Recovery point objective via dynamic usage of bind segments in a global mirror environment

Abstract

Embodiments for an approach to resource optimization during consistency group formation associated with a global mirror environment is provided. The approach detects when a primary volume associated with the consistency group has completed transmitting its out-of-sync (OOS) data towards its associated secondary volume. A command is sent to create a next consistency group sidefile so further writes can be sent to the sidefile rather than queueing at the primary volume. The approach repeats this process for each primary volume associated with the global mirror environment until all primary volumes are complete. Commands are sent to disable writes to the associated sidefiles and to merge the data stored in the sidefiles into normal cache as the next consistency group becomes the current consistency group.

Description

BACKGROUND OF THE INVENTION

[0001] The present invention relates generally to global mirror environment optimization and more specifically, to improved recover point objective (RPO) based on increased bandwidth utilization.

[0002] A global mirror environment is a multi-site data replication solution that scales and provides a cross volume / storage subsystem with data integrity and data consistency. During consistency group formation, different primary volumes complete their data transfers at different times, such that primary volumes that have completed their transfers will not transfer any more data until all the primary volumes associated with the current consistency group have completed their data transfer. Accordingly, writes to primary volumes that have completed their data transfers for the current consistency group are queued, leading to an increase in idle bandwidth as a greater number of primary volumes complete their data transfer for the current consistency group.SUMMARY

[0003] ...

Images