Opening a Lagging Standby Database (to verify data ?)

 As shown in my previous blog post, you can create a Standby Database that lags the Primary by not applying Redo immediately but "waiting" for a specified interval.  It continues to receive and  ArchiveLogs but simply applies each only after the "wait interval".

So, first, the status at the Primary:

oracle19c>echo $ORACLE_SID
ORCLCDB
oracle19c>sqlplus '/ as sysdba'

SQL*Plus: Release 19.0.0.0.0 - Production on Sat Jul 24 18:03:12 2021
Version 19.3.0.0.0

Copyright (c) 1982, 2019, Oracle.  All rights reserved.


Connected to:
Oracle Database 19c Enterprise Edition Release 19.0.0.0.0 - Production
Version 19.3.0.0.0

SQL> select systimestamp, database_role, current_scn from v$database;

SYSTIMESTAMP                                                   DATABASE_ROLE    CURRENT_SCN
-------------------------------------------------------------- ---------------- -----------
24-JUL-21 06.03.32.106863 PM +08:00                            PRIMARY             13258062

SQL> alter session set container=ORCLPDB1;

Session altered.

SQL> select * from hemant.job_tracking_tbl order by 1;

JOB_END_TIME                                                       JOB_ID     DB_SCN
-------------------------------------------------------------- ---------- ----------
24-JUL-21 04.57.21.676949 PM                                            1   13239134
24-JUL-21 05.04.29.870877 PM                                            2   13241261
24-JUL-21 05.30.17.962275 PM                                            3   13246616
24-JUL-21 05.39.10.912969 PM                                            4   13247859
24-JUL-21 05.40.20.865467 PM                                            5   13248159
24-JUL-21 05.50.23.930352 PM                                            6   13252182
24-JUL-21 06.00.27.037797 PM                                            7   13257658

7 rows selected.

SQL> archive log list;
Database log mode              Archive Mode
Automatic archival             Enabled
Archive destination            /opt/oracle/archivelog/ORCLCDB
Oldest online log sequence     310
Next log sequence to archive   312
Current log sequence           312
SQL>

Now, the status at STDB2 (the Standby that is lagging with an enforced 60minutes delay

From the alert.log :
2021-07-24T17:51:15.716720+08:00
PR00 (PID:2924): Media Recovery Log /opt/oracle/archivelog/STDB21_303_1036108814.dbf
PR00 (PID:2924): Media Recovery Delayed for 59 minute(s) T-1.S-304
2021-07-24T17:57:26.299295+08:00
PR00 (PID:2924): Media Recovery Log /opt/oracle/archivelog/STDB21_304_1036108814.dbf
2021-07-24T17:57:44.580258+08:00
PR00 (PID:2924): Media Recovery Delayed for 60 minute(s) T-1.S-305
2021-07-24T18:00:32.550708+08:00
 rfs (PID:3452): Archived Log entry 52 added for B-1036108814.T-1.S-311 ID 0xa7521ccd LAD:3
2021-07-24T18:00:33.444329+08:00
 rfs (PID:3452): Selected LNO:4 for T-1.S-312 dbid 2778483057 branch 1036108814


oracle19c>sqlplus '/ as sysdba'

SQL*Plus: Release 19.0.0.0.0 - Production on Sat Jul 24 18:05:53 2021
Version 19.3.0.0.0

Copyright (c) 1982, 2019, Oracle.  All rights reserved.


Connected to:
Oracle Database 19c Enterprise Edition Release 19.0.0.0.0 - Production
Version 19.3.0.0.0

SQL> alter database recover managed standby database cancel;

Database altered.

SQL> alter database open read only;

Database altered.

SQL> select systimestamp, database_role, open_mode, current_scn from v$database;

SYSTIMESTAMP                                        DATABASE_ROLE    OPEN_MODE    CURRENT_SCN
--------------------------------------------------- ---------------- ------------ -----------
24-JUL-21 06.06.51.313616 PM +08:00                 PHYSICAL STANDBY READ ONLY       13239390

SQL>
SQL> alter pluggable database orclpdb1 open read only;

Pluggable database altered.

SQL> alter session set container=ORCLPDB1;

Session altered.

SQL> select * from hemant.job_tracking_tbl order by 1;

CURRENT_TIMESTAMP                                                  JOB_ID CURRENT_SCN
-------------------------------------------------------------- ---------- -----------
24-JUL-21 04.57.21.676949 PM                                            1    13239134

SQL>

The Primary database is at Log Sequence#312. This Standby has applied only Sequence#304. Let me resume Recovery for some more time and then check the Standby again.

SQL> exit
Disconnected from Oracle Database 19c Enterprise Edition Release 19.0.0.0.0 - Production
Version 19.3.0.0.0
oracle19c>echo $ORACLE_SID
STDB2
oracle19c>sqlplus '/ as sysdba'

SQL*Plus: Release 19.0.0.0.0 - Production on Sat Jul 24 18:11:12 2021
Version 19.3.0.0.0

Copyright (c) 1982, 2019, Oracle.  All rights reserved.


Connected to:
Oracle Database 19c Enterprise Edition Release 19.0.0.0.0 - Production
Version 19.3.0.0.0

SQL> shutdown immediate;
Database closed.
Database dismounted.
ORACLE instance shut down.
SQL> startup mount;
ORACLE instance started.

Total System Global Area 1207955552 bytes
Fixed Size                  9134176 bytes
Variable Size             318767104 bytes
Database Buffers          872415232 bytes
Redo Buffers                7639040 bytes
Database mounted.
SQL> alter database recover managed standby database using archived logfile disconnect from session;

Database altered.

SQL>
SQL> quit
Disconnected from Oracle Database 19c Enterprise Edition Release 19.0.0.0.0 - Production
Version 19.3.0.0.0
oracle19c>


Later ... from the STDB2 alert log :

2021-07-24T18:40:22.531574+08:00
PR00 (PID:29990): Media Recovery Log /opt/oracle/archivelog/STDB21_309_1036108814.dbf
PR00 (PID:29990): Media Recovery Delayed for 59 minute(s) T-1.S-310
2021-07-24T18:40:43.574486+08:00
 rfs (PID:531): No SRLs available for T-1
2021-07-24T18:40:43.743506+08:00
 rfs (PID:531): Opened log for T-1.S-317 dbid 2778483057 branch 1036108814
2021-07-24T18:40:43.762715+08:00
ARC3 (PID:29836): Archived Log entry 57 added for T-1.S-316 ID 0xa7521ccd LAD:1
2021-07-24T18:40:43.762785+08:00
ARC3 (PID:29836): Archive log for T-1.S-316 available in 60 minute(s)
2021-07-24T18:49:27.636427+08:00
PR00 (PID:29990): Media Recovery Log /opt/oracle/archivelog/STDB21_310_1036108814.dbf
PR00 (PID:29990): Media Recovery Delayed for 60 minute(s) T-1.S-311
2021-07-24T18:50:45.257290+08:00
 rfs (PID:531): Archived Log entry 58 added for B-1036108814.T-1.S-317 ID 0xa7521ccd LAD:3
2021-07-24T18:50:46.045279+08:00
 rfs (PID:531): Selected LNO:4 for T-1.S-318 dbid 2778483057 branch 1036108814


oracle19c>echo $ORACLE_SID
STDB2
oracle19c>sqlplus '/ as sysdba'

SQL*Plus: Release 19.0.0.0.0 - Production on Sat Jul 24 18:51:27 2021
Version 19.3.0.0.0

Copyright (c) 1982, 2019, Oracle.  All rights reserved.


Connected to:
Oracle Database 19c Enterprise Edition Release 19.0.0.0.0 - Production
Version 19.3.0.0.0

SQL> alter database recover managed standby database cancel;

Database altered.

SQL> alter database open read only;

Database altered.

SQL> alter pluggable database orclpdb1 open read only;

Pluggable database altered.

SQL> alter session set container=orclpdb1;

Session altered.

SQL>
SQL> select * from hemant.job_tracking_tbl order by 1;

JOB_END_TIME                                                       JOB_ID     DB_SCN
-------------------------------------------------------------- ---------- ----------
24-JUL-21 04.57.21.676949 PM                                            1   13239134
24-JUL-21 05.04.29.870877 PM                                            2   13241261
24-JUL-21 05.30.17.962275 PM                                            3   13246616
24-JUL-21 05.39.10.912969 PM                                            4   13247859
24-JUL-21 05.40.20.865467 PM                                            5   13248159
24-JUL-21 05.50.23.930352 PM                                            6   13252182

6 rows selected.

SQL>
SQL> select systimestamp from dual;

SYSTIMESTAMP
---------------------------------------------------------------------------
24-JUL-21 06.53.31.159094 PM +08:00

SQL>

So, now at 18:47, STDB2 has applied Sequence#310 and the database now shows data that came through that ArchiveLog. Upto JOB_ID=6, JOB_END_TIME=05:50:23pm
The Primary has already progressed further.

SQL> l
  1* select * from hemant.job_tracking_tbl order by 1
SQL> /

JOB_END_TIME                                                       JOB_ID     DB_SCN
-------------------------------------------------------------- ---------- ----------
24-JUL-21 04.57.21.676949 PM                                            1   13239134
24-JUL-21 05.04.29.870877 PM                                            2   13241261
24-JUL-21 05.30.17.962275 PM                                            3   13246616
24-JUL-21 05.39.10.912969 PM                                            4   13247859
24-JUL-21 05.40.20.865467 PM                                            5   13248159
24-JUL-21 05.50.23.930352 PM                                            6   13252182
24-JUL-21 06.00.27.037797 PM                                            7   13257658
24-JUL-21 06.10.33.163203 PM                                            8   13259223
24-JUL-21 06.20.36.839944 PM                                            9   13261275
24-JUL-21 06.22.46.972310 PM                                           10   13261560
24-JUL-21 06.30.39.787880 PM                                           11   13262799
24-JUL-21 06.37.18.623659 PM                                           12   13263658
24-JUL-21 06.40.41.713016 PM                                           13   13264263
24-JUL-21 06.50.43.755835 PM                                           14   13265798

14 rows selected.

SQL>

So, the operative methods at the Standby are :

 For Recovery :

1.  alter database recover managed standby database USING ARCHIVED LOGFILE disconnect from session

To Open and Query :

1. alter database recover managed standby database CANCEL

2. alter database OPEN READ ONLY

3. alter pluggable database <pdbname>   OPEN READ ONLY

To resume Recovery :

1. shutdown immediate

2. startup mount

3. alter database recover managed standby database USING ARCHIVED LOGFILE disconnect from session

While the Primary must specify a DELAY value in the log_archive_dest_n parameter for this destination Standby

A Standby that lags the Primary by a deliberate Delay

 As I noted in my previous blog post, with multiple Standby databases, you can have one or more of them, lagging the Primary. This allows the organisation a database that can be quickly opened for data recovery in case someone makes a mistake and deletes data or drops tables/objects from the Primary and the delete/drop has already been replicated to the first Standby.

Here is a quick demo.

At the Primary I have :

SQL> alter system archive log current;

System altered.

SQL> archive log list;
Database log mode              Archive Mode
Automatic archival             Enabled
Archive destination            /opt/oracle/archivelog/ORCLCDB
Oldest online log sequence     280
Next log sequence to archive   282
Current log sequence           282
SQL>

At the first Standby "STDBYDB", I have :

2021-07-20T17:50:05.870763+08:00
PR00 (PID:2912): Media Recovery Waiting for T-1.S-282 (in transit)
2021-07-20T17:50:06.354006+08:00
ARC3 (PID:2736): Archived Log entry 35 added for T-1.S-281 ID 0xa7521ccd LAD:1
2021-07-20T17:50:06.396543+08:00
 rfs (PID:3263): Archival of T-1.S-281 complete
2021-07-20T17:50:06.527483+08:00
 rfs (PID:3263): Selected LNO:4 for T-1.S-282 dbid 2778483057 branch 1036108814
2021-07-20T17:50:07.008298+08:00
Recovery of Online Redo Log: Thread 1 Group 4 Seq 282 Reading mem 0
  Mem# 0: /opt/oracle/oradata/STDBYDB/stdbredo01.log

At the second Standby "STDB2", I have :

2021-07-20T17:50:09.484608+08:00
PR00 (PID:2975): Media Recovery Waiting for T-1.S-282 (in transit)
2021-07-20T17:50:09.500278+08:00
 rfs (PID:3216): Opened log for T-1.S-282 dbid 2778483057 branch 1036108814
2021-07-20T17:50:09.527462+08:00
ARC3 (PID:2867): Archived Log entry 22 added for T-1.S-281 ID 0xa7521ccd LAD:1

To introduce a delay in applying ArchiveLogs at STDBY, I specify the DELAY parameter at the Primary database  :

SQL> show parameter log_archive_dest_3

NAME                                 TYPE        VALUE
------------------------------------ ----------- ------------------------------
log_archive_dest_3                   string      SERVICE=STDB2 ASYNC VALID_FOR=
                                                 (ONLINE_LOGFILES,PRIMARY_ROLE)
                                                  DB_UNIQUE_NAME=STDB2
log_archive_dest_30                  string
log_archive_dest_31                  string
SQL>  alter system set log_archive_dest_3='SERVICE=STDB2 DELAY=60 ASYNC VALID_FOR=(ONLINE_LOGFILES,PRIMARY_ROLE) DB_UNIQUE_NAME=STDB2';

System altered.

SQL> show parameter log_archive_dest_3

NAME                                 TYPE        VALUE
------------------------------------ ----------- ------------------------------
log_archive_dest_3                   string      SERVICE=STDB2 DELAY=60 ASYNC V
                                                 ALID_FOR=(ONLINE_LOGFILES,PRIM
                                                 ARY_ROLE) DB_UNIQUE_NAME=STDB2
log_archive_dest_30                  string
log_archive_dest_31                  string
SQL>

So, I have introduced a lag of 60minutes for STDB2.  Over the next 60minutes, the Primary will continue generating Redo and ArchiveLogs and both Standbys will be receiving them.  But STDB2 will apply them only after 60minutes.  

However, to enforce this, I must also cause the Standby to *not* use Real Time Apply, so I must change the RECOVER command at the Standby  Note, however, that this must be done on the Standby first !  The Standby must start it's recovery with "USING ARCHIVED LOGFILE" *before* the Primary sets a DELAY value for the target log_archive_dest_n

oracle19c>echo $ORACLE_SID
STDB2
oracle19c>sqlplus '/ as sysdba'

SQL*Plus: Release 19.0.0.0.0 - Production on Tue Jul 20 18:14:26 2021
Version 19.3.0.0.0

Copyright (c) 1982, 2019, Oracle.  All rights reserved.


Connected to:
Oracle Database 19c Enterprise Edition Release 19.0.0.0.0 - Production
Version 19.3.0.0.0

SQL> alter database recover managed standby database cancel;

Database altered.

SQL> alter database recover managed standby database using archived logfile disconnect from session;

Database altered.

SQL>

If I do not change the RECOVER command at the Standby to explicitly specify "USING ARCHIVED LOGFILE", it still defaults to Real Time Apply and ignores the DELAY specified by the Primary. I get the message in STDB2 alert log :

 rfs (PID:13712): WARN: Managed Standby Recovery started with REAL TIME APPLY
 rfs (PID:13712): WARN: DELAY 60 minutes specified at primary ignored

I review the alert log files for all 3 databases about an hour later.

This is the first Standby (STDBYDB)

2021-07-20T19:10:57.161885+08:00
PR00 (PID:2912): Media Recovery Waiting for T-1.S-297 (in transit)
2021-07-20T19:10:57.324337+08:00
 rfs (PID:3263): Selected LNO:4 for T-1.S-297 dbid 2778483057 branch 1036108814
2021-07-20T19:10:58.401720+08:00
Recovery of Online Redo Log: Thread 1 Group 4 Seq 297 Reading mem 0
  Mem# 0: /opt/oracle/oradata/STDBYDB/stdbredo01.log
2021-07-20T19:14:39.910894+08:00
ARC3 (PID:2736): Archived Log entry 51 added for T-1.S-297 ID 0xa7521ccd LAD:1
2021-07-20T19:14:39.943728+08:00
 rfs (PID:3263): Standby controlfile consistent with primary
2021-07-20T19:14:40.136187+08:00
 rfs (PID:3263): Selected LNO:4 for T-1.S-298 dbid 2778483057 branch 1036108814
2021-07-20T19:14:40.136811+08:00
PR00 (PID:2912): Media Recovery Waiting for T-1.S-298 (in transit)
2021-07-20T19:14:41.180355+08:00
Recovery of Online Redo Log: Thread 1 Group 4 Seq 298 Reading mem 0
  Mem# 0: /opt/oracle/oradata/STDBYDB/stdbredo01.log

SQL> show parameter db_unique_name

NAME                                 TYPE        VALUE
------------------------------------ ----------- ------------------------------
db_unique_name                       string      STDBYDB
SQL> select current_scn from v$database;

CURRENT_SCN
-----------
   13123569

SQL>

And this is the Second Standby (STDB2) :

2021-07-20T19:10:58.180405+08:00
 rfs (PID:21331): No SRLs available for T-1
2021-07-20T19:10:58.319036+08:00
ARC0 (PID:2857): Archived Log entry 37 added for T-1.S-296 ID 0xa7521ccd LAD:1
2021-07-20T19:10:58.319358+08:00
ARC0 (PID:2857): Archive log for T-1.S-296 available in 60 minute(s)
2021-07-20T19:10:58.320321+08:00
 rfs (PID:21331): Opened log for T-1.S-297 dbid 2778483057 branch 1036108814
2021-07-20T19:14:40.363888+08:00
 rfs (PID:21331): Archived Log entry 38 added for B-1036108814.T-1.S-297 ID 0xa7521ccd LAD:3
2021-07-20T19:14:40.782081+08:00
 rfs (PID:21331): Selected LNO:4 for T-1.S-298 dbid 2778483057 branch 1036108814
2021-07-20T19:15:12.430015+08:00
PR00 (PID:26793): Media Recovery Log /opt/oracle/archivelog/STDB21_286_1036108814.dbf
PR00 (PID:26793): Media Recovery Delayed for 60 minute(s) T-1.S-287

SQL> show parameter db_unique_name

NAME                                 TYPE        VALUE
------------------------------------ ----------- ------------------------------
db_unique_name                       string      STDB2
SQL> select current_scn from v$database;

CURRENT_SCN
-----------
   13108691

SQL>

So, while the first Standby (STDBYDB) has already applied and archived Sequence#297 and is currently applying Sequence#298 from the Standby logfile, the second Standby (STDB2) has archived Sequence#297 and received Sequence#298 but notifies that Sequence#286 is currently being applied and the apply Sequence#297 is delayed by 60minutes.
I can also run an SCN_TO_TIMESTAMP query in the Primary (this cannot be executed on a Standby that is not OPEN) :

SQL> select scn_to_timestamp(13123569) At_STDBYDB from dual;

AT_STDBYDB
---------------------------------------------------------------------------
20-JUL-21 07.13.39.000000000 PM

SQL> select scn_to_timestamp(13108691) At_STDB2 from dual;

AT_STDB2
---------------------------------------------------------------------------
20-JUL-21 06.15.07.000000000 PM

SQL>

This shows that STDB2 is lagging by about 60minutes

So, if any "bad action"  (deletion of data or dropping of objects) is detected at the Primary (and a Flashback option is not available), the Standby can be OPENed Read Only to view the data as it was 1hour ago.  

I'll show that option in my next blog post.

Creating [1 or more] Additional Standby Database[s]

 In my previous demos, I've shown 1 Standby (STDBYDB) for the Primary (ORCLCDB).

However, an organisation may choose to have additional Standby Databases.  

Note that each Standby Database must be licensed.  If the Primary is licensed by Processor count (say 32 processors), each Standby must also be licensed by Processor count, although the Standbys may be running on servers with fewer processors, thus needing smaller licensing than the Primary.

Why would an organisation have additional Standbys ?  Multiple reasons

1.  Having additional Disaster Recovery Data Centres.  Thus, with 2 DRCs in addition to the Production, there may be 2 Standby Databases.

2. Choosing to have a Standby that can be used to run Reporting Queries when it is OPENed READ ONLY for, say, 4hours a day.  [If the organisation has the Active Dataguard Licence, the Standby can be running Reporting Queries all 24hours]  Without the Active Dataguard Licence, Recovery has to be stopped at this Standby for those 4hours, though it will continue to receive and Archive Redo from the Primary.  Once the 4hour window is closed, Recovery can resumed and the Standby will apply all the "pending" Redo [from the ArchiveLogs that it has received] to "catch-up" to the Primary

3. To test D.R. scenarios, whereby a secondary Standby is opened for Read-Write operations, without impacting the actual Production and first Standby instance which continue to be in-sync throughout the D.R. testing.  The Standby may be opened as a Snapshot Standby

4. To have an environment that can be opened day-time hours for UAT while it is resynced (i.e. Redo Apply done) with the Primary every night.  This will require the Standby to have Flashback Restore Points created and reverted to each day.

5. To have a Standby that is lagging the Primary by, say, 30minutes or 1hour or 4hours deliberately.  This allows the organisation a database that can be quickly opened for data recovery in case someone makes a mistake and deletes data or drops tables/objects from the Primary and the delete/drop has already been replicated to the first Standby.

6. To run [additional] Database Backups at a Standby DRC.  I prefer that the Primary and each Standby be running local database backups.

7. To migrate the Production Primary to a new Data Centre via a Standby built and refreshed over a week and then a Switchover executed during a weekend downtime.

8. To migrate the Production Primary to new  Server Hardware via a Standby built and refreshed over a week and then a Switchover executed during a weekend downtime.

I have built my Second Standby with these parameters :

initSTDB2.ora :

*.audit_file_dest='/opt/oracle/admin/STDB2/adump'
*.audit_sys_operations=false
*.audit_trail='none'
*.commit_logging='batch'
*.commit_wait='nowait'
*.compatible='19.0.0'
*.control_files='/opt/oracle/oradata/STDB2/control01.ctl','/opt/oracle/oradata/STDB2/control02.ctl'
*.db_block_size=8192
*.db_create_file_dest='/opt/oracle/oradata'
*.db_file_name_convert='/opt/oracle/oradata/ORCLCDB','/opt/oracle/oradata/STDB2'
*.db_name='ORCLCDB'
*.db_recovery_file_dest_size=10G
*.db_recovery_file_dest='/opt/oracle/FRA/STDB2'
*.db_unique_name='STDB2'
*.diagnostic_dest='/opt/oracle'
*.dispatchers='(PROTOCOL=TCP) (SERVICE=STDB2XDB)'
*.enable_pluggable_database=true
*.fal_server='ORCLCDB'
*.filesystemio_options='setall'
*.local_listener='LISTENER_STDB2'
*.log_archive_config='DG_CONFIG=(ORCLCDB, STDBYDB, STDB2)'
*.log_archive_dest_1='LOCATION=/opt/oracle/archivelog/STDB2'
###*.log_archive_dest_2='SERVICE=ORCLCDB SYNC AFFIRM VALID_FOR=(ONLINE_LOGFILES,PRIMARY_ROLE) DB_UNIQUE_NAME=ORCLCDB'
*.log_file_name_convert='/opt/oracle/oradata/ORCLCDB','/opt/oracle/oradata/STDB2'
*.nls_language='AMERICAN'
*.nls_territory='AMERICA'
*.open_cursors=300
*.pga_aggregate_target=384m
*.processes=300
*.remote_login_passwordfile='EXCLUSIVE'
*.sga_target=1152m
*.standby_file_management='AUTO'
*.undo_tablespace='UNDOTBS1'


and listener.ora :
LISTENER_STDB2 =
(DESCRIPTION_LIST =
  (DESCRIPTION =
    (ADDRESS = (PROTOCOL = TCP)(HOST = 0.0.0.0)(PORT = 1532))
  )
)

SID_LIST_LISTENER_STDB2 =
 (SID_LIST=
  (SID_DESC =
   (ORACLE_HOME = /opt/oracle/product/19c/dbhome_1)
   (SID_NAME = STDB2)
   )
 )

and the password file copied from the Primary -- although Oracle will copy the password file in recent versions, I still prefer to do it myself
oracle19c>pwd
/opt/oracle/product/19c/dbhome_1/dbs
oracle19c>ls orapw$ORACLE_SID
orapwSTDB2
oracle19c>

Note how I have disabled log_archive_dest_2.  In this case, this Standby will never be part of a Role Reversal done either by Switchover or Failover [where it could become a Primary in the D.R. Data Centre and ship Redo to the old Primary in the Production Data Centre]

oracle19c>lsnrctl start listener_STDB2

LSNRCTL for Linux: Version 19.0.0.0.0 - Production on 15-JUL-2021 22:07:14

Copyright (c) 1991, 2019, Oracle.  All rights reserved.

Starting /opt/oracle/product/19c/dbhome_1/bin/tnslsnr: please wait...

TNSLSNR for Linux: Version 19.0.0.0.0 - Production
System parameter file is /opt/oracle/product/19c/dbhome_1/network/admin/listener.ora
Log messages written to /opt/oracle/diag/tnslsnr/ora19cs1/listener_stdb2/alert/log.xml
Listening on: (DESCRIPTION=(ADDRESS=(PROTOCOL=tcp)(HOST=0.0.0.0)(PORT=1532)))

Connecting to (DESCRIPTION=(ADDRESS=(PROTOCOL=TCP)(HOST=0.0.0.0)(PORT=1532)))
STATUS of the LISTENER
------------------------
Alias                     listener_STDB2
Version                   TNSLSNR for Linux: Version 19.0.0.0.0 - Production
Start Date                15-JUL-2021 22:07:14
Uptime                    0 days 0 hr. 0 min. 0 sec
Trace Level               off
Security                  ON: Local OS Authentication
SNMP                      OFF
Listener Parameter File   /opt/oracle/product/19c/dbhome_1/network/admin/listener.ora
Listener Log File         /opt/oracle/diag/tnslsnr/ora19cs1/listener_stdb2/alert/log.xml
Listening Endpoints Summary...
  (DESCRIPTION=(ADDRESS=(PROTOCOL=tcp)(HOST=0.0.0.0)(PORT=1532)))
Services Summary...
Service "STDB2" has 1 instance(s).
  Instance "STDB2", status UNKNOWN, has 1 handler(s) for this service...
The command completed successfully
oracle19c>
oracle19c>echo $ORACLE_SID
STDB2
oracle19c>sqlplus '/ as sysdba'

SQL*Plus: Release 19.0.0.0.0 - Production on Thu Jul 15 22:21:47 2021
Version 19.3.0.0.0

Copyright (c) 1982, 2019, Oracle.  All rights reserved.

Connected to an idle instance.

SQL> startup nomount;
ORACLE instance started.

Total System Global Area 1207955552 bytes
Fixed Size                  9134176 bytes
Variable Size             318767104 bytes
Database Buffers          872415232 bytes
Redo Buffers                7639040 bytes
SQL>

Then, at the Primary :

SQL> show parameter log_archive_config

NAME                                 TYPE        VALUE
------------------------------------ ----------- ------------------------------
log_archive_config                   string      DG_CONFIG=(ORCLCDB, STDBYDB)
SQL> alter system set log_archive_config='DG_CONFIG=(ORCLCDB, STDBYDB, STDB2)';  -- note that this update is also done at the first Standy STDBYDB2

System altered.

SQL> show parameter log_archive_dest_2

NAME                                 TYPE        VALUE
------------------------------------ ----------- ------------------------------
log_archive_dest_2                   string      SERVICE=STDBYDB SYNC AFFIRM VA
                                                 LID_FOR=(ONLINE_LOGFILES,PRIMA
                                                 RY_ROLE) DB_UNIQUE_NAME=STDBYD
                                                 B
log_archive_dest_20                  string
log_archive_dest_21                  string
log_archive_dest_22                  string
log_archive_dest_23                  string
log_archive_dest_24                  string
log_archive_dest_25                  string
log_archive_dest_26                  string
log_archive_dest_27                  string
log_archive_dest_28                  string
log_archive_dest_29                  string
SQL> show parameter log_archive_dest_3

NAME                                 TYPE        VALUE
------------------------------------ ----------- ------------------------------
log_archive_dest_3                   string
log_archive_dest_30                  string
log_archive_dest_31                  string
SQL> alter system set log_archive_dest_3='SERVICE=STDB2 ASYNC VALID_FOR=(ONLINE_LOGFILES,PRIMARY_ROLE) DB_UNIQUE_NAME=STDB2';

System altered.

SQL>

Note that log_archive_dest_2 is set to SYNC AFFIRM but log_archive_dest_3 is set to ASYNC and [NOAFFIRM]. Because I can afford a slight lag (latency) for this Standby.  Maybe it is at a Data Centre that is much further away, increasing the latency.

Then, I copy the Primary to the new Standby using RMAN (see the previous example)

-- assuming that the tnsnames.ora at the Primary is updated to include "STDB2"

oracle19c>rman target sys/manager auxiliary sys/manager@STDB2

Recovery Manager: Release 19.0.0.0.0 - Production on Thu Jul 15 22:24:00 2021
Version 19.3.0.0.0

Copyright (c) 1982, 2019, Oracle and/or its affiliates.  All rights reserved.

connected to target database: ORCLCDB (DBID=2778483057)
connected to auxiliary database: ORCLCDB (not mounted)

RMAN> duplicate target database for standby from active database dorecover;

Starting Duplicate Db at 15-JUL-21
using target database control file instead of recovery catalog
allocated channel: ORA_AUX_DISK_1
channel ORA_AUX_DISK_1: SID=21 device type=DISK
allocated channel: ORA_AUX_DISK_2
channel ORA_AUX_DISK_2: SID=256 device type=DISK
current log archived

contents of Memory Script:
{
   backup as copy reuse
   passwordfile auxiliary format  '/opt/oracle/product/19c/dbhome_1/dbs/orapwSTDB2'   ;
}
executing Memory Script

Starting backup at 15-JUL-21
allocated channel: ORA_DISK_1
channel ORA_DISK_1: SID=273 device type=DISK
allocated channel: ORA_DISK_2
channel ORA_DISK_2: SID=274 device type=DISK

....
....
....
....
....

Finished Duplicate Db at 15-JUL-21

RMAN>

After that I shutdown and restart the new Standby and enable Recovery

SQL> shutdown immediate;
ORA-01109: database not open


Database dismounted.
ORACLE instance shut down.
SQL> startup mount;
ORACLE instance started.

Total System Global Area 1207955552 bytes
Fixed Size                  9134176 bytes
Variable Size             318767104 bytes
Database Buffers          872415232 bytes
Redo Buffers                7639040 bytes
Database mounted.
SQL> alter database recover managed standby database disconnect from session;

Database altered.

SQL>

Let me now verify that the Primary is shipping Redo to *2* Standbys

2021-07-15T22:49:55.328276+08:00
ALTER SYSTEM ARCHIVE LOG
2021-07-15T22:49:56.801388+08:00
LGWR (PID:6404): SRL selected to archive T-1.S-272
LGWR (PID:6404): SRL selected for T-1.S-272 for LAD:2
LGWR (PID:6404): SRL selected to archive T-1.S-272. However, all SRLs are currently active
2021-07-15T22:49:57.005260+08:00
Thread 1 advanced to log sequence 272 (LGWR switch)
  Current log# 2 seq# 272 mem# 0: /opt/oracle/oradata/ORCLCDB/redo02.log
2021-07-15T22:49:57.973514+08:00
NET  (PID:28568): Archived Log entry 588 added for T-1.S-271 ID 0xa7521ccd LAD:1
2021-07-15T22:49:59.487773+08:00
TT03 (PID:8344): SRL selected for T-1.S-272 for LAD:3
2021-07-15T22:50:06.667866+08:00
ALTER SYSTEM ARCHIVE LOG
2021-07-15T22:50:07.669103+08:00
LGWR (PID:6404): SRL selected to archive T-1.S-273
LGWR (PID:6404): SRL selected for T-1.S-273 for LAD:2
LGWR (PID:6404): SRL selected to archive T-1.S-273. However, all SRLs are currently active
2021-07-15T22:50:07.862847+08:00
Thread 1 advanced to log sequence 273 (LGWR switch)
  Current log# 3 seq# 273 mem# 0: /opt/oracle/oradata/ORCLCDB/redo03.log
2021-07-15T22:50:08.191087+08:00
NET  (PID:28568): Archived Log entry 592 added for T-1.S-272 ID 0xa7521ccd LAD:1
2021-07-15T22:52:42.492619+08:00
ALTER SYSTEM ARCHIVE LOG
2021-07-15T22:52:42.988590+08:00
LGWR (PID:6404): SRL selected to archive T-1.S-274
LGWR (PID:6404): SRL selected for T-1.S-274 for LAD:2
LGWR (PID:6404): SRL selected to archive T-1.S-274. However, all SRLs are currently active
2021-07-15T22:52:43.059342+08:00
Thread 1 advanced to log sequence 274 (LGWR switch)
  Current log# 1 seq# 274 mem# 0: /opt/oracle/oradata/ORCLCDB/redo01.log
2021-07-15T22:52:43.142797+08:00
NET  (PID:28568): Archived Log entry 594 added for T-1.S-273 ID 0xa7521ccd LAD:1
2021-07-15T22:52:43.917398+08:00
TT03 (PID:8344): SRL selected for T-1.S-274 for LAD:3



and

SQL> select protection_mode from v$database;

PROTECTION_MODE
--------------------
MAXIMUM AVAILABILITY

SQL>

Thus, we can see that the Primary is shipping Redo to both "LAD2" and "LAD3"

While the first Standby (STDBYDB) shows :

ARC2 (PID:23577): Archived Log entry 26 added for T-1.S-272 ID 0xa7521ccd LAD:1
2021-07-15T22:50:07.448944+08:00
 rfs (PID:23634): Archival of T-1.S-272 complete
2021-07-15T22:50:07.533082+08:00
PR00 (PID:23765): Media Recovery Waiting for T-1.S-273 (in transit)
2021-07-15T22:50:07.668714+08:00
 rfs (PID:23634): Selected LNO:4 for T-1.S-273 dbid 2778483057 branch 1036108814
2021-07-15T22:50:09.298541+08:00
Recovery of Online Redo Log: Thread 1 Group 4 Seq 273 Reading mem 0
  Mem# 0: /opt/oracle/oradata/STDBYDB/stdbredo01.log
2021-07-15T22:52:42.921683+08:00
 rfs (PID:23634): Standby controlfile consistent with primary
 rfs (PID:23634): No SRLs available for T-1
 rfs (PID:23634): Waiting for T-1.S-273 archival to complete
2021-07-15T22:52:42.929839+08:00
ARC0 (PID:23569): Archived Log entry 27 added for T-1.S-273 ID 0xa7521ccd LAD:1
2021-07-15T22:52:42.935444+08:00
 rfs (PID:23634): Archival of T-1.S-273 complete
2021-07-15T22:52:42.988242+08:00
 rfs (PID:23634): Selected LNO:4 for T-1.S-274 dbid 2778483057 branch 1036108814
2021-07-15T22:52:43.001151+08:00
PR00 (PID:23765): Media Recovery Waiting for T-1.S-274 (in transit)
2021-07-15T22:52:43.002924+08:00
Recovery of Online Redo Log: Thread 1 Group 4 Seq 274 Reading mem 0
  Mem# 0: /opt/oracle/oradata/STDBYDB/stdbredo01.log


and

SQL> select protection_mode from v$database;

PROTECTION_MODE
--------------------
MAXIMUM AVAILABILITY

SQL> 

And the new Standby (STDB2) shows :

Recovery of Online Redo Log: Thread 1 Group 4 Seq 272 Reading mem 0
  Mem# 0: /opt/oracle/oradata/STDB2/stdbredo01.log
2021-07-15T22:50:08.467719+08:00
 rfs (PID:23637): No SRLs available for T-1
2021-07-15T22:50:08.530396+08:00
PR00 (PID:23655): Media Recovery Waiting for T-1.S-273 (in transit)
2021-07-15T22:50:08.532241+08:00
 rfs (PID:23637): Opened log for T-1.S-273 dbid 2778483057 branch 1036108814
2021-07-15T22:50:08.532340+08:00
ARC0 (PID:23450): Archived Log entry 13 added for T-1.S-272 ID 0xa7521ccd LAD:1
2021-07-15T22:52:43.299392+08:00
 rfs (PID:23637): Archived Log entry 14 added for B-1036108814.T-1.S-273 ID 0xa7521ccd LAD:3
2021-07-15T22:52:43.917098+08:00
 rfs (PID:23637): Selected LNO:4 for T-1.S-274 dbid 2778483057 branch 1036108814
2021-07-15T22:52:44.137685+08:00
PR00 (PID:23655): Media Recovery Log /opt/oracle/archivelog/STDB21_273_1036108814.dbf
PR00 (PID:23655): Media Recovery Waiting for T-1.S-274 (in transit)
2021-07-15T22:52:44.576855+08:00
Recovery of Online Redo Log: Thread 1 Group 4 Seq 274 Reading mem 0
  Mem# 0: /opt/oracle/oradata/STDB2/stdbredo01.log


and

SQL> select protection_mode from v$database;

PROTECTION_MODE
--------------------
MAXIMUM PERFORMANCE

SQL>

So, now the Primary is shipping Redo to 2 Standbys. The first Standby (STDBYDB) is in MAXIMUM AVAILABILITY mode, the second Standby (STDB2) is in MAXIMUM PERFORMANCE mode.

(For simplicity, I haven't created additional Standby Redo Logs, I have only Standby Redo Log at Primary and each of the Standbys --- it is group#4, filename "stdbredo01.log" in the listing above)

    

In the next blog post, I will explore how to have STDB2 deliberately lagging the Primary by 30minutes 60minutes (use case 5 in the list at the beginning of this blog post)  {updated to 60minutes)

Failover to a Standby with Data Loss

 One scenario I consider when testing a Standby Database -- actually testing a Disaster Recovery  site --  is to simulate complete loss of the Primary Database (at the Production site)

Previously, I have demonstrated two methods of "Destructive" Disaster Recovery site testing :

1. Being able to Switchover to the Standby (at a Disaster Recovery site), verifying that transactions created at the Disaster Recovery site do actually flow back to the Production site and, finally, reverting both databases to their pre-test image)

2. Opening a Standby Database as a Snapshot Standby, verify that it is writable and, finally, reverting it to being a "normal" Standby

But what if you want to simulate complete loss of the Production site (Primary database) -- signifying that you may have even lost some transactions, but needing to open the Standby for Read-Write operations ?  OR you actually encounter such a situation where the Primary goes down when the Standby is lagging ? 

On my Primary I had these configured for MAXIMUM AVAILABILITY (with SYNC and AFFIRM for log_archive_dest_2) :

SQL>alter system set log_archive_dest_2='SERVICE=STDBYDB SYNC AFFIRM VALID_FOR=(ONLINE_LOGFILES,PRIMARY_ROLE) DB_UNIQUE_NAME=STDBYDB';

System altered.

SQL>
SQL> shutdown immediate;
Database closed.
Database dismounted.
ORACLE instance shut down.
SQL> startup mount;
ORACLE instance started.

Total System Global Area 1207955552 bytes
Fixed Size                  9134176 bytes
Variable Size             822083584 bytes
Database Buffers          369098752 bytes
Redo Buffers                7639040 bytes
Database mounted.
SQL>
SQL> alter database SET STANDBY DATABASE TO MAXIMIZE AVAILABILITY ;

Database altered.

SQL> shutdown
ORA-01109: database not open


Database dismounted.
ORACLE instance shut down.
SQL> startup;
ORACLE instance started.

Total System Global Area 1207955552 bytes
Fixed Size                  9134176 bytes
Variable Size             822083584 bytes
Database Buffers          369098752 bytes
Redo Buffers                7639040 bytes
Database mounted.
Database opened.
SQL>
SQL> select protection_mode, open_mode, database_role from v$database;

PROTECTION_MODE      OPEN_MODE            DATABASE_ROLE
-------------------- -------------------- ----------------
MAXIMUM AVAILABILITY READ WRITE           PRIMARY

SQL>

This is information of the Primary Database (at the Production site)  when Primary and Standby are in sync :

SQL> archive log list;
Database log mode              Archive Mode
Automatic archival             Enabled
Archive destination            /opt/oracle/archivelog/ORCLCDB
Oldest online log sequence     235
Next log sequence to archive   237
Current log sequence           237
SQL>

And this is at the Standby Database (at the Disaster Recovery site) :

 
 rfs (PID:4415): Changing standby controlfile to MAXIMUM AVAILABILITY mode
 rfs (PID:4415): Changing standby controlfile to RESYNCHRONIZATION level
 rfs (PID:4415): Standby controlfile consistent with primary
 
Recovery of Online Redo Log: Thread 1 Group 5 Seq 237 Reading mem 0
  Mem# 0: /opt/oracle/oradata/STDBYDB/stdbredo02.log

My Standby database / site goes down for a short while.  

So the Primary reports :

LGWR (PID:4280): Failed to archive LNO:1 T-1.S-238, error=3113
LGWR (PID:4280): Error 1041 disconnecting from LAD:2 standby host 'STDBYDB'
2021-07-10T17:58:05.264661+08:00
Thread 1 advanced to log sequence 239 (LGWR switch)
  Current log# 2 seq# 239 mem# 0: /opt/oracle/oradata/ORCLCDB/redo02.log
2021-07-10T17:58:05.390312+08:00
ARC1 (PID:4345): Archived Log entry 524 added for T-1.S-238 ID 0xa7521ccd LAD:1


and sometime later

SQL> archive log list;
Database log mode              Archive Mode
Automatic archival             Enabled
Archive destination            /opt/oracle/archivelog/ORCLCDB
Oldest online log sequence     242
Next log sequence to archive   244
Current log sequence           244
SQL> select current_scn from v$database;

CURRENT_SCN
-----------
   12768459

SQL>

We now know that the Standby (last applying Online Redo Sequence#237 [it went on to 238 before the Instance/Server crashed]) is lagging behind the Primary (current Redo Log 244, SCN 12768459 .

Reminder : The configuration is in MAXIMUM AVAILABILITY mode (not MAXIMUM PROTECTION).

Now, my Primary goes down.

So the FAILOVER must be issued from the Standby when I bring it up.  Let my try that now

SQL> startup mount;
ORACLE instance started.

Total System Global Area 1207955552 bytes
Fixed Size                  9134176 bytes
Variable Size             436207616 bytes
Database Buffers          754974720 bytes
Redo Buffers                7639040 bytes
Database mounted.
SQL> create restore point before_failover guarantee flashback database;

Restore point created.

SQL>
SQL> select protection_mode, open_mode, database_role from v$database;

PROTECTION_MODE      OPEN_MODE            DATABASE_ROLE
-------------------- -------------------- ----------------
MAXIMUM AVAILABILITY MOUNTED              PHYSICAL STANDBY

SQL> show parameter db_unique_name

NAME                                 TYPE        VALUE
------------------------------------ ----------- ------------------------------
db_unique_name                       string      STDBYDB
SQL>
SQL> alter database failover to STDBYDB;

Database altered.

SQL>
SQL> select protection_mode, open_mode, database_role from v$database;

PROTECTION_MODE      OPEN_MODE            DATABASE_ROLE
-------------------- -------------------- ----------------
MAXIMUM PERFORMANCE  MOUNTED              PRIMARY

SQL> alter database open;

Database altered.

SQL>

Let's see what the alert log on the Standby shows

ARC1 (PID:6205): FAL: Error 12541 connecting to ORCLCDB for fetching gap sequence

2021-07-10T18:04:34.507622+08:00
RVWR started with pid=45, OS id=11261
2021-07-10T18:04:38.704492+08:00
Allocating 8388608 bytes in shared pool for flashback generation buffer.
Allocated 5430064 bytes in shared pool for flashback generation buffer
2021-07-10T18:04:39.802983+08:00
Created guaranteed restore point BEFORE_FAILOVER

2021-07-10T18:05:50.402988+08:00
alter database failover to STDBYDB
2021-07-10T18:05:50.403086+08:00
.... (PID:6240): The Time Management Interface (TMI) is being enabled for role transition
.... (PID:6240): information.  This will result in messages beingoutput to the alert log
.... (PID:6240): file with the prefix 'TMI: '.  This is being enabled to make the timing of
.... (PID:6240): the various stages of the role transition available for diagnostic purposes.
.... (PID:6240): This output will end when the role transition is complete.
TMI: dbsdrv failover to target BEGIN 2021-07-10 18:05:50.403742
Terminal Recovery requested in process 6240
TMI: adbdrv termRecovery BEGIN 2021-07-10 18:05:50.408509
2021-07-10T18:05:50.494984+08:00
Attempt to do a Terminal Recovery (STDBYDB)
TMI: adbdrv termRecovery END 2021-07-10 18:05:50.495053
2021-07-10T18:05:50.495642+08:00
Media Recovery Start: Managed Standby Recovery (STDBYDB)
 Started logmerger process
2021-07-10T18:05:50.532463+08:00
PR00 (PID:11393): Managed Standby Recovery not using Real Time Apply
max_pdb is 5
2021-07-10T18:05:51.094491+08:00
Parallel Media Recovery started with 2 slaves
2021-07-10T18:05:51.233396+08:00
stopping change tracking
2021-07-10T18:05:51.583393+08:00
PR00 (PID:11393): Begin: SRL archival
PR00 (PID:11393): End: SRL archival
PR00 (PID:11393): Terminal Recovery timestamp is '07/10/2021 18:05:51'
PR00 (PID:11393): Terminal Recovery: applying standby redo logs.
PR00 (PID:11393): Terminal Recovery: thread 1 seq# 238 redo required
2021-07-10T18:05:51.682050+08:00
PR00 (PID:11393): Terminal Recovery:
2021-07-10T18:05:51.685886+08:00
Recovery of Online Redo Log: Thread 1 Group 4 Seq 238 Reading mem 0
  Mem# 0: /opt/oracle/oradata/STDBYDB/stdbredo01.log
Terminal Recovery finished with No-Data-Loss
2021-07-10T18:05:51.985957+08:00
Incomplete Recovery applied until change 12767732 time 07/10/2021 17:57:59
2021-07-10T18:05:52.074427+08:00
Media Recovery Complete (STDBYDB)
2021-07-10T18:05:52.655358+08:00
Terminal Recovery: successful completion
PR00 (PID:11393): Forcing ARSCN to IRSCN for TR SCN:0x0000000000c2d1f4
PR00 (PID:11393): Attempt to set limbo arscn SCN:0x0000000000c2d1f4 irscn SCN:0x0000000000c2d1f4
PR00 (PID:11393): Resetting standby activation ID 2807176397 (0xa7521ccd)
stopping change tracking
2021-07-10T18:05:52.873082+08:00
ALTER DATABASE SWITCHOVER TO PRIMARY (STDBYDB)
2021-07-10T18:05:52.873684+08:00
Maximum wait for role transition is 15 minutes.
TMI: kcv_commit_to_so_to_primary wait for MRP to finish BEGIN 2021-07-10 18:05:52.873693
TMI: kcv_commit_to_so_to_primary wait for MRP to finish END 2021-07-10 18:05:52.873771
TMI: kcv_commit_to_so_to_primary Switchover from physical BEGIN 2021-07-10 18:05:52.874148
Backup controlfile written to trace file /opt/oracle/diag/rdbms/stdbydb/STDBYDB/trace/STDBYDB_ora_6240.trc
2021-07-10T18:05:53.571015+08:00
Standby terminal recovery start SCN: 12767670
RESETLOGS after complete recovery through change 12767732
NET  (PID:6240): ORL pre-clearing operation disabled by switchover
Online log /opt/oracle/oradata/STDBYDB/redo01.log: Thread 1 Group 1 was previously cleared
Online log /opt/oracle/oradata/STDBYDB/redo02.log: Thread 1 Group 2 was previously cleared
Online log /opt/oracle/oradata/STDBYDB/redo03.log: Thread 1 Group 3 was previously cleared
2021-07-10T18:05:53.907884+08:00
Standby became primary SCN: 12767669
2021-07-10T18:05:53.908157+08:00
Setting recovery target incarnation to 4
2021-07-10T18:05:53.966494+08:00
NET  (PID:6240): RT: Role transition work is not done
NET  (PID:6240): The Time Management Interface (TMI) is being enabled for role transition
NET  (PID:6240): information.  This will result in messages beingoutput to the alert log
NET  (PID:6240): file with the prefix 'TMI: '.  This is being enabled to make the timing of
NET  (PID:6240): the various stages of the role transition available for diagnostic purposes.
NET  (PID:6240): This output will end when the role transition is complete.
NET  (PID:6240): Redo network throttle feature is disabled at mount time
2021-07-10T18:05:54.087343+08:00
NET  (PID:6240): Database role cleared from PHYSICAL STANDBY [kcvs.c:1030]
Switchover: Complete - Database mounted as primary
TMI: kcv_commit_to_so_to_primary Switchover from physical END 2021-07-10 18:05:54.088069
TMI: dbsdrv failover to target END 2021-07-10 18:05:54.088135
Failover completed with No-Data-Loss.
Completed: alter database failover to STDBYDB
2021-07-10T18:06:33.754882+08:00
alter database open
2021-07-10T18:06:33.755022+08:00
TMI: adbdrv open database BEGIN 2021-07-10 18:06:33.754898
Ping without log force is disabled:
  instance mounted in exclusive mode.
2021-07-10T18:06:33.945688+08:00
Crash Recovery excluding pdb 2 which was cleanly closed.
2021-07-10T18:06:33.945807+08:00
Crash Recovery excluding pdb 4 which was cleanly closed.
Endian type of dictionary set to little
2021-07-10T18:06:33.999216+08:00
Assigning activation ID 2848626420 (0xa9ca96f4)
2021-07-10T18:06:34.009104+08:00
LGWR (PID:6045): LAD:2 is UNSYNCHRONIZED
2021-07-10T18:06:34.503600+08:00
Thread 1 advanced to log sequence 2 (thread open)
2021-07-10T18:06:34.503635+08:00
ARC3 (PID:6209): Becoming the 'no SRL' ARCH
2021-07-10T18:06:34.503832+08:00
Redo log for group 2, sequence 2 is not located on DAX storage
Thread 1 opened at log sequence 2
  Current log# 2 seq# 2 mem# 0: /opt/oracle/oradata/STDBYDB/redo02.log
Successful open of redo thread 1
2021-07-10T18:06:34.521347+08:00
MTTR advisory is disabled because FAST_START_MTTR_TARGET is not set
stopping change tracking
2021-07-10T18:06:34.662361+08:00
TT03 (PID:11461): Sleep 5 seconds and then try to clear SRLs in 2 time(s)
2021-07-10T18:06:34.883461+08:00
ARC0 (PID:6199): Archived Log entry 259 added for T-1.S-1 ID 0xa9ca96f4 LAD:1
2021-07-10T18:06:35.784688+08:00
Undo initialization recovery: err:0 start: 1407469 end: 1407660 diff: 191 ms (0.2 seconds)
[6240] Successfully onlined Undo Tablespace 2.
Undo initialization online undo segments: err:0 start: 1407660 end: 1408339 diff: 679 ms (0.7 seconds)
Undo initialization finished serial:0 start:1407469 end:1408406 diff:937 ms (0.9 seconds)
Dictionary check beginning
2021-07-10T18:06:36.927538+08:00
Dictionary check complete
Verifying minimum file header compatibility for tablespace encryption for pdb 1..
Verifying file header compatibility for tablespace encryption completed for pdb 1
Database Characterset is AL32UTF8
No Resource Manager plan active
2021-07-10T18:06:39.177858+08:00


TT00 (PID:6201): Error 12541 received logging on to the standby
TT00 (PID:6201): Check whether the listener is up and running.
2021-07-10T18:06:39.185611+08:00
TT03 (PID:11461): Sleep 5 seconds and then try to clear SRLs in 3 time(s)
2021-07-10T18:06:39.733886+08:00
joxcsys_required_dirobj_exists: directory object exists with required path /opt/oracle/product/19c/dbhome_1/javavm/admin/, pid 6240 cid 1
replication_dependency_tracking turned off (no async multimaster replication found)
2021-07-10T18:06:42.855357+08:00
LOGSTDBY: Validating controlfile with logical metadata
LOGSTDBY: Validation complete
Starting background process AQPC
2021-07-10T18:06:43.050601+08:00
AQPC started with pid=50, OS id=11483
PDB$SEED(2):Autotune of undo retention is turned on.
2021-07-10T18:06:44.060956+08:00
PDB$SEED(2):Endian type of dictionary set to little
2021-07-10T18:06:44.314613+08:00
TT03 (PID:11461): Enabling archival of deferred physical standby SRLs
2021-07-10T18:06:44.532641+08:00
TT03 (PID:11461): Archived Log entry 260 added for T-1.S-238 ID 0xa7521ccd LAD:1
2021-07-10T18:06:44.681960+08:00
PDB$SEED(2):Undo initialization finished serial:0 start:1416557 end:1416557 diff:0 ms (0.0 seconds)
PDB$SEED(2):Pluggable database PDB$SEED dictionary check beginning
2021-07-10T18:06:48.986311+08:00
PDB$SEED(2):Pluggable Database PDB$SEED Dictionary check complete
PDB$SEED(2):Database Characterset for PDB$SEED is AL32UTF8
2021-07-10T18:07:04.127503+08:00
PDB$SEED(2):Opening pdb with no Resource Manager plan active
2021-07-10T18:07:05.037256+08:00
QPI: opatch file present, opatch
QPI: qopiprep.bat file present
2021-07-10T18:07:06.501677+08:00
Starting background process CJQ0
2021-07-10T18:07:06.511870+08:00
CJQ0 started with pid=47, OS id=11538
Completed: alter database open

Note the messages 

 PR00 (PID:11393): Terminal Recovery:
2021-07-10T18:05:51.685886+08:00
Recovery of Online Redo Log: Thread 1 Group 4 Seq 238 Reading mem 0
  Mem# 0: /opt/oracle/oradata/STDBYDB/stdbredo01.log
Terminal Recovery finished with No-Data-Loss
2021-07-10T18:05:51.985957+08:00
Incomplete Recovery applied until change 12767732 time 07/10/2021 17:57:59
2021-07-10T18:05:52.074427+08:00
Media Recovery Complete (STDBYDB)
2021-07-10T18:05:52.655358+08:00
Terminal Recovery: successful completion

Standby terminal recovery start SCN: 12767670

Standby became primary SCN: 12767669

Failover completed with No-Data-Loss.
Completed: alter database failover to STDBYDB

The Standby completed Recovery to Sequence#238 and SCN 12767669 which are much lower then the last known ArchiveLog (Sequence#243) and SCN (something higher than 12768459).

I had expected to see an ORA-16472 error requiring the usage of the "ALTER DATABASE ACTIVATE STANDBY DATABASE" command.  Unfortunately, in this case, the Standby assumes that there are no transactions from the Primary and assumes that there is no Data Loss !

So, the next time you do a Failover, you have to be careful to identify whether there is actual Data Loss if you are unable to access the Primary database.

Now, at the Standby I can query

SQL> select scn_to_timestamp(12767669) from dual;

SCN_TO_TIMESTAMP(12767669)
---------------------------------------------------------------------------
10-JUL-21 05.48.25.000000000 PM

SQL>

which certainly tells me that the last transaction at the Standby is as of  05:48pm (approximately,  the SCN_TO_TIMESTAMP provides us an approximate time only), even though I startup the Standby at 06:04pm.  

Fortunately, when I can access the Primary and read the alert.log, I find that the Primary went down after 06:02pm, with the last ArchiveLog generated at 6:00pm

2021-07-10T18:00:12.509208+08:00
Thread 1 advanced to log sequence 244 (LGWR switch)
  Current log# 1 seq# 244 mem# 0: /opt/oracle/oradata/ORCLCDB/redo01.log
2021-07-10T18:00:12.634091+08:00
ARC2 (PID:4347): Archived Log entry 529 added for T-1.S-243 ID 0xa7521ccd LAD:1

So I can confirm that there IS Data Loss !

FYI , ORA-16472 is :

oracle19c>oerr ora 16472
16472, 0000, "failover failed due to data loss"
// *Cause:  The database was in MaxAvailability or MaxProtection mode but
//          terminal recovery finished with data loss.
// *Action: For a data loss failover, issue ALTER DATABASE ACTIVATE STANDBY
//          DATABASE command.
oracle19c>

This signifies that a FAILOVER command has not succeeded and an ALTER DATABASE ACTIVATE STANDBY DATABASE must be issued.