EHC Disaster Recovery as-a-Service Overview

The Enterprise Hybrid Cloud DRaaS solution leverages DellEMC RecoverPoint for Virtual Machines-based disaster recovery to restart or recover VMware virtual machine workloads and cloud management infrastructure to another converged infrastructure platform on an alternate site.

Workloads using this protection service in vRealize Automation (vRA) may operate on either of the two sites participating in any given DR relationship such that:

  • Virtual machine workloads may be individually recovered on either of the two sites.
  • Virtual machine backup images are replicated to the other site and are available for restore

The diagram below displays the architecture for a VxBlock or VxRack dual-site EHC DRaaS with RP4VM.

Badger - VCEOverlay - 03 - RP4VMs - VxBlock (Fig 42)

What gets Replicated?

In this architecture, the only objects replicated and recovered between sites are the EHC Automation Cluster and EHC Tenant Workload Cluster components.

What does NOT get Replicated?

The AMP/Core cluster and NEI/Edge cluster components must exist simultaneously on each site, while the RP4VM vRPAs must also reside locally on separate ESXi clusters on their respective sites. No replication or recovery takes place between these clusters or components.

How is VMware NSX used in this solution?

This EHC DRaaS solution, using RP4VM, permits the use of cross-vCenter VMware NSX networking objects. The cross-vCenter NSX feature allows multiple NSX managers to be federated together in a primary/secondary relationship, thus enabling the deployment of NSX network and security components across multiple vCenters. This solution supports the use of both Universal and standard/local objects. Please refer to the EHC 4.1.1 Concepts and Architecture solution guide for more detail

How are the Virtual Machines protected?

RP4VM protects up to 2048 Virtual Machines at the VM level, using Virtual RecoverPoint Appliances (vRPAs) to split write operations at the VM kernel layer, copying the data from the local to the remote vRPA.

A vRP-wha?

  • A vRPA is a virtualised RecoverPoint appliance. RP running in a VM.
  • Each vRPA must run in a vRPA cluster (min:2, max:8 vRPAs per Cluster)
  • A vRPA system consists of paired vRPA clusters between 2 sites (2 vCenter Servers)
  • Each vRPA system can contain 128 Consistency Groups (up to total of 1024 CGs)
  • A total 0f 2048 VMs can be protected (currently)

vRPA_overview_01

Note: Diagram illustrates vRPA system allocation to protect EHC Pod components. vRPAs are not located in these EHC Pods. vRPAs reside in separate vSphere cluster to EHC Mgmt and Workload Pods

  • The EHC spec for a vRPA is 8 vCPU and 8GB RAM
  • A vRPA uses iSCSI connectivity between ESXi hosts
  • a vRPA uses an RP Splitter located in the ESXi kernel

More details on VIB installation and ESXi iSCSI Initiator Configuration via @DavidCRing

What happens to local Backups after a Failover?

Tenant workload clusters are logically paired across vCenters and sites to ensure networking and backup policies operate correctly both before and after failover. All local VM backups and policies are replicated between sites, meaning that after a failover, all existing backups continue to be available for restore purposes. All automated/scheduled backups also resume as normal, and will be replicated back to the original site when it is restored/available again.

How are the VMs Failed over?

The failover of the virtual machines is controlled by a vCenter administative user via the RecoverPoint for VM console in the VMware vCenter Web UI. From there the user initiates the failover at an RP4VM Consistency Group level.

tenant_vc01

The granularity of the fail over depends on how many VMs are protected in each Consistency Group.

How are the VMs protected initially?

A vRA Virtual Machine Blueprint must be created, within which the RP4VM Build Profile needs to be assigned. At deployment time, when a vRA user then selects to deploy a VM from that RP4VM-enabled VM blueprint, they can select the relevant RP4VM Policy and Consistency Group , as shown belowRP4VM_BP_buildProfile01

DRaaS protection can be applied to existing vRA VMs as well as new VMs. All managed through the vRA portal.

Please refer to the EHC 4.1.1 Solution Admin Guide for more details covering:

  • Creation of RP4VM Policies,
  • Creation of DR-protected VM Blueprints
  • Automated DR-protected protection of Virtual Machines at deployment time in vRA
  • Applying DRaaS protection to existing Virtual Machines in vRA.

 

Related: For RP4VM integration with vRealize Operations Manager, see here

 

Hope that helps!

 

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s