Forum Stats

  • 3,854,538 Users
  • 2,264,379 Discussions
  • 7,905,721 Comments

Discussions

Downgrade database with standby

4331679
4331679 Member Posts: 3
edited Sep 24, 2020 5:40AM in Database Upgrade

Hello all,

Please help me understanding the following

In oracle documentation it is mentioned

ref.

#https://docs.oracle.com/en/database/oracle/oracle-database/12.2/sbydb/upgrading-patching-downgrading-oracle-data-guard-configuration.html#GUID-FDB0FA2C-EA66-423C-AFD9-51F3FA4899F8

#Data Guard configuration that does not contain a logical standby database.

#Ensure that all physical standby databases are mounted,

#but not open. Do not open the standby database(s) until prim and standby are in sync

#Downgrade the primary database

#keeping the following in mind:

#At each step of the downgrade procedure where a script is executed,

#execute the script only at the primary database.

#Do not perform the next downgrade step until all redo generated by

#the execution of the script at the primary database has been applied

#to each physical standby database.

#At each step of the downgrade procedure where an action other than

#running a script is performed, perform the step at the primary database

#first and then at each physical standby database.

#Do not perform the next downgrade step at the primary database

#until the action has been performed at each physical standby database.

Instead can this be done and is there any disadvantage or impact to do it ?

shutdown standby databases, downgrade primary database, Then later mount standby database from old home and apply all the logs generated during downgrade of primary

Answers

  • Daniel Overby Hansen-Oracle
    Daniel Overby Hansen-Oracle Posts: 93 Employee
    edited Sep 24, 2020 5:40AM

    Hi,

    Your approach won't work. You have to follow the procedure from the documentation. Example: you want to downgrade from e.g. 19c to 11.2.0.4.

    When you downgrade the database is started on 19c binaries, and are generating 19c-format redo logs. Those redo logs can be applied by a standby database running on the 19c binaries. If you start the standby database using 11.2.0.4 binaries (even after primary has successfully downgrade, and running on 11.2.0.4 binaries), then the standby database won't understand the format for the 19c-format redo logs.

    After you have downgraded, you should shut down the primary database. Then ensure that all 19c-format redo are applied at all standby databases - and shut them down as well. Next, start primary in 11.2.0.4 and execute "catreload". Then start standby databases on 11.2.0.4 binaries and let the "catreload" propagate to the standby databases via redo apply.

    Regards,

    Daniel