Failover and Failback in Windows XP and Windows Server 2003
Applies To: Windows 10, Windows 7, Windows 8, Windows 8.1, Windows Server 2008, Windows Server 2008 R2, Windows Server 2012, Windows Server 2012 R2, Windows Server Technical Preview, Windows Vista
Microsoft Windows XP, Windows Server 2003, and subsequent versions are designed to handle recovery of the DTC when a node in the cluster fails for any reason. The two mechanisms it uses to perform recovery are failover and failback.
Failover
When the node that is designated as the owner of the DTC resource goes down, Windows XP, Windows Server 2003, and subsequent versions automatically designate another node in the cluster as owner of the resource and the DTC transaction manager is automatically restarted on another node in the cluster. During this operation, the DTC resource and its dependencies are automatically moved to the new owner. This automatic process of moving the DTC resource and its dependencies to another node in the cluster is called failover.
The newly restarted DTC transaction manager reads the DTC log file on the shared disk to determine the outcome of pending and recently completed transactions. Resource managers reconnect to the DTC transaction manager and perform recovery to determine the outcome of in-doubt transactions. Applications reconnect to the DTC so that they can initiate new transactions.
For example, in the following illustration, the DTC transaction manager is active on System B. The application program and resource manager on System A call the DTC proxy. The DTC proxy on System A forwards all DTC calls to the DTC transaction manager on System B.
.jpeg "<No Change>")
If System B fails, the DTC transaction manager on System A takes over. It reads the DTC log file on the shared disk, performs recovery, and then serves as the transaction manager for the entire cluster.
In choosing an alternate node for failover, Windows XP and Windows Server 2003 are configured by default to adhere to a defined failover policy. If you use the default policy, the selection of the failover node appears to be random. If you want more control over which node is chosen for failover, you can use the Cluster Administrator tool to alter the policy.
Failback
When a failed DTC owner in the cluster comes back up later, you might want the DTC resource and its dependency resources to be returned to the preferred node, usually the node on which they resided prior to the failover. The process of moving a DTC resource and its dependencies to the preferred node is called failback.
By default, the DTC resource and its dependencies remain on the node to which they were moved during failover. However, you can use the Cluster Administrator tool to configure Windows XP, Windows Server 2003, and subsequent versions to perform failback automatically.
See Also
Changing DTC Resource Dependencies
Cluster Administrator and DTC Configurations
DTC Transaction Manager on a Windows Server Cluster
Managing DTC Resources in Windows XP and Windows Server 2003
Managing a DTC Resource When a Node Is Evicted
Prerequisite Considerations for Adding a DTC Resource to a Cluster