Transactional failover of data sets

摘要

A network storage server implements a method to perform transactional failover of data sets. Multiple storage objects are organized into primary and secondary data sets, and a disaster recovery policy is configured for failing-over a primary data set to a secondary data set. A failover operation is defined for the disaster recovery policy. The failover operation includes multiple failover actions. During a failover situation, the failover operation is invoked to fail-over the primary data set. The failover operation is transactionally processed to ensure that all failover actions of the failover operation are performed in a single transaction.

主权项

1. A method comprising:
replicating a primary data set stored in a first storage unit of a network storage system to a secondary data set stored in a second storage unit of the network storage system, wherein the primary data set and the secondary data set include data associated with data services provided to clients by a primary storage server and a secondary storage server, respectively, of the network storage system; receiving a signal to failover the primary data set to the secondary data set; identifying a predefined failover operation created specifically for the primary data set based on a disaster recovery (DR) policy, wherein the predefined failover operation includes multiple failover actions to failover the primary data set to the secondary data set including stopping execution of the data services on the primary data set, processing a failover script that has been predefined by a user, requesting the secondary data set to be in a ready state for the data services, breaking the replication relationship between the primary data set and the secondary data set, and activating the data services on the secondary data set; continuously maintaining a state of the primary data set during the failover using an internal parameter of the primary data set, wherein the state of the primary data set indicates a status of the predefined failover; and performing the predefined failover operation to failover the primary data set to the secondary data set in an event the state of the primary data set is in a ready state, wherein performing the predefined operation includes executing the failover actions when the predefined failover operation is invoked, and wherein in an event that any one of the failover actions fails to complete successfully, the state of the primary data set is set to a failover error state, the predefined failover operation is aborted, and the primary data set and the secondary data set are reverted to conditions of the primary data set and the secondary data set, respectively, prior to the performance of the predefined operation.