To create a replay schedule:

1. Use the BEGIN_REPLAY_SCHEDULE procedure:

   DBMS_WORKLOAD_REPLAY.BEGIN_REPLAY_SCHEDULE (
      schedule_name  IN VARCHAR2);
   

2. Set the schedule_name parameter to the name of this replay schedule.

This example shows how to create a replay schedule named peak_schedule.

EXEC DBMS_WORKLOAD_REPLAY.BEGIN_REPLAY_SCHEDULE ('peak_schedule');

To add workload captures to a replay schedule:

1. Use the ADD_CAPTURE function:

   DBMS_WORKLOAD_REPLAY.ADD_CAPTURE (
      capture_dir_name    IN VARCHAR2,
      start_delay_seconds IN NUMBER  DEFAULT 0,
      stop_replay         IN BOOLEAN DEFAULT FALSE,
      take_begin_snapshot IN BOOLEAN DEFAULT FALSE,
      take_end_snapshot   IN BOOLEAN DEFAULT FALSE,
      query_only          IN BOOLEAN DEFAULT FALSE)
   RETURN NUMBER;
   

   DBMS_WORKLOAD_REPLAY.ADD_CAPTURE (
      capture_dir_name    IN VARCHAR2,
      start_delay_seconds IN NUMBER,
      stop_replay         IN VARCHAR2,
      take_begin_snapshot IN VARCHAR2 DEFAULT 'N',
      take_end_snapshot   IN VARCHAR2 DEFAULT 'N',
      query_only          IN VARCHAR2 DEFAULT 'N')
   RETURN NUMBER;
   

   This function returns an unique identifier that identifies the workload capture in this replay schedule.

   See:

   "About Query-Only Database Replay" for information about query-only database replays.

   Note:

   Query-only database replays are meant to be used and executed in test environments only.

   • Do not use query-only database replays on production systems.

   • Divergence is expected during query-only database replays.

2. Set the capture_dir_name parameter to the name of the directory object that points to the workload capture under the top-level replay directory.

   The directory must contain a valid workload capture that is preprocessed on a system running the same database version as the replay system.

3. Set the other parameters, which are optional, as appropriate.

   For information about these parameters, see Oracle Database PL/SQL Packages and Types Reference.

The following example shows how to add a workload capture named peak_wkld to a replay schedule by using the ADD_CAPTURE function in a SELECT statement.

SELECT DBMS_WORKLOAD_REPLAY.ADD_CAPTURE ('peak_wkld')
  FROM dual;

You can also use the DBMS_WORKLOAD_REPLAY package to remove workload captures from a replay schedule.

To remove workload captures from a replay schedule:

1. Use the REMOVE_CAPTURE procedure:

   DBMS_WORKLOAD_REPLAY.REMOVE_CAPTURE (
      schedule_capture_number IN NUMBER);
   

2. Set the schedule_capture_number parameter to the unique identifier that identifies the workload capture in this replay schedule.

   The unique identifier is the same identifier that was returned by the ADD_CAPTURE function when the workload capture was added to the replay schedule.

To add schedule orders to a replay schedule:

1. Use the ADD_SCHEDULE_ORDERING function:

   DBMS_WORKLOAD_REPLAY.ADD_SCHEDULE_ORDERING (
      schedule_capture_id IN NUMBER,
      waitfor_capture_id IN NUMBER)
   RETURN NUMBER;
   

   This function adds a schedule order between two workload captures that have been added to the replay schedule. If a schedule order cannot be added, it returns a nonzero error code.

2. Set the schedule_capture_id parameter to the workload capture that you want to wait in this schedule order.

3. Set the wait_for_capture_id parameter to the workload capture that you want to be completed before the other workload capture can start in this schedule order.

You can also use the DBMS_WORKLOAD_REPLAY package to remove schedule orders from a replay schedule.

To remove schedule orders from a replay schedule:

1. Use the REMOVE_SCHEDULE_ORDERING procedure:

   DBMS_WORKLOAD_REPLAY.REMOVE_SCHEDULE ORDERING (
      schedule_capture_id IN VARCHAR2,
      wait_for_capture_id IN VARCHAR2);
   

2. Set the schedule_capture_id parameter to the workload capture waiting in this schedule order.

3. Set the wait_for_capture_id parameter to the workload capture that needs to be completed before the other workload capture can start in this schedule order.

To view schedule orders:

• Use the DBA_WORKLOAD_SCHEDULE_ORDERING view.

To save a replay schedule:

• Use the END_REPLAY_SCHEDULE procedure:

  DBMS_WORKLOAD_REPLAY.END_REPLAY_SCHEDULE;
  

  This procedure completes the creation of a replay schedule. The replay schedule is saved and associated with the replay directory. Once a replay schedule is saved, you can use it for a consolidated replay.

To view replay schedules:

• Use the DBA_WORKLOAD_REPLAY_SCHEDULES view.

Initializing the replay data performs the following operations:

• Puts the database state in initialized mode for the replay of a consolidated workload.

• Points to the replay directory that contains all workload captures participating in the replay schedule.

• Loads the necessary metadata into tables required for replay.

  For example, captured connection strings are loaded into a table where they can be remapped for replay.

To initialize Consolidated Database Replay:

1. Use the INITIALIZE_CONSOLIDATED_REPLAY procedure:

   DBMS_WORKLOAD_REPLAY.INITIALIZE_CONSOLIDATED_REPLAY (
      replay_name    IN VARCHAR2,
      schedule_name  IN VARCHAR2,
      plsql_mode     IN VARCHAR2 DEFAULT 'TOP_LEVEL');
   

2. Set the replay_name parameter to the name of the consolidated replay.

3. Set the schedule_name parameter to the name of the replay schedule to use.

   The schedule_name parameter is the name of the replay schedule used during its creation, as described in "Creating Replay Schedules Using APIs".

The optional plsql_mode parameter specifies the PL/SQL replay mode.

These two values can be set for the plsql_mode parameter:

• top_level: Only top-level PL/SQL calls. This is the default value.

• extended: SQL executed inside PL/SQL or top-level PL/SQL if there is no SQL recorded inside. All captures must have been done in the ‘extended’ PL/SQL mode. Non-PL/SQL calls will be replayed in the usual manner.

The following example shows how to initialize a consolidated replay named peak_replay using the replay schedule named peak_schedule.

EXEC DBMS_WORKLOAD_REPLAY.INITIALIZE_CONSOLIDATED_REPLAY ('peak_replay',
     'peak_schedule');

To remap connection strings:

1. Use the REMAP_CONNECTION procedure:

   DBMS_WORKLOAD_REPLAY.REMAP_CONNECTION (
      schedule_cap_id   IN NUMBER,
      connection_id     IN NUMBER,
      replay_connection IN VARCHAR2);
   

   This procedure remaps the captured connection to a new connection string for all participating workload captures in the replay schedule.

2. Set the schedule_capture_id parameter to a participating workload capture in the current replay schedule.

   The schedule_capture_id parameter is the unique identifier returned when adding the workload capture to the replay schedule, as described in "Adding Workload Captures to Replay Schedules Using APIs".

3. Set the connection_id parameter to the connection to be remapped.

   The connection_id parameter is generated when replay data is initialized and corresponds to a connection from the workload capture.

4. Set the replay_connection parameter to the new connection string that will be used during replay.

To remap users:

1. Use the SET_USER_MAPPING procedure:

   DBMS_WORKLOAD_REPLAY.SET_USER_MAPPING (
      schedule_cap_id IN NUMBER,
      capture_user    IN VARCHAR2,
      replay_user     IN VARCHAR2);
   

2. Set the schedule_capture_id parameter to a participating workload capture in the current replay schedule.

   The schedule_capture_id parameter is the unique identifier returned when adding the workload capture to the replay schedule, as described in "Adding Workload Captures to Replay Schedules Using APIs".

3. Set the capture_user parameter to the username of the user or schema captured during the time of the workload capture.

4. Set the replay_user parameter to the username of a new user or schema to which the captured user is remapped during replay.

   If this parameter is set to NULL, then the mapping is disabled.

This example shows how to remap the PROD user used during capture to the TEST user during replay for the workload capture identified as 1001.

EXEC DBMS_WORKLOAD_REPLAY.SET_USER_MAPPING (1001, 'PROD', 'TEST');

Preparing a consolidated replay performs the following operations:

• Specifies the replay options, such as synchronization mode, session connection rate, and session request rate.

• Puts the database state in replay mode.

• Enables the start of replay clients.

To prepare a consolidated replay:

• Use the PREPARE_CONSOLIDATED_REPLAY procedure:

  DBMS_WORKLOAD_REPLAY.PREPARE_CONSOLIDATED_REPLAY (
     synchronization         IN VARCHAR2 DEFAULT 'SCN',
     connect_time_scale      IN NUMBER   DEFAULT 100,
     think_time_scale        IN NUMBER   DEFAULT 100,
     think_time_auto_correct IN BOOLEAN  DEFAULT TRUE,
     capture_sts             IN BOOLEAN  DEFAULT FALSE,
     sts_cap_interval        IN NUMBER   DEFAULT 300);
  

  For information about these parameters and how to set them, see "Specifying Replay Options".

To start a consolidated replay:

• Use the START_CONSOLIDATED_REPLAY procedure:

  DBMS_WORKLOAD_REPLAY.START_CONSOLIDATED_REPLAY;