration/cluster-env.xml 41140ab 
  ambari-server/src/main/resources/stacks/HDP/2.0.6/hooks/before-ANY/scripts/params.py 17c748c 
  ambari-server/src/main/resources/stacks/HDP/2.0.6/hooks/before-ANY/scripts/shared_initialization.py 35db022 
  ambari-server/src/test/java/org/apache/ambari/server/upgrade/UpgradeCatalog212Test.java fae8985 
  ambari-server/src/test/python/stacks/2.0.6/configs/default.json 12cc74d 
  ambari-web/app/data/HDP2/site_properties.js 060e3b5 
  ambari-web/app/utils/config.js abd1e08 

Diff: https://reviews.apache.org/r/37579/diff/


Testing
-------

Unit tests passed


Thanks,

Dmytro Sen


Re: Review Request 37579: Ambari should not reset UIDs for existing users in a cluster

Posted by Dmytro Grinenko <ha...@gmail.com>.

> On Aug. 21, 2015, 9:56 a.m., Dmytro Grinenko wrote:
> > ambari-server/src/main/java/org/apache/ambari/server/upgrade/UpgradeCatalog212.java, line 181
> > <https://reviews.apache.org/r/37579/diff/2/?file=1046588#file1046588line181>
> >
> >     config.getProperties() could return null, in that case we get here NPE

According to us code, we assume that if config catalog exist, then properties should be present(at least empty map)


- Dmytro


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/37579/#review96036
-----------------------------------------------------------


On Aug. 21, 2015, 9:29 a.m., Dmytro Sen wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/37579/
> -----------------------------------------------------------
> 
> (Updated Aug. 21, 2015, 9:29 a.m.)
> 
> 
> Review request for Ambari, Myroslav Papirkovskyy and Vitalyi Brodetskyi.
> 
> 
> Bugs: AMBARI-12808
>     https://issues.apache.org/jira/browse/AMBARI-12808
> 
> 
> Repository: ambari
> 
> 
> Description
> -------
> 
> If a user wishes to set the UIDs for the users in an HDP cluster, due to some internal security requirements or standards, problems with rolling upgrades can occur, since the upgrade process will create files and directories with a UID of "1001" for "ambari-qa", even if the user had already set the UID for this user to a different value.
> The setup_users() method in the following stack script:
> ambari-server/src/main/resources/stacks/HDP/2.0.6/hooks/before-ANY/scripts/shared_initialization.py
> will always override the UID of a given cluster user, even if the customer has already customized this.
> The shared initialization scripts should check to see if a user has already been created with a separate UID, and provide a way to skip the process of setting the new UID. The Ambari scripts set the UID in order to comply with Hadoop and Kerberos standards, so any attempt to override this behavior should provide some config logging
> 
> 
> Diffs
> -----
> 
>   ambari-server/src/main/java/org/apache/ambari/server/upgrade/UpgradeCatalog212.java 6f20423 
>   ambari-server/src/main/resources/common-services/HBASE/0.96.0.2.0/configuration/hbase-env.xml 213b95f 
>   ambari-server/src/main/resources/stacks/HDP/2.0.6/configuration/cluster-env.xml 41140ab 
>   ambari-server/src/main/resources/stacks/HDP/2.0.6/hooks/before-ANY/scripts/params.py 17c748c 
>   ambari-server/src/main/resources/stacks/HDP/2.0.6/hooks/before-ANY/scripts/shared_initialization.py 35db022 
>   ambari-server/src/test/java/org/apache/ambari/server/upgrade/UpgradeCatalog212Test.java fae8985 
>   ambari-server/src/test/python/stacks/2.0.6/configs/default.json 12cc74d 
>   ambari-web/app/data/HDP2/site_properties.js 060e3b5 
>   ambari-web/app/utils/config.js abd1e08 
> 
> Diff: https://reviews.apache.org/r/37579/diff/
> 
> 
> Testing
> -------
> 
> Unit tests passed
> 
> 
> Thanks,
> 
> Dmytro Sen
> 
>


Re: Review Request 37579: Ambari should not reset UIDs for existing users in a cluster

Posted by Dmytro Grinenko <ha...@gmail.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/37579/#review96036
-----------------------------------------------------------



ambari-server/src/main/java/org/apache/ambari/server/upgrade/UpgradeCatalog212.java (line 181)
<https://reviews.apache.org/r/37579/#comment151224>

    config.getProperties() could return null, in that case we get here NPE


- Dmytro Grinenko


On Aug. 21, 2015, 9:29 a.m., Dmytro Sen wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/37579/
> -----------------------------------------------------------
> 
> (Updated Aug. 21, 2015, 9:29 a.m.)
> 
> 
> Review request for Ambari, Myroslav Papirkovskyy and Vitalyi Brodetskyi.
> 
> 
> Bugs: AMBARI-12808
>     https://issues.apache.org/jira/browse/AMBARI-12808
> 
> 
> Repository: ambari
> 
> 
> Description
> -------
> 
> If a user wishes to set the UIDs for the users in an HDP cluster, due to some internal security requirements or standards, problems with rolling upgrades can occur, since the upgrade process will create files and directories with a UID of "1001" for "ambari-qa", even if the user had already set the UID for this user to a different value.
> The setup_users() method in the following stack script:
> ambari-server/src/main/resources/stacks/HDP/2.0.6/hooks/before-ANY/scripts/shared_initialization.py
> will always override the UID of a given cluster user, even if the customer has already customized this.
> The shared initialization scripts should check to see if a user has already been created with a separate UID, and provide a way to skip the process of setting the new UID. The Ambari scripts set the UID in order to comply with Hadoop and Kerberos standards, so any attempt to override this behavior should provide some config logging
> 
> 
> Diffs
> -----
> 
>   ambari-server/src/main/java/org/apache/ambari/server/upgrade/UpgradeCatalog212.java 6f20423 
>   ambari-server/src/main/resources/common-services/HBASE/0.96.0.2.0/configuration/hbase-env.xml 213b95f 
>   ambari-server/src/main/resources/stacks/HDP/2.0.6/configuration/cluster-env.xml 41140ab 
>   ambari-server/src/main/resources/stacks/HDP/2.0.6/hooks/before-ANY/scripts/params.py 17c748c 
>   ambari-server/src/main/resources/stacks/HDP/2.0.6/hooks/before-ANY/scripts/shared_initialization.py 35db022 
>   ambari-server/src/test/java/org/apache/ambari/server/upgrade/UpgradeCatalog212Test.java fae8985 
>   ambari-server/src/test/python/stacks/2.0.6/configs/default.json 12cc74d 
>   ambari-web/app/data/HDP2/site_properties.js 060e3b5 
>   ambari-web/app/utils/config.js abd1e08 
> 
> Diff: https://reviews.apache.org/r/37579/diff/
> 
> 
> Testing
> -------
> 
> Unit tests passed
> 
> 
> Thanks,
> 
> Dmytro Sen
> 
>


Re: Review Request 37579: Ambari should not reset UIDs for existing users in a cluster

Posted by Dmytro Sen <ds...@hortonworks.com>.
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/37579/
-----------------------------------------------------------

(Updated Авг. 21, 2015, 9:29 д.п.)


Review request for Ambari, Myroslav Papirkovskyy and Vitalyi Brodetskyi.


Bugs: AMBARI-12808
    https://issues.apache.org/jira/browse/AMBARI-12808


Repository: ambari


Description
-------

If a user wishes to set the UIDs for the users in an HDP cluster, due to some internal security requirements or standards, problems with rolling upgrades can occur, since the upgrade process will create files and directories with a UID of "1001" for "ambari-qa", even if the user had already set the UID for this user to a different value.
The setup_users() method in the following stack script:
ambari-server/src/main/resources/stacks/HDP/2.0.6/hooks/before-ANY/scripts/shared_initialization.py
will always override the UID of a given cluster user, even if the customer has already customized this.
The shared initialization scripts should check to see if a user has already been created with a separate UID, and provide a way to skip the process of setting the new UID. The Ambari scripts set the UID in order to comply with Hadoop and Kerberos standards, so any attempt to override this behavior should provide some config logging


Diffs (updated)
-----

  ambari-server/src/main/java/org/apache/ambari/server/upgrade/UpgradeCatalog212.java 6f20423 
  ambari-server/src/main/resources/common-services/HBASE/0.96.0.2.0/configuration/hbase-env.xml 213b95f 
  ambari-server/src/main/resources/stacks/HDP/2.0.6/configuration/cluster-env.xml 41140ab 
  ambari-server/src/main/resources/stacks/HDP/2.0.6/hooks/before-ANY/scripts/params.py 17c748c 
  ambari-server/src/main/resources/stacks/HDP/2.0.6/hooks/before-ANY/scripts/shared_initialization.py 35db022 
  ambari-server/src/test/java/org/apache/ambari/server/upgrade/UpgradeCatalog212Test.java fae8985 
  ambari-server/src/test/python/stacks/2.0.6/configs/default.json 12cc74d 
  ambari-web/app/data/HDP2/site_properties.js 060e3b5 
  ambari-web/app/utils/config.js abd1e08 

Diff: https://reviews.apache.org/r/37579/diff/


Testing
-------

Unit tests passed


Thanks,

Dmytro Sen