If you’re a frequent visitor of my blog you might have read my article “Future of Disaster Recovery with NextGen Site Recovery Manager“. In this article I discussed vSphere Metro Storage Cluster (vMSC) support for SRM. With the release of SRM 6.1 vMSC or “stretched cluster” support is here, next to some other additions like Storage Policy Based Protection Groups and enhanced NSX integration. Read on to learn more about these great improvements!
Stretched Storage Support
Let’s just jump straight into the hottest improvement in SRM 6.1: stretched cluster support! Stretched cluster configurations remain a popular option for specifically disaster avoidance scenarios. Stretched cluster configurations are less suitable for disaster recovery scenarios, because a stretched cluster is relying on VMware High Availability for recovery. Another common challenge with a stretched cluster configuration is that you have just one vCenter Server and that DRS is not site aware. When you compare a stretched cluster setup with an SRM implementation, you will also notice the lack of orchestration and testability options.
SRM on its turn is very good for disaster recovery, has two vCenters in place and includes a repeatable (orchestrated) DR plan. There are also some challenges with SRM; such as no options for cross-site load balancing, disaster avoidance requires downtime and there’s no option to move applications between sites without downtime.
With SRM 6.1 the best of both worlds come together; SRM offers the integration of a stretched cluster configuration. This results in a unified plan for disaster avoidance, disaster recovery and mobility. You can now do disaster avoidance without downtime; in case disaster avoidance is not possible you can switch over to the pre-configured disaster recovery scenario. Note that a vMotion between sites in case of a disaster is only possible using the planned migration option.
I will sure look from some additional info on this very interesting topic, so stay tuned. For now also check last year’s article on this topic.
Storage Policy Protection Groups
Site Recovery Manager 6.1 now supports Storage Policy Protection Groups, or SPPGs. This works as follows:
- First, tag a virtual machine;
- Then tag a datastore;
- Now protect the datastore with SRM.
Each VM that now has the tag that is linked to the datastore protected by SRM will be automatically moved to the datastore using storage vMotion. SRM will pick up the virtual and automatically include the virtual machine in the protection plan.
To unprotect a virtual machine just remove the tag from the virtual machine. Removing the tag will move the virtual machine from the datastore; protection is automatically removed as well.
SPPG is only supported for array based replication. You can create recovery plans solely based on SPPGs.
Enhanced VMware NSX Integration
On August 20 VMware release NSX 6.2, VMware’s network virtualization solution. NSX 6.2 includes several features we have all been waiting for and all have to do with cross vCenter network virtualization. NSX 6.2 with vSphere 6 supports cross vCenter NSX, consistent firewall policy across vCenters and cross vCenter vMotion with the Distributed FireWall. New constructs in NSX 6.2 are Universal Security Groups, Universal Logical Switch and Universal Distributed Logical Router.
NSX 6.2 will make a lot better fit with SRM. NSX 6.2 will now also support a secondary NSX manager. You can loose the primary manager; the secondary will take over the primary manager. All networking/firewalling information will be synchronized between NSX manager instances. Together with Universal Logical Switch and Universal Distributed Logical Router you can now create stretched network across multiple vCenters which make your life much more easier when implementing SRM. The following image illustrates the concept of the a Universal Logical Switch:
vSphere Replication RPO Improvements
vSphere Replication now support an RPO of 5 minutes when replication between VSAN datastores.