HOL 19c - Fallback - Issues After Upgrade

« PREVIOUS «

Protection for issues after upgrade

In this part you’ll use again two techniques to protect your database, but this time for issues happening after the upgrade. You can call this also “Downgrade“.

1. Downgrade with a Full Database Export and Import

For this part you’ll just start the export from the 19c database after upgrading.

Run the full database export:
.

. ftex19
sqlplus / as sysdba

insert into SYSTEM.TRACKING_TAB values (4,'full export downgrade');
commit;
select * from TRACKING_TAB;
exit

expdp system/oracle DIRECTORY=EXP18 DUMPFILE=down.dmp LOGFILE=down.log VERSION=12.2 FULL=Y REUSE_DUMPFILES=Y EXCLUDE=STATISTICS LOGTIME=ALL

The important part is that the VERSION parameter tells Data Pump to create an export in the format a database of “VERSION” will understand.
In this case you will downgrade into a 12.2 Pluggable Database. PDB3 has been created already and should be started as well.
.

. cdb1
sqlplus / as sysdba

startup
alter session set container=pdb3;
create directory IMP19 as '/home/oracle/IMP';
grant read, write on directory IMP19 to public;
exit

impdp system/oracle@PDB3 DIRECTORY=IMP19 DUMPFILE=down.dmp LOGFILE=impdown.log LOGTIME=ALL

This was a quick exercise. Of course it would take longer the more data and objects your database contains. Especially LOB data types can be crucial.

2. Downgrade with the downgrade scripts

HOL 19c - Fallback - Issues After Upgrade

In this final exercise you’ll use a powerful technique, the database downgrade with downgrade scripts.

Set a marker in the database
.

. ftex19
sqlplus / as sysdba

insert into SYSTEM.TRACKING_TAB values (5,'database downgrade');
commit;
Run the downgrade script
.

shutdown immediate
startup downgrade
set echo on termout on serveroutput on timing on
spool /home/oracle/logs/downgrade.log
@?/rdbms/admin/catdwgrd.sql
shutdown immediate
exit
Switch to the source environment and start he bootstrap reload script
.

. ftex
sqlplus / as sysdba

startup upgrade
set echo on termout on timing on
spool /home/oracle/logs/relod.log
@?/rdbms/admin/catrelod.sql
shutdown immediate
Final steps and checks
.

startup
@?/rdbms/admin/utlrp.sql
select * from TRACKING_TAB;
select count(*) from DBA_OBJECTS where STATUS=’INVALID’;
select COMP_ID, STATUS from DBA_REGISTRY order by COMP_ID;

The downgrade should have removed the XDB component as well. Did it?