VMware has done a great job of maximising the up time of applications with technologies such as vMotion and HA, but historically they have not been much use when it came to protecting against the loss of all hosts, the network or a storage array. Their answer was to leverage storage array replication combined with vCenter Site Recovery Manager (SRM) to automate disaster recovery fail-over, fail-back and testing.
Certainly a much better solution than what we had before, but it was still a manual process that required the passive storage replica to be promoted to be active and the VMs re-started – downtime therefore could not be avoided. Ideally we should be able to:
- Automate site fail-over with vSphere HA thus avoiding deploying SRM
- Move VMs between sites to avoid downtime using vSphere vMotion
- Re-direct storage requests to the remote copy in the event of a failure
VMware’s answer to the problem is a vSphere Metro Storage Cluster (vMSC) which is a certified solution that combines synchronous replication with array-based clustering. The goal of the solution is to reap the same benefits that a single site/array vSphere high-availability cluster provides and extend it across two storage arrays, typically deployed at two data centres in different locations. At its core, a vMSC infrastructure is a stretched cluster that enables 24×7 continuous availability, including support for:
- vSphere vMotion, HA, DRS and Fault Tolerance over distance
- Transparent storage failure protection (i.e. RAID Group, shelf or complete system)
Once you have your vMSC certified storage and vSphere cluster you just need to add a stretched layer 2 network and away you go.
There are a few scenarios for deploying a vMSC as follows:
- Between racks
- Will protect against complete storage array failure
- Between buildings within the same site
- Will protect against complete storage array failure along with building related issues such as power, flood, fire, etc.
- Between data centres up to a few hundred kms apart
- Will protect against complete storage array failure along with site related issues such as power, flood, fire, etc.
To learn more download the VMware vSphere 6.0 Metro Storage Cluster Recommended Practices white paper here
I am convinced that as these solutions become more affordable we will start to see legacy active/passive storage replication technology being replaced with metro storage clusters.
Related Posts
- Cost-effective Metro Storage Cluster solution now available from NetApp
- Comparing NetApp MetroCluster with EMC VPLEX Metro Continuous Availability solutions
- An introduction to VMware NSX Software-Defined Networking technology
by 