Oracle Golden Gate 21.3 Microservice Replication Setup walk through

In this blog we will walk through the steps involved in Oracle Golden Gate 21.3 replication setup.

Browser based action:

As per oracle instruction, we take look at the admin service port

 

Click on the url opens admin console.

 

Overview page looks like below...

Click three dash [application navigation] and go to configuration

Add new credential

Enter the user id in the form of EZConnect. I have created the credential for CDB connection, Source PDB and target standalone setup.

 

The ggadmin credentials needs to be tested as shown below. All the connection credentials to be tested; to be sure it works. Then remove the lock

Like below

 

Add TRANDATA (source) ,CHECKPOINT (target)  & HeartBeat by following the below screenshots:

Connect to source PDB by clicking on the action button corresponding to the ggadmin_opdb1 connection alias:

Search for the schema to be used  for replication (TRANDATA):

Irrespective, if your using a CDB connection prefix PDB name, if we are using PDB connection. My search with or without prefix schema search works just fine.

Add checkpoint table (target):

 

Add HeartBeat tables to both source and target (connect to source and target one by one add them):

Like below..

 

After HeartBeat is added, the icons will be greyed out:

Step Extract:

Go back to overview section in Admin UI

Click on add symbol near EXTRACT:

 

Leave the default selection “Integrated Extract”

Click next and fill highlighted fields. Give a name to the extract source process, location where the trail files gets stored and the username to use for extract. Remember in source if it is Container deployment, we have to use CDB connection, hence we chose cggadmin credential alias.

Choose PDB for replication as well.

By default you get the below on next screen:

We have to add the below settings..

From guide:
SOURCECATALOG DBEAST
DDL INCLUDE MAPPED
TABLE hr.*;
 
Actual:
SOURCECATALOG OPDB1
DDL INCLUDE MAPPED
TABLE OPDB1_PUBLIC_USER.*;

Click create and run to launch extract.

1 Extract process launched. We can see extract trail files as well.

[oracle@vcentos79-oracle-sa1 src]$ ls -altr

total 4

drwxr-x---. 3 oracle oinstall   17 Jun 18 20:22 ..

drwxr-x---. 2 oracle oinstall   25 Jun 18 20:22 .

-rw-r-----. 1 oracle oinstall 1329 Jun 18 20:22 es000000000

[oracle@vcentos79-oracle-sa1 src]$ pwd

/oggdata/oggdep01/var/lib/data/src

[oracle@vcentos79-oracle-sa1 src]$

 

Add Replicat process now:

Ensure checkpoint table is already added
Ensure to copy the schema to target db before (initial loading should be done)
[note: in a low memory system, you may hit UDE-31623 , in such case stop GG extract and retry datadump during initial load]

Click on + sign to add new replicat process:

Select parallel non-integerated replication option

 

The highlighted options need to be modified..

 

Next brings the below default info populated

 

Ex. (as shown in guide):
 
REPLICAT repe
USERIDALIAS ggwest DOMAIN OracleGoldenGate
--DDL EXCLUDE ALL
DDLERROR default discard
REPERROR (default,discard)
DDLOPTIONS REPORT
SOURCECATALOG DBEAST
MAP hr.*, TARGET hr.*;
 
Actual:
REPLICAT rept
USERIDALIAS ggadmin_DBCASLNT DOMAIN OracleGoldenGate
--DDL EXCLUDE ALL
DDLERROR default discard
REPERROR (default,discard)
DDLOPTIONS REPORT
SOURCECATALOG OPDB1
MAP OPDB1_PUBLIC_USER.*, TARGET OPDB1_PUBLIC_USER.*;

 

Added the below param and just created it and not run yet. Since my intial data copy is still inprogress.

 

 

Param looks like below..

Now the inital sync is done:

Source:

18-JUN-23 20:56:57.459:   /oraarch/OPDB1_PUBLIC_USER_18Jun23_01.dmp

18-JUN-23 20:56:57.513: Job "SYS"."SYS_EXPORT_SCHEMA_01" successfully completed at Sun Jun 18 20:56:57 2023 elapsed 0 00:03:07

 

Target:

18-JUN-23 21:18:17.062: W-1      Completed 11 SCHEMA_EXPORT/TABLE/TABLE_DATA objects in 785 seconds

18-JUN-23 21:18:17.266: Job "SYS"."SYS_IMPORT_SCHEMA_01" successfully completed at Sun Jun 18 21:18:17 2023 elapsed 0 00:13:48

 

Now let us kick off both extract and replicat process:

 

Extract messages in source alert log:

OGG Capture client successfully attached to GoldenGate Capture OGG$CAP_EXTS to receive uncommitted changes with pid=60 OS id=28730.

GoldenGate OGG$CAP_EXTS Capture sends LCRs using streaming mode (OCI call duration=60 secs, bufsize=1024000).

LOGMINER: session#=1 (OGG$CAP_EXTS), reader H001 pid=65 OS id=28743 sid=108 started

LOGMINER: Begin mining logfile for session 1 thread 1 sequence 76, /oraarch/ORA19C/archivelog/2023_06_18/o1_mf_1_76_l8yr6hmj_.arc

LOGMINER: End   mining logfile for session 1 thread 1 sequence 76, /oraarch/ORA19C/archivelog/2023_06_18/o1_mf_1_76_l8yr6hmj_.arc

LOGMINER: Begin mining logfile for session 1 thread 1 sequence 77, /oradata/ORA19C/redo02.log

 

 

We can see both the Extract and Replicat are running.

Golden Gate is complaining about streams pool not configured. But the target is very small, hence I left it as it is.
 
Test replication:

Source:
SQL> insert into "OPDB1_PUBLIC_USER"."JT1"
  2  select * from "OPDB1_PUBLIC_USER"."JT1" where rownum < 2;
 
1 row created.
 
SQL> commit;
 
Commit complete.
 
SQL>

 

Extract statistics:

Replication statistics:

Source:
SQL> select count(1) from "OPDB1_PUBLIC_USER"."JT1";
 
  COUNT(1)
----------
  25000001
 
SQL>

 

Target:
SQL> select count(1) from "OPDB1_PUBLIC_USER"."JT1";
 
  COUNT(1)
----------
  25000001
 
SQL>

 

This completes replication setup – a quick setup.

We saw a DML. Next we will check how a DDL is handled.