Failback moves the primary node’s resources from the secondary node back to the primary node and starts the primary Adaptive Server. IRIX FailSafe and Sybase’s Failover software offer two different methods to fail back:
Automatic – Failback controlled by IRIX FailSafe.
Manual – Failback handled by IRIX FailSafe in conjunction with the DBA issued sp_companion commands.
The following table describes the transition from an initial state of failover, to failback, and then to steady-state. This example uses two resources groups, pri_rg (the primary node) and sec_rg (the secondary node). The pri_rg is on the primary node and the sec_rg is on the secondary node.
Table 15-1 describes the automatic transition states.
Action |
Primary node |
Secondary node |
Primary server |
Secondary server |
|---|---|---|---|---|
Initial state |
The node is down |
The node is up and the pri_rg and sec_rg are online |
Not running |
Adaptive Server is running the Primary and Secondary databases. |
FailSafe Services started on the Primary Node |
The node is up with no resource groups assigned |
No change |
No change |
No change. |
The DBA issues a prepare_failback on the secondary server |
No change |
No changes |
No change |
The primary server’s databases and devices are released by the secondary server. |
The primary resource group pri_rg is taken offline |
No change |
The primary server’s resources are dismounted |
No change |
No change. |
The primary resource group pri_rg is brought online |
The node is up and the pri_rg is online |
The node is up and the sec_rg is online |
Adaptive Server is started |
No change. |
The DBA issues a resume on the primary companion. |
No change |
No changes |
The user databases are brought back online. |
No change. |
Final State |
The node is up and the pri_rg is online |
The node is up and the sec_rg is online |
Adaptive Server is running the primary databases |
Adaptive Server is running the secondary databases. |
Preform the following steps to move from the failed-over state to a normal state.
