Delayed fallback timer

You can use a delayed fallback timer to set the time for a resource group to fall back to a higher priority node. You can configure the fallback behavior for a resource group to occur at a predefined recurring time (daily, weekly, monthly, or a specific date).

The delayed fallback timer has the following characteristics:

• Specifies the time when a resource group that is online and residing on a nonhome or low priority node falls back its home node or a higher priority node.
• Affects the movement of the resource group to another node. For example, if you move the nonconcurrent resource group (that has a fallback timer attribute) to another node by using the resource group management (RGM) utility (clRGmove), the group stays on the destination node (unless you reboot the cluster, which is rarely done). If the destination node goes down and then is reintegrated, the resource group also falls back to this node at the specified time.

Node reintegration with a delayed fallback timer set

The resource group does not immediately fall back to its higher priority node under the following condition.

• You have configured a delayed fallback timer for a resource group.
• A higher priority node joins the cluster.

  At the time specified in the Delayed Fallback Timer attribute, one of two scenarios takes place:

• A higher priority node is found. If a higher priority node is available for the resource group, PowerHA® SystemMirror® attempts to move the resource group to this node when the fallback timer expires. If the acquisition is successful, the resource group is acquired on that node.

  However, if the acquisition of the resource group on the node fails, PowerHA SystemMirror attempts to move the resource group to the next higher priority node in the group node list, and so on. If the acquisition of the resource group on the last node that is available fails, the resource group goes into an error state. You must take action to fix the error and bring such a resource group back online.

• A higher priority node is not found. If there are no higher priority nodes available for a resource group, the resource group remains online on the same node until the fallback timer expires again. for example, if a daily fallback timer expires at 11:00 p.m. and there are no higher priority nodes available for the resource group to fallback on, the fallback timer recurs the next night at 11:00 p.m.

  A fallback timer that is set to a specific date does not recur.