DB2 10.5 for Linux, UNIX, and Windows

Removing members from a high availability disaster recovery (HADR) setup

Removing a member from your DB2® pureScale® instance requires you to reinitialize the standby based on the primary's updated topology.

About this task

To drop a member, you need to stop HADR and the DB2 pureScale instance. You cannot drop the last member in the instance using this procedure.

To remove a member from an HADR setup in a DB2 pureScale instance:

Remove the member from the primary cluster. You must do this from a host that will still belong to the instance after the member is dropped.
1. Stop HADR on the primary database using the STOP HADR command.
2. Stop the DB2 pureScale instance using the db2stop command.
3. Drop the member by running the following command:
```
db2iupdt -drop -m member_ID instance_name
```
Note: You cannot directly drop a member from an HADR standby database.
Remove the member from the standby cluster. You must do this from a host that will still belong to the instance after the member is dropped.
1. Deactivate the database on the standby database using the DEACTIVATE DATABASE command.
```
DEACTIVATE DATABASE db_name
```
2. Drop the database using the following command:
```
DROP DATABASE db_name
```
3. Drop the member by running the following command:
```
db2iupdt -drop -m member_ID instance_name
```
  You must use the same member ID that you specified when removing the member from the primary cluster.
Create the standby database by restoring a backup image or by initializing a split mirror, based on the primary's updated topology after step 1.
1. On the primary, issue the following command:
```
BACKUP DB dbname
```
2. Restore the standby by issuing the following command:
```
RESTORE DB dbname
```
Update the HADR-specific database configuration parameters on the standby cluster.
Start HADR on the primary:
```
START HADR AS PRIMARY
```
Start HADR on the standby:
```
START HADR AS STANDBY
```