When a resource goes from a running to a non-running state, while the intent to have it running remains unchanged, this transition is called a resource failure. At this point, Oracle Clusterware applies a resource state recovery procedure that may try to restart the resource locally, relocate it to another server, or just stop the dependent resources, depending on the high availability policy for resources and the state of entities at the time.
When two or more resources depend on each other, a failure of one of them may end up causing the other to fail, as well. In most cases, it is difficult to control or even predict the order in which these failures are detected. For example, even if resource A depends on resource B, Oracle Clusterware may detect the failure of resource B after the failure of resource A.
This lack of failure order predictability can cause Oracle Clusterware to attempt to restart dependent resources in parallel, which, ultimately, leads to the failure to restart some resources, because the resources upon which they depend are being restarted out of order.
In this case, Oracle Clusterware reattempts to restart the dependent resources locally if either or both the hard stop and pullup dependencies are used. For example, if resource A has either a hard stop dependency or pullup dependency, or both, on resource B, and resource A fails because resource B failed, then Oracle Clusterware may end up trying to restart both resources at the same time. If the attempt to restart resource A fails, then as soon as resource B successfully restarts, Oracle Clusterware reattempts to restart resource A.