By default, the view records the most recent 10,000 events. However, you can change the number of recorded events by calling DBMS_LOGSTDBY.APPLY_SET() PL/SQL procedure. If SQL Apply stops unexpectedly, the reason for the problem is also recorded in this view.

The view can be customized to contain other information, such as which DDL transactions were applied and which were skipped. For example:

SQL> ALTER SESSION SET NLS_DATE_FORMAT  = 'DD-MON-YY HH24:MI:SS';
Session altered.
SQL> COLUMN STATUS FORMAT A60
SQL> SELECT EVENT_TIME, STATUS, EVENT FROM DBA_LOGSTDBY_EVENTS -
> ORDER BY EVENT_TIMESTAMP, COMMIT_SCN, CURRENT_SCN;

EVENT_TIME         STATUS
------------------------------------------------------------------------------
EVENT
-------------------------------------------------------------------------------
23-JUL-02 18:20:12 ORA-16111: log mining and apply setting up
23-JUL-02 18:25:12 ORA-16128: User initiated shut down successfully completed
23-JUL-02 18:27:12 ORA-16112: log mining and apply stopping
23-JUL-02 18:55:12 ORA-16128: User initiated shut down successfully completed
23-JUL-02 18:57:09 ORA-16111: log mining and apply setting up
23-JUL-02 20:21:47 ORA-16204: DDL successfully applied
create table hr.test_emp (empno number, ename varchar2(64))
23-JUL-02 20:22:55 ORA-16205: DDL skipped due to skip setting 
create database link link_to_boston connect to system identified by change_on_inst
7 rows selected.

This query shows that SQL Apply was started and stopped a few times. It also shows what DDL was applied and skipped.

For example:

SQL> COLUMN DICT_BEGIN FORMAT A10;
SQL> SET NUMF 99999999;
SQL> SELECT FILE_NAME, SEQUENCE# AS SEQ#, FIRST_CHANGE# AS F_SCN#, -
>  NEXT_CHANGE# AS N_SCN#, TIMESTAMP, -
>  DICT_BEGIN AS BEG, DICT_END AS END, -
>  THREAD# AS THR#, APPLIED FROM DBA_LOGSTDBY_LOG -
>  ORDER BY SEQUENCE#;

FILE_NAME                 SEQ# F_SCN    N_SCN TIMESTAM BEG END THR# APPLIED
------------------------- ---- ------- ------- -------- --- --- --- ---------
/oracle/dbs/hq_nyc_2.log  2     101579  101588 11:02:58 NO  NO  1     YES
/oracle/dbs/hq_nyc_3.log  3     101588  142065 11:02:02 NO  NO  1     YES
/oracle/dbs/hq_nyc_4.log  4     142065  142307 11:02:10 NO  NO  1     YES
/oracle/dbs/hq_nyc_5.log  5     142307  142739 11:02:48 YES YES 1     YES
/oracle/dbs/hq_nyc_6.log  6     142739  143973 12:02:10 NO  NO  1     YES
/oracle/dbs/hq_nyc_7.log  7     143973  144042 01:02:11 NO  NO  1     YES
/oracle/dbs/hq_nyc_8.log  8     144042  144051 01:02:01 NO  NO  1     YES
/oracle/dbs/hq_nyc_9.log  9     144051  144054 01:02:16 NO  NO  1     YES
/oracle/dbs/hq_nyc_10.log 10    144054  144057 01:02:21 NO  NO  1     YES
/oracle/dbs/hq_nyc_11.log 11    144057  144060 01:02:26 NO  NO  1  CURRENT
/oracle/dbs/hq_nyc_12.log 12    144060  144089 01:02:30 NO  NO  1  CURRENT
/oracle/dbs/hq_nyc_13.log 13    144089  144147 01:02:41 NO  NO  1       NO

The YES entries in the BEG and END columns indicate that a LogMiner dictionary build starts at log file sequence number 5. The most recent archived redo log file is sequence number 13, and it was received at the logical standby database at 01:02:41.The APPLIED column indicates that SQL Apply has applied all redo before SCN 144057. Since transactions can span multiple archived log files, multiple archived log files may show the value CURRENT in the APPLIED column.

This view provides information related to the failover characteristics of the logical standby database, including:

• The time to failover (apply finish time)

• How current is the committed data in the logical standby database (apply lag)

• What the potential data loss will be in the event of a disaster (transport lag).

For example:

SQL> COL NAME FORMAT A20
SQL> COL VALUE FORMAT A12
SQL> COL UNIT FORMAT A30
SQL> SELECT NAME, VALUE, UNIT FROM V$DATAGUARD_STATS;
 
NAME                 VALUE        UNIT
-------------------- ------------ ------------------------------
apply finish time    +00 00:00:00 day(2) to second(1) interval
apply lag            +00 00:00:00 day(2) to second(0) interval
transport lag        +00 00:00:00 day(2) to second(0) interval

This output is from a logical standby database that has received and applied all redo generated from the primary database.

This view provides information about the current state of the various processes involved with SQL Apply, including;

• Identifying information (sid | serial# | spid)

• SQL Apply process: COORDINATOR, READER, BUILDER, PREPARER, ANALYZER, or APPLIER (type)

• Status of the process's current activity (status_code | status)

• Highest redo record processed by this process (high_scn)

For example:

SQL> COLUMN SERIAL# FORMAT 9999
SQL> COLUMN SID FORMAT 9999
SQL> SELECT SID, SERIAL#, SPID, TYPE, HIGH_SCN FROM V$LOGSTDBY_PROCESS;
 
  SID   SERIAL#   SPID         TYPE            HIGH_SCN
  ----- -------   ----------- ---------------- ----------
   48        6    11074        COORDINATOR     7178242899
   56       56    10858        READER          7178243497
   46        1    10860        BUILDER         7178242901
   45        1    10862        PREPARER        7178243295
   37        1    10864        ANALYZER        7178242900
   36        1    10866        APPLIER         7178239467
   35        3    10868        APPLIER         7178239463
   34        7    10870        APPLIER         7178239461
   33        1    10872        APPLIER         7178239472
 
9 rows selected.

The HIGH_SCN column shows that the reader process is ahead of all other processes, and the PREPARER and BUILDER process ahead of the rest.

SQL> COLUMN STATUS FORMAT A40
SQL> SELECT TYPE, STATUS_CODE, STATUS FROM V$LOGSTDBY_PROCESS;
 
TYPE             STATUS_CODE STATUS
---------------- ----------- -----------------------------------------
COORDINATOR            16117 ORA-16117: processing
READER                 16127 ORA-16127: stalled waiting for additional
                             transactions to be applied
BUILDER                16116 ORA-16116: no work available
PREPARER               16116 ORA-16117: processing
ANALYZER               16120 ORA-16120: dependencies being computed for
                             transaction at SCN 0x0001.abdb440a
APPLIER                16124 ORA-16124: transaction 1 13 1427 is waiting
                             on another transaction
APPLIER                16121 ORA-16121: applying transaction with commit
                             SCN 0x0001.abdb4390
APPLIER                16123 ORA-16123: transaction 1 23  1231 is waiting
                             for commit approval
APPLIER                16116 ORA-16116: no work available

The output shows a snapshot of SQL Apply running. On the mining side, the READER process is waiting for additional memory to become available before it can read more, the PREPARER process is processing redo records, and the BUILDER process has no work available. On the apply side, the COORDINATOR is assigning more transactions to APPLIER processes, the ANALYZER is computing dependencies at SCN 7178241034, one APPLIER has no work available, while two have outstanding dependencies that are not yet satisfied.

This view provides detailed information regarding progress made by SQL Apply, including:

• SCN and time at which all transactions that have been committed on the primary database have been applied to the logical standby database (applied_scn, applied_time)

• SCN and time at which SQL Apply would begin reading redo records (restart_scn, restart_time) on restart

• SCN and time of the latest redo record received on the logical standby database (latest_scn, latest_time)

• SCN and time of the latest record processed by the BUILDER process (mining_scn, mining_time)

For example:

SQL> SELECT APPLIED_SCN, LATEST_SCN, MINING_SCN, RESTART_SCN -
> FROM V$LOGSTDBY_PROGRESS;
 
APPLIED_SCN  LATEST_SCN MINING_SCN RESTART_SCN
----------- ----------- ---------- -----------
 7178240496  7178240507 7178240507  7178219805

According to the output:

• SQL Apply has applied all transactions committed on or before SCN of 7178240496

• The latest redo record received at the logical standby database was generated at SCN 7178240507

• The mining component has processed all redo records generate on or before SCN 7178240507

• If SQL Apply stops and restarts for any reason, it will start mining redo records generated on or after SCN 7178219805

SQL> ALTER SESSION SET NLS_DATE_FORMAT='yy-mm-dd hh24:mi:ss';
Session altered
 
SQL> SELECT APPLIED_TIME, LATEST_TIME, MINING_TIME, RESTART_TIME - 
> FROM V$LOGSTDBY_PROGRESS;
 
APPLIED_TIME      LATEST_TIME       MINING_TIME       RESTART_TIME     
----------------- ----------------- ----------------- -----------------
05-05-12 10:38:21 05-05-12 10:41:53 05-05-12 10:41:21 05-05-12 10:09:30

According to the output:

• SQL Apply has applied all transactions committed on or before the time 05-05-12 10:38:21 (APPLIED_TIME)

• The last redo was generated at time 05-05-12 10:41:53 at the primary database (LATEST_TIME)

• The mining engine has processed all redo records generated on or before 05-05-12 10:41:21 (MINING_TIME)

• In the event of a restart, SQL Apply will start mining redo records generated after the time 05-05-12 10:09:30

This view provides a synopsis of the current state of SQL Apply, including:

• The DBID of the primary database (primary_dbid).

• The LogMiner session ID allocated to SQL Apply (session_id).

• Whether or not SQL Apply is applying in real time (realtime_apply).

For example:

SQL> COLUMN REALTIME_APPLY FORMAT a15
SQL> COLUMN STATE FORMAT a16
SQL> SELECT * FROM V$LOGSTDBY_STATE;

PRIMARY_DBID SESSION_ID REALTIME_APPLY  STATE
------------ ---------- --------------- ----------------
  1562626987          1 Y               APPLYING

The output shows that SQL Apply is running in the real-time apply mode and is currently applying redo data received from the primary database, the primary database's DBID is 1562626987 and the LogMiner session identifier associated the SQL Apply session is 1.

The V$LOGSTDBY_STATS view displays statistics, current state, and status information related to SQL Apply. No rows are returned from this view when SQL Apply is not running. This view is only meaningful in the context of a logical standby database.

For example:

 SQL> ALTER SESSION SET NLS_DATE_FORMAT='dd-mm-yyyy hh24:mi:ss';
 Session altered

 SQL> SELECT SUBSTR(name, 1, 40) AS NAME, SUBSTR(value,1,32) AS VALUE FROM V$LOGSTDBY_STATS;
 
 NAME                                     VALUE
 ---------------------------------------- --------------------------------
 logminer session id                      1
 number of preparers                      1
 number of appliers                       5
 server processes in use                  9
 maximum SGA for LCR cache (MB)           30
 maximum events recorded                  10000
 preserve commit order                    TRUE
 transaction consistency                  FULL
 record skipped errors                    Y
 record skipped DDLs                      Y
 record applied DDLs                      N
 record unsupported operations            N
 realtime apply                           Y
 apply delay (minutes)                    0
 coordinator state                        APPLYING
 coordinator startup time                 19-06-2007 09:55:47
 coordinator uptime (seconds)             3593
 txns received from logminer              56
 txns assigned to apply                   23
 txns applied                             22
 txns discarded during restart            33
 large txns waiting to be assigned        2
 rolled back txns mined                   4
 DDL txns mined                           40
 CTAS txns mined                          0
 bytes of redo mined                      60164040
 bytes paged out                          0
 pageout time (seconds)                   0
 bytes checkpointed                       4845
 checkpoint time (seconds)                0
 system idle time (seconds)               2921
 standby redo logs mined                  0
 archived logs mined                      5
 gap fetched logs mined                   0
 standby redo log reuse detected          1
 logfile open failures                    0
 current logfile wait (seconds)           0
 total logfile wait (seconds)             2910
 thread enable mined                      0
 thread disable mined                     0
 .
 40 rows selected. 

Initializing State

When you start SQL Apply by issuing an ALTER DATABASE START LOGICAL STANDBY APPLY statement, it goes into the initializing state.

To determine the current state of SQL Apply, query the V$LOGSTDBY_STATE view. For example:

SQL> SELECT SESSION_ID, STATE FROM V$LOGSTDBY_STATE;

SESSION_ID    STATE
----------    -------------
1             INITIALIZING

The SESSION_ID column identifies the persistent LogMiner session created by SQL Apply to mine the archived redo log files generated by the primary database.

Waiting for Dictionary Logs

The first time the SQL Apply is started, it needs to load the LogMiner dictionary captured in the redo log files. SQL Apply stays in the WAITING FOR DICTIONARY LOGS state until it has received all redo data required to load the LogMiner dictionary.

Loading Dictionary State

This loading dictionary state can persist for a while. Loading the LogMiner dictionary on a large database can take a long time. Querying the V$LOGSTDBY_STATE view returns the following output when loading the dictionary:

SQL> SELECT SESSION_ID, STATE FROM V$LOGSTDBY_STATE;

SESSION_ID    STATE
----------    ------------------
1             LOADING DICTIONARY

Only the COORDINATOR process and the mining processes are spawned until the LogMiner dictionary is fully loaded. Therefore, if you query the V$LOGSTDBY_PROCESS at this point, you do not see any of the APPLIER processes. For example:

SQL> SELECT SID, SERIAL#, SPID, TYPE FROM V$LOGSTDBY_PROCESS;

SID     SERIAL#     SPID       TYPE
------  ---------   ---------  ---------------------
47      3           11438      COORDINATOR
50      7           11334      READER
45      1           11336      BUILDER
44      2           11338      PREPARER
43      2           11340      PREPARER

You can get more detailed information about the progress in loading the dictionary by querying the V$LOGMNR_DICTIONARY_LOAD view. The dictionary load happens in three phases:

1. The relevant archived redo log files or standby redo logs files are mined to gather the redo changes relevant to load the LogMiner dictionary.

2. The changes are processed and loaded in staging tables inside the database.

3. The LogMiner dictionary tables are loaded by issuing a series of DDL statements.

For example:

SQL> SELECT PERCENT_DONE, COMMAND -
> FROM V$LOGMNR_DICTIONARY_LOAD -
> WHERE SESSION_ID = (SELECT SESSION_ID FROM V$LOGSTDBY_STATE);

PERCENT_DONE     COMMAND
-------------    -------------------------------
40               alter table SYSTEM.LOGMNR_CCOL$ exchange partition 
                 P101 with table SYS.LOGMNRLT_101_CCOL$ excluding
                 indexes without validation

If the PERCENT_DONE or the COMMAND column does not change for a long time, query the V$SESSION_LONGOPS view to monitor the progress of the DDL transaction in question.

Applying State

In this state, SQL Apply has successfully loaded the initial snapshot of the LogMiner dictionary, and is currently applying redo data to the logical standby database.

For detailed information about the SQL Apply progress, query the V$LOGSTDBY_PROGRESS view:

SQL> ALTER SESSION SET NLS_DATE_FORMAT = 'DD-MON-YYYY HH24:MI:SS';
SQL> SELECT APPLIED_TIME, APPLIED_SCN, MINING_TIME, MINING_SCN -
> FROM V$LOGSTDBY_PROGRESS;

APPLIED_TIME            APPLIED_SCN   MINING_TIME           MINING_SCN
--------------------    -----------   --------------------  -----------
10-JAN-2005 12:00:05    346791023     10-JAN-2005 12:10:05  3468810134

All committed transactions seen at or before APPLIED_SCN (or APPLIED_TIME) on the primary database have been applied to the logical standby database. The mining engine has processed all redo records generated at or before MINING_SCN (and MINING_TIME) on the primary database. At steady state, the value of MINING_SCN (and MINING_TIME) is always ahead of APPLIED_SCN (and APPLIED_TIME).

Waiting On Gap State

This state occurs when SQL Apply has mined and applied all available redo records, and is waiting for a new log file (or a missing log file) to be archived by the RFS process.

SQL> SELECT STATUS FROM V$LOGSTDBY_PROCESS WHERE TYPE = 'READER';

STATUS
------------------------------------------------------------------------
ORA-16240: waiting for log file (thread# 1, sequence# 99)

Idle State

SQL Apply enters this state once it has applied all redo generated by the primary database.

By default the last 10,000 events are remembered in the event view. You can change the number of events logged by invoking the DBMS_LOGSTDBY.APPLY_SET procedure. For example, to ensure that the last 100,000 events are recorded, you can issue the following statement:

SQL> EXECUTE DBMS_LOGSTDBY.APPLY_SET ('MAX_EVENTS_RECORDED', '100000');

Errors that cause SQL Apply to stop are always recorded in the DBA_LOGSTDBY_EVENTS view (unless there is insufficient space in the SYSTEM tablespace). These events are always put into the alert file as well, with the keyword LOGSTDBY included in the text. When querying the view, select the columns in order by EVENT_TIME, COMMIT_SCN, and CURRENT_SCN. This ordering ensures a shutdown failure appears last in the view.

The following examples show DBMS_LOGSTDBY subprograms that specify events to be recorded in the view.

Example 1: Determining if DDL Statements Have Been Applied

For example, to record applied DDL transactions to the DBA_LOGSTDBY_EVENTS view, issue the following statement:

SQL> EXECUTE DBMS_LOGSTDBY.APPLY_SET ('RECORD_APPLIED_DDL', 'TRUE');

Example 2: Checking the DBA_LOGSTDBY_EVENTS View for Unsupported Operations

To capture information about transactions running on the primary database that are not supported by a logical standby database, issue the following statements:

SQL> ALTER DATABASE STOP LOGICAL STANDBY APPLY;SQL> EXEC DBMS_LOGSTDBY.APPLY_SET('RECORD_UNSUPPORTED_OPERATIONS', 'TRUE');SQL> ALTER DATABASE START LOGICAL STANDBY APPLY IMMEDIATE; 

Then, check the DBA_LOGSTDBY_EVENTS view for any unsupported operations. Usually, an operation on an unsupported table is silently ignored by SQL Apply. However, during rolling upgrade (while the standby database is at a higher version and mining redo generated by a lower versioned primary database), if you performed an unsupported operation on the primary database, the logical standby database may not be the one to which you want to perform a switchover. Oracle Data Guard logs at least one unsupported operation per table in the DBA_LOGSTDBY_EVENTS view. Using SQL Apply to Upgrade the Oracle Database provides detailed information about rolling upgrades.

You can change the default behavior by specifying rules to skip applying modifications to specific tables. For example, to omit changes to the HR.EMPLOYEES table, you can specify rules to prevent application of DML and DDL changes to the specific table. For example:

For example, if the file system organization in the logical standby database is different than that in the primary database, you can write a DBMS_LOGSTDBY.SKIP procedure to transparently handle DDL transactions with file specifications.

The following procedure can handle different file system organization between the primary database and standby database, as long as you use a specific naming convention for your file-specification string.

You can also use this procedure to enable SQL Apply on a table that was formerly skipped.

Before you can create a table, it must meet the requirements described in Ensure Table Rows in the Primary Database Can Be Uniquely Identified. Then, you can use the following steps to re-create a table named HR.EMPLOYEES and resume SQL Apply. The directions assume that there is already a database link BOSTON defined to access the primary database.

The following list shows how to re-create a table and restart SQL Apply on that table:

1. Stop SQL Apply:

   SQL> ALTER DATABASE STOP LOGICAL STANDBY APPLY;
   

2. Ensure no operations are being skipped for the table in question by querying the DBA_LOGSTDBY_SKIP view:

   SQL> SELECT * FROM DBA_LOGSTDBY_SKIP;
   
   ERROR  STATEMENT_OPT        OWNER          NAME                PROC
   -----  -------------------  -------------  ----------------    -----
   N      SCHEMA_DDL           HR             EMPLOYEES
   N      DML                  HR             EMPLOYEES
   N      SCHEMA_DDL           OE             TEST_ORDER
   N      DML                  OE             TEST_ORDER
   

   Because you already have skip rules associated with the table that you want to re-create on the logical standby database, you must first delete those rules. You can accomplish that by calling the DBMS_LOGSTDBY.UNSKIP procedure. For example:

   SQL> EXECUTE DBMS_LOGSTDBY.UNSKIP(stmt => 'DML', -
   > schema_name => 'HR', -
   > object_name => 'EMPLOYEES');
   

   SQL> EXECUTE DBMS_LOGSTDBY.UNSKIP(stmt => 'SCHEMA_DDL', -
   > schema_name => 'HR', -
   > object_name => 'EMPLOYEES');
   

3. Re-create the table HR.EMPLOYEES with all its data in the logical standby database by using the DBMS_LOGSTDBY.INSTANTIATE_TABLE procedure. For example:

   SQL> EXECUTE DBMS_LOGSTDBY.INSTANTIATE_TABLE(schema_name => 'HR', -
   > table_name => 'EMPLOYEES', -
   > dblink => 'BOSTON');
   

4. Start SQL Apply:

   SQL> ALTER DATABASE START LOGICAL STANDBY APPLY IMMEDIATE;
   

   See Also:

   Oracle Database PL/SQL Packages and Types Reference for information about the DBMS_LOGSTDBY.UNSKIP and the DBMS_LOGSTDBY.INSTANTIATE_TABLE procedures

To ensure a consistent view across the newly instantiated table and the rest of the database, wait for SQL Apply to catch up with the primary database before querying this table. You can do this by performing the following steps:

Perform the following steps:

For logical standby databases, no manual intervention is required if the standby database did not apply redo data past the new resetlogs SCN (past the start of the new branch of redo data)

The following table describes how to resynchronize the standby database with the primary database branch.

See Oracle Database Backup and Recovery User's Guide for more information about database incarnations, recovering through an OPEN RESETLOGS operation, and Flashback Database.

On the logical standby database, create an index on the columns that make up the primary key. The following query generates a list of tables with no index information that can be used by a logical standby database to apply to uniquely identify rows. By creating an index on the following tables, performance can be improved significantly.

SQL> SELECT OWNER, TABLE_NAME FROM DBA_TABLES -
> WHERE OWNER NOT IN (SELECT OWNER FROM DBA_LOGSTDBY_SKIP -
> WHERE STATEMENT_OPT = 'INTERNAL SCHEMA') -
> MINUS -
> SELECT DISTINCT TABLE_OWNER, TABLE_NAME FROM DBA_INDEXES -
> WHERE INDEX_TYPE NOT LIKE ('FUNCTION-BASED%') -
> MINUS -
> SELECT OWNER, TABLE_NAME FROM DBA_LOGSTDBY_UNSUPPORTED;

You can add a rely primary key constraint to a table on the primary database, as follows:

To determine whether increasing memory allocated to the LCR cache would be beneficial, perform the following steps:

Ideally, the pageout activity should not consume more than 5 percent of the total uptime. If you continue to take snapshots over an extended interval and you find the pageout activities continue to consume a significant portion of the apply time, increasing the memory size may provide some benefits. You can increase the memory allocated to SQL Apply by setting the memory allocated to LCR cache (for this example, the SGA is set to 1 GB):

SQL> EXECUTE DBMS_LOGSTDBY.APPLY_SET('MAX_SGA', 1024);
PL/SQL procedure successfully completed

The strict default order of committing transactions allows any application to run transparently on the logical standby database.

However, many applications do not require such strict ordering among all transactions. Such applications do not require transactions containing non-overlapping sets of rows to be committed in the same order that they were committed at the primary database. This less strict ordering typically results in higher apply rates at the logical standby database. You can change the default order of committing transactions by performing the following steps:

1. Stop SQL Apply:

   SQL> ALTER DATABASE STOP LOGICAL STANDBY APPLY;
   Database altered
   

2. Issue the following to allow transactions to be applied out of order from how they were committed on the primary databases:

   SQL> EXECUTE DBMS_LOGSTDBY.APPLY_SET('PRESERVE_COMMIT_ORDER', 'FALSE');
   PL/SQL procedure successfully completed
   

3. Start SQL Apply:

   SQL> ALTER DATABASE START LOGICAL STANDBY APPLY IMMEDIATE;
   Database altered
   

You can change back the apply mode as follows: