Metro-cluster based on synchronous replication of virtualized storage processors

摘要

A technique for managing failover from a first data storage system to a second data storage system includes configuring replication settings for data objects associated with VSPs (Virtualized Storage Processors) on a per-VSP basis, such that all of the data objects associated with each VSP share common replication settings. As the first data storage system operates a VSP, operations replicate data written to any of the data objects of the VSP to the second data storage system such that the second data storage system acquires and stores a replica of the VSP. In the event of a failure at the first data storage system, operations fail over the entire VSP, including all of its data objects, to the second data storage system, to restore host access to the data objects from the VSP replicas. A metro-cluster may be formed from multiple data storage systems that use this failover approach.

主权项

1. A method of managing failover of VSPs (Virtualized Storage Processors) from a first data storage system to a second data storage system, comprising:
establishing synchronous replication settings for each of a set of VSPs in the first data storage system on a per-VSP basis; operating the set of VSPs in the first data storage system, each of the set of VSPs including multiple data objects, each data object being any of a file system, a LUN (logical storage unit), and a vVOL (virtual volume); receiving IO requests at the first data storage system designating data to be written to data objects of the set of VSPs operating in the first data storage system; synchronously replicating the IO requests received at the first data storage system to the second data storage system in accordance with the synchronous replication settings for each respective one of the set of VSPs to create a passive replica of each of the set of VSPs at the second data storage system; and in response to a failure of the first data storage system, activating the replica of each of the set of VSPs at the second data storage system to provide active operation of the set of VSPs at the second data storage system, wherein, when receiving the IO requests, each of the set of VSPs (i) includes a respective grouping of data objects and a respective set of host interfaces and (ii) is operated on a physical storage processor in the first data storage system, wherein each of the set of VSPs interacts with multiple host computing devices via the set of host interfaces to process IO requests received from the host computing devices, and wherein activating the replica includes starting the set of host interfaces of each of the set of VSPs on a physical storage processor in the second data storage system to enable each VSP to process IO requests from the host computing device directed to the respective data objects.