11 DBA Operations in an Oracle Database Vault Environment

Propagating Oracle Database Vault Policies to Other Databases

You can propagate Database Vault policies to other Database Vault-protected databases.

1. From Cloud Control, log into Oracle Database Vault Administrator as a user who has been granted the DV_OWNER or DV_ADMIN role.

   "Logging into Oracle Database Vault" explains how to log in.

2. In the Database Vault home page, under Policy Propagation, select Database Vault Policy Propagation.

   The Available Policies area in the Policy Propagation subpage lists a summary of the Oracle Database Vault policies that were created for the current database. From here, you can propagate these policies to another database.

3. Under Available Policies, select each policy that you want to propagate to another database.

   By default, all policies are selected.

   
   Description of the illustration dv_oem.gif
   
   

4. Under Destination Databases, click the Add button.

5. Under Search and Select: Database Vault Enabled Destination Databases, search for the destination databases, and then select each database to which you want to propagate the policies. Then click the Select button.

6. Under Destination Databases, do the following:

   1. Under Apply credentials across destination database(s), enter the user name and password of the administrator of the Database Vault database that contains the policies you want to propagate.

      This feature applies the Database Vault administrator's user name and password to all of the selected destination databases.

   2. Select each database to which you want to propagate the policies.

   3. Enter the Database Vault administrator user name and password for each database.

   4. Click the Apply button.

7. In the Propagate Options page, select from the following options.

   Any changes made to the seeded realms, command rules, rule sets, and so on will not be propagated to the destination databases. Only custom-created data are propagated.

   • Restore on failure: If the policy propagation encounters errors, then the propagation is rolled back. That is, the original policies on the destination database are restored. If you do not select this option, then the policy propagation on the destination database continues and ignores any errors.

   • Skip propagation if user defined policies exist: If the destination databases already have the user-defined policies, then the policy propagation is not attempted. If you do not select this option, then regardless of whether user-defined policies exist on the destination database, all the existing policies are cleared, and the policies from the source database are applied to the destination database.

   • Propagate Enterprise Manager metric thresholds for database vault metrics: If the source database has Oracle Database Vault metric thresholds set, then these thresholds are also propagated to the destination databases. If you do not select this option, then only policies are propagated and not the Oracle Database Vault thresholds.

8. Click the OK button.

9. In the Confirmation window, click OK.

   A message indicating success or failure appears. If the propagation succeeds, then the policies are active right away in their destination databases.

Enterprise Manager Cloud Control Alerts for Oracle Database Vault Policies

Cloud Control generates Oracle Database Vault-specific alerts. To view these alerts, you must be granted the DV_OWNER, DV_ADMIN, or DV_SECANALYST role.

The alerts are as follows:

• Database Vault Attempted Realm Violations. This alert helps the Oracle Database Vault security analyst (DV_SECANALYST role) to monitor violation attempts on the Database Vault database. This user can select the realms to be affected by the alert and filter these realms based on the different types of attempts by using error codes. You can enable this metric from the Metrics and Policy Settings page. By default, the attempted realm violations are collected every 24 hours.

• Database Vault Attempted Command Rule Violations. The functionality for this alert is the same as for Database Vault Attempted Realm Violations, except that it focuses on violations on command rules.

• Database Vault Realm Configuration Issues. This metric tracks and raises an alert if users misconfigure realms. This metric is enabled when you install Oracle Database vault, and by default it collects data every one hour.

• Database Vault Command Rule Configuration Issues. This functionality for this alert is that same as Database Vault Realm Configuration Issues, except that it focuses on configuration changes to command rules.

• Database Vault Policy Changes. This metric raises an alert on any change to any Database Vault policy, such as policies for realms and command rules. It provides a detailed policy changes report.

Oracle Database Vault-Specific Reports in Enterprise Manager Cloud Control

From the Database Vault home page, you can find information about the specific types of violations.

These violations are as follows:

• Top five attempted violations on realm and command rule

• Top five attempted violations by database users and client host

• Time series-based graphical reports on attempted violations for more detailed analysis

To have full access to the Database Vault reports, you must log into Database Vault Administrator as a user who has been granted the DV_OWNER, DV_ADMIN, or DV_SECANALYST role.

Changing the DBSNMP Account Password in an Oracle Database Vault Environment

Before you can change the password for the DBSNMP user account, you must revoke the DV_MONITOR role from this account.

In an Oracle Database Vault environment, the DBSNMP user account is granted the DV_MONITOR role. (The DBSNMP user can change his or her own password directly, without having to have the DV_MONITOR role revoked first.)

1. Log into the database instance using an account that has been granted the DV_OWNER role.

2. Revoke the DV_MONITOR role from the DBSNMP user account.

3. Connect as a user who has been granted the DV_ACCTMGR role and then change the DBSNMP user account password.

4. Connect as the DV_OWNER user and then grant the DV_MONITOR role back to the DBSNMP user account.

About Using Oracle Data Pump with Oracle Database Vault

Database administrators who want to use Oracle Data Pump must have Oracle Database Vault-specific authorization, in addition to the standard Oracle Data Pump privileges, if they want to export and import data in an Oracle Database Vault environment.

If these users want to perform Oracle Data Pump transportable tablespace operations, then they must have special authorization. You can check a user's authorizations for using Data Pump in an Oracle Database Vault environment by querying the DVSYS.DBA_DV_DATAPUMP_AUTH data dictionary view.

Authorizing Users for Oracle Data Pump Regular Export and Import Operations

You can grant a range of different authorization types for administrators who are responsible for performing regular export and import operations in a Database Vault environment.

Topics:

• About Authorizing Users for Oracle Data Pump Regular Operations

• Levels of Database Vault Authorization for Oracle Data Pump Regular Operations

• Authorizing Users for Oracle Data Pump Regular Operations in Database Vault

• Revoking Oracle Data Pump Authorization from Users

Authorizing Users for Oracle Data Pump Transportable Export and Import Operations

You can grant a variety of authorization levels for users who are responsible for using Oracle Data Pump transportable operations.

Topics:

• About Authorizing Users for Oracle Data Pump Transportable Operations

• Levels of Database Vault Authorization for Data Pump Transportable Operations

• Authorizing Users for Data Pump Transportable Operations in Database Vault

• Revoking Transportable Tablespace Authorization from Users

Guidelines for Exporting or Importing Data in an Oracle Database Vault Environment

After you have granted the database administrator who is using Oracle Data Pump the proper authorization, this user is ready to perform any export or import operations that are necessary.

Before this user begins work, he or she should follow these guidelines:

• Create a full backup of the database datafiles. This way, if you or other users do not like the newly-imported data, then you easily can revert the database to its previous state. This guideline is especially useful if an intruder had managed to modify Oracle Data Pump exported data to use his or her own policies.

• Decide how to handle exporting and importing multiple schemas or tables. You cannot specify multiple schemas or tables in the DBMS_MACADM.AUTHORIZE_DATAPUMP_USER procedure, but you can use either of the following methods to accomplish this task:

  • Run the DBMS_MACADM.AUTHORIZE_DATAPUMP_USER procedure for each schema or table, and then specify the list of these objects in the SCHEMAS or TABLES parameter of the EXPDP and IMPDP utilities.

  • Perform a full database export or import operation. If so, see the next guideline.

• When performing an export or import operation for an entire database, set the EXPDP or IMPDP FULL option to Y. Remember that this setting will capture the DVSYS schema, so ensure that the user has been granted the DV_OWNER role. For detailed information about Oracle Data Pump, see Oracle Database Utilities.

Note the following:

• You cannot use the legacy EXP and IMP utilities with the direct path option (direct=y) if Oracle Database Vault is enabled.

• Users who have been granted Database Vault-specific Oracle Data Pump authorization through the DBMS_MACADM.AUTHORIZE_DATAPUMP_USER procedure or transportable tablespace authorization through the DBMS_MACADM.AUTHORIZE_TTS_USER procedure can export and import database objects, but they cannot perform other activities, such as SELECT queries on schema tables to which they normally do not have access. Similarly, users are not permitted to perform Data Pump operations on objects outside the designated data objects.

• You must grant the DV_OWNER role to users who want to export or import an entire database, because a full database export requires access to the DVSYS schema, which stores the Oracle Database Vault policies. However, you cannot export the DVSYS schema itself. Data Pump only exports the protection definitions. The target database must have the DVSYS schema in it and Database Vault enabled before you can begin the import process.) Conversely, for a Data Pump import operation to apply the imported policies to the target database, it internally uses the DBMS_MACADM PL/SQL package, which in turn requires the Data Pump user to have the DV_OWNER role.

About Using Oracle Scheduler with Oracle Database Vault

The level of authorization that you must grant depends on schema in which the administrator wants to perform a task.

Possible scenarios are as follows:

• An administrator wants to schedule a job in his or her own schema. An administrator who has been granted privileges to schedule database jobs can continue to do so without any Oracle Database Vault-specific authorizations, unless this schema is protected by a realm. In that case, ensure that this user is authorized to access the realm. See "About Realm Authorization" for instructions on granting a user realm authorization.

• An administrator wants to run a job in another schema, but this job does not access any Oracle Database Vault realm or command rule protected object. In this case, this user only needs job related system privileges, not the Oracle Database Vault privileges.

• An administrator wants to run a job under the schema of another user, including any schema in the database or a remote database. If this job accesses an Oracle Database Vault realm or command rule protected object, then you must grant this user Database Vault-specific authorization by using the DBMS_MACADM.AUTHORIZE_SCHEDULER_USER procedure. This authorization applies to both background and foreground jobs. For background jobs, the authorization applies to the last user who created or modified the job. In addition, ensure that the schema owner (the protected schema in which the job is created) authorized to the realm.

  Later on, you can revoke this authorization by using the DBMS_MACADM.UNAUTHORIZE_SCHEDULER_USER procedure. If the schema is not protected by a realm, then you do not need to run the DBMS_MACADM.AUTHORIZE_SCHEDULER_USER procedure for the user.

Granting a Job Scheduling Administrator Authorization for Oracle Database Vault

You can authorize a user to schedule database jobs in a Database Vault environment.

1. Log into the database instance as a user who has been granted the DV_OWNER or DV_ADMIN role.

   Only a user who has been granted either of these roles can grant the necessary authorization.

2. Ensure that the user to whom you want to grant authorization has been granted system privileges to schedule database jobs.

   These privileges include any of the following: CREATE JOB, CREATE ANY JOB, CREATE EXTERNAL JOB, EXECUTE ANY PROGRAM, EXECUTE ANY CLASS, MANAGE SCHEDULER. The DBA and SCHEDULER_ADMIN roles provide these privileges; however, when Oracle Database Vault is enabled, the privileges are revoked from these roles.

   For example:

   SELECT GRANTEE, PRIVILEGE FROM DBA_SYS_PRIVS 
    WHERE PRIVILEGE IN ('CREATE JOB', 'CREATE ANY JOB');
   

3. Grant this user Oracle Database Vault authorization.

   For example, to authorize the user job_mgr to schedule jobs for any schema in the database:

   EXEC DBMS_MACADM.AUTHORIZE_SCHEDULER_USER('JOB_MGR');
   

   Optionally, you can restrict job_mgr's activities to a specific schema, as follows:

   EXEC DBMS_MACADM.AUTHORIZE_SCHEDULER_USER('JOB_MGR', 'HR');
   

   See "AUTHORIZE_SCHEDULER_USER Procedure" for detailed information about this procedure.

4. Ensure that the user has been authorized by querying the DVSYS.DBA_DV_JOB_AUTH data dictionary view as follows:

   SELECT GRANTEE,SCHEMA FROM DVSYS.DBA_DV_JOB_AUTH WHERE GRANTEE = 'user_name';
   

   See "DVSYS.DBA_DV_JOB_AUTH View" for more information about this view.

Revoking Authorization from Job Scheduling Administrators

You can revoke authorization from a user for scheduling database jobs.

1. Query the DVSYS.DBA_DV_JOB_AUTH data dictionary view to find the user's authorization.

   SELECT GRANTEE, SCHEMA FROM DVSYS.DBA_DV_JOB_AUTH WHERE GRANTEE='username';
   

2. Use the information you gathered from Step 1 to build the DBMS_MACADM.UNAUTHORIZE_SCHEDULER_USER command.

   For example:

   EXEC DBMS_MACADM.UNAUTHORIZE_SCHEDULER_USER('JOB_MGR');
   

   Ensure that this unauthorization complements the original authorization action. In other words, if you originally gave job_mgr authorization over the entire database, then the following command will not work:

   EXEC DBMS_MACADM.UNAUTHORIZE_SCHEDULER_USER('JOB_MGR', 'HR');
   

   See "UNAUTHORIZE_SCHEDULER_USER Procedure" for more information.

About Data Masking in an Oracle Database Vault Enabled Database

In an Oracle Database Vault-enabled database, only users who have Database Vault authorizations can mask data in Database Vault-protected database objects.

In a non-Database Vault environment, users who have been granted the SELECT_CATALOG_ROLE and DBA roles can perform data masking. However, with Database Vault, users must have additional privileges. This section describes three ways that you can use to enable users to mask data in Database Vault-protected objects.

If users do not have the correct privileges, then the following errors can occur while creating the masking definition or when the job is executing:

ORA-47400: Command Rule violation for string on string 

ORA-47401: Realm violation for string on string. 

ORA-47408: Realm violation for the EXECUTE command 

ORA-47409: Command Rule violation for the EXECUTE command 

ORA-01301: insufficient privileges

Adding Data Masking Users to the Data Dictionary Realm Authorizations

You can add data masking users to the Oracle Default Component Protection realm to give them data dictionary realm authorizations.

The Oracle Data Dictionary controls access to the Oracle Database catalog schemas, such as SYS and SYSTEM. (See "Default Realms" for a full list of these schemas.) It also controls the ability to grant system privileges and database administrator roles. If you add users to the Oracle Default Component Protection realm, and assuming these users already have the privileges associated with the Oracle Data Dictionary, then these users will have these same privileges in a Database Vault environment. Therefore, if you do add a user to this realm, ensure that this user is a trusted user.

• To add a user to the Oracle Default Component Protection realm, use the DBMS_MACADM.ADD_AUTH_TO_REALM procedure.

For example:

BEGIN
 DBMS_MACADM.ADD_AUTH_TO_REALM(
  realm_name   => 'Oracle Default Component Protection Realm', 
  grantee      => 'DBA_JSMITH', 
  auth_options => DBMS_MACUTL.G_REALM_AUTH_PARTICIPANT);
END;
/

Giving Users Access to Tables or Schemas That They Want to Mask

To give users access to tables or schemas that they want to mask, you must authorize them for the appropriate realm.

If the table or schema of a table that is to be data masked is in a realm, then you must add the user responsible for data masking to the realm authorization as a participant or owner. If the table or schema has dependent objects that are in other realm-protected tables, then you must grant the user participant or owner authorization for those realms as well.

• To authorize users for data masking to a realm that protects the objects they want to data mask, use the DBMS_MACADM.ADD_AUTH_TO_REALM procedure.

The following example shows how to grant user DBA_JSMITH authorization for the HR.EMPLOYEES table, which is protected by a realm called Business Apps Realm:

BEGIN
 DBMS_MACADM.ADD_AUTH_TO_REALM(
  realm_name   => 'Business Apps Realm', 
  grantee      => 'DBA_JSMITH', 
  auth_options => DBMS_MACUTL.G_REALM_AUTH_PARTICIPANT;
END;
/

Creating a Command Rule to Control Data Masking Privileges

For data masking, users must have the CREATE TABLE, SELECT TABLE, ALTER TABLE, and DROP TABLE privileges for the masking objects and if there are any dependent objects to be created, the user must have the appropriate privileges such as CREATE PACKAGE, CREATE TRIGGER, and so on.

You can create command rules to control data masking privileges at a granular level. To do so, create a command rule that can either prevent or allow the user access to objects that must have to be data masked. For example, you can create a command rule called Allow Data Masking that checks if the user is in a list of users who are responsible for data masking. If the user logging in is one of these users, then the command rule evaluates to true and the user is permitted to create the data mask for the protected object.

To create this type of command rule:

1. Create the rule set rule.

   For example:

   BEGIN
    DBMS_MACADM.CREATE_RULE(
     rule_name  => 'Is HDRISCOLL or DBA_JSMITH User', 
     rule_expr  =>'USER IN(''HDRISCOLL'',''DBA_JSMITH'')';
   END;
   /
   

2. Create a rule set and then add the rule to it:

   BEGIN
    DBMS_MACADM.CREATE_RULE_SET(
     rule_set_name    => 'Allow Data Masking', 
     description      => 'Allows users HDRISCOLL and DBA_JSMITH access', 
     enabled          => 'Y',
     eval_options     => 1,
     audit_options    => 1,
     fail_options     => 1,
     fail_message     => 'You do not have access to this object.',
     fail_code        => 20461,
     handler_options  => 0, 
     is_static        => TRUE);
   END;
   /
   BEGIN
    DBMS_MACADM.ADD_RULE_TO_RULE_SET(
     rule_set_name => 'Allow Data Masking', 
     rule_name     => 'Is HDRISCOLL or DBA_JSMITH User'),
     rule_order    => 1);
   END;
   /
   

3. Create a command rule and then add this rule to it:

   BEGIN
    DBMS_MACADM.CREATE_COMMAND_RULE(
     command         => 'CREATE TABLE', 
     rule_set_name   => 'Allow Data Masking', 
     object_owner    => 'HR', 
     object_name     => 'EMPLOYEES', 
     enabled         => DBMS_MACUTL.G_YES);
   END; 
   /