Failing Back to the Primary Companion

When you register the logical hosts, disable the automatic failback option. Failback should be a planned event.

Failback moves the primary companion’s shared disks from the secondary node back to the primary node and starts the primary companion on the primary node.

After your primary host is ready to take over the primary companion, issue this command from the secondary companion:
```
sp_companion primary_companion_name, prepare_failback
```
Where primary_companion_name is the name of primary companion server.

This command moves the primary companion’s logical host to the primary host.

For example, to fail back the primary companion MONEY1, issue this command from the secondary companion PERSONEL1:
```
sp_companion MONEY1, prepare_failback
```
Make sure the primary companion’s logical host is moved successfully to the primary host by issuing this command:
```
haget -f mastered
```
The output shows the primary host monitoring the logical host of the primary companion.
Start the primary companion:
```
hasybase start primary_companion_name
```
For example to start the primary companion MONEY1:
```
hasybase start MONEY1
```
To resume normal companion mode, issue the following from the primary companion.
```
sp_companion secondary_companion_name, resume
```
Where secondary_companion_name is the name of the secondary companion server. For example, to resume normal companion mode for primary companion MONEY1:
```
sp_companion PERSONEL1, resume
```

You cannot connect clients with the failover property (for example isql -Q) until you issue sp_companion resume. If you do try to reconnect them after issuing sp_companion prepare_failback, the client hangs until you issue sp_companion resume.