The administration journal contains a log of past global administration actions, i.e. the history of the administration actions. openUTM sets up the administration journal under the filebase name of the associated UTM cluster application the first time the first node application is started (see also the section "UTM cluster files" (Files)).
Like all files that are global to the cluster, the administration journal is located on a storage medium which can be accessed by all node applications (see the section "Runtime environment"). The UTM system code handles concurrent accesses via NFS4 locks.
All node applications reconstruct the administrative changes that have global application on the basis of the administration journal.
Running applications apply these actions with minimal delay. They do this at the latest before they initiate global administration actions themselves. Depending on the load on a node, this will generally be done within a few seconds.
They are notified of the need to do so by the node application that was administered directly.A network problem can occasionally cause this notification to be lost. For this reason, and depending on the
CHECK-ALIVE-TIMER-SEC operand of the CLUSTER statement, the administration journal is checked at regular intervals by the running node applications.Node applications that are subsequently started apply the changes during the startup phase.