Understanding Pausing Nodes

A pause time is an interval of time during which the node becomes inactive. When the pause time begins, the node is shut down until the pause time is scheduled to end.

You might schedule a pause time to perform maintenance tasks or devote server resources to an important batch run. For example, say that you have a complex batch program that runs on the same server as a particular node every Monday morning from 12:05 a.m. to 3:30 a.m. To make sure that the batch program has enough memory devoted to it, you can set a pause time for the node that runs from 12 a.m. to 4 a.m.

During a pause time, transactions are not published or received by the local system. When the system is paused, the node cannot accept service operations sent to it. Consequently, the publishing node must attempt to send transactions again later. The publishing node continues to send transactions until it exceeds the local time-out period. When this happens, the transaction assumes a Timeout status in the publisher’s queue. The time-out period is an attribute of the publication queue, not the subscription queue.

If the system attempts to send a transaction while the node is paused, the system writes the transaction to the publication and subscription queues, but it cannot publish the transaction until the system is no longer in the paused state.

Note: Pause times do not appear in PeopleSoft Application Designer upgrade projects; you cannot upgrade them.

If you have domain failover configured for your integration system, it is disabled when a node is paused. However, there is a chance that failover could occur when you pause a node, due to the independent timing between domain failover logic and the PeopleSoft Integration Broker runtime environment. The chance of failover occurring during a node pause time increases if the failover time check (IB Failover Time) on the Failover Configuration page is set to a low number of minutes.