Method and system for stateful recovery and self-healing

摘要

The ability to perform non-destructive editing of files and models requires the generation and persistence of input deltas that capture changes that are made to a base starting point. Reconstitution of saved state may be achieved through the application of deltas. This capability is useful for failover remediation in client/server environments since the client has access to the deltas, such that in the event that a stateful service becomes unresponsive (and therefore, no longer available), the service may be taken offline and a new resource may be assigned as a replacement. In such an event, the service is directed to load the baseline data and any changes may be reapplied, restoring the service state.

主权项

1. A computer-implemented method of managing process plant switchover data for a user group of at least one user comprising:
a. forming a set of change groups, made through a user interface by a given user of the user group, wherein a given change group of a set of change groups is associated with the given user of the user group, wherein the given user of the user group makes changes to a model, starting from a baseline that is retrieved from a storage group, and the model is associated with an application instance; b. storing the baseline in a shared file storage included in the storage group; c. storing, upon a save operation, a first set of changes from the given change group, to a user-authenticated content storage included in the storage group; d. continuously tracking changes made after the save operation; e. storing both a remaining set of changes from the given change group, including the changes made after the save operation, and the first set of changes from the given change group, to a local persistent memory included in the storage group; and f. detecting a switchover event, through a resource manager that identifies the application instance as an application instance identified for removal, the resource manager communicating with an adapter paired with the application instance; the adapter intercommunicating with both the application instance and the resource manager, the adapter performing one or more of terminating, restarting, or repairing the application instance asynchronously with respect to the resource manager obtaining a new application instance, the adapter monitoring both the adapter itself and the application instance, and the adapter detecting an additional switchover event and terminating the application instance in response to said detecting the additional switchover event.