DISASTER RECOVERY FRAMEWORK FOR CLOUD DELIVERY

摘要

In an example embodiment, an instance of a first database and an instance of a first application in a first data center of a managed cloud are replicated to a second data center of the managed cloud. Then state information regarding the first application is stored in a network file system in the first data center. Interactions between a user and the first application are directed to the instance of the first application in the first data center. The state information is then updated based on the interactions, and any changes to the instance of the first database based on the interactions are replicated to the second data center. Then a disaster is detected in the first data center, and all interactions between the user and the first application are redirected to a second instance of the first application in the second data center caused by the replication of the instance of the first application in the first data center.

主权项

1. A method comprising:
causing replication of an instance of a first database in a first data center of a managed cloud of a particular customer network to a second data center of the managed cloud of the particular customer network; causing mirroring of an instance of a first application in the first data center to the second data center, thereby replicating completely any application setup and/or file system structure to the second data center;storing state information regarding the first application in a network file system in the first data center;directing interactions between a user and the first application to the instance of the first application in the first data center;
updating the state information based on the interactions, the updating being performed on a database level as well as on an application-specific repository level;causing replication to the second data center of any changes to the instance of the first database based on the interactions;
causing replication to the second data center of any changes to an instance of the first Application specific repository level based on the interactions; detecting a disaster in the first data center; and in response to the detecting of the disaster, redirecting all interactions between the user and the first application to a second instance of the first application in the second data center caused by the replication of the instance of the first application in the first data center, after making them the primary once in the course of the failover.