Scenario 7: VM7 is deployed to Cluster 7 VM1, VM2, VM3, VM4, VM5 and VM6 exist
4. Specific to DR:
a. Day 2 storage DRS migrations between datastores would break the Site Recovery Manager protection for the virtual machines moved.
b. Day 2 storage DRS migrations between datastores would result in re- replicating the entire virtual machine to the secondary site.
Multiple vCenter endpoints are supported within Federation Enterprise Hybrid Cloud. However depending on the topology chosen, there are certain considerations as outlined in this section.
These topologies can support:
Only one vCenter per tenant with the ability to execute STaaS services.
More than one vCenter per tenant as long as the second and subsequent vCenter endpoints require IaaS services only.
Enabling additional STaaS-enabled vCenter endpoints
To enable additional STaaS-enabled vCenter endpoints, these topologies require a separate vRealize tenant per vCenter endpoint for the following reasons:
Federation Enterprise Hybrid Cloud STaaS catalog items use vCenter Orchestrator through the vRealize Automation advanced server configuration, which only allows one vCenter Orchestrator to be configured.
This vCenter Orchestrator stores important vCenter configuration details gathered during the Federation Enterprise Hybrid Cloud Foundation installation process.
To store additional vCenter configuration details, an additional vCenter Orchestrator is required.
Specifying the additional vCenter Orchestrator as an advanced server configuration requires an additional tenant.
Each vCenter endpoint requires its own independent vCenter Orchestrator server and NSX Manager instance.
The vCenter Orchestrator consideration is based on the additional tenant consideration above.
The NSX Manager requirement is based on the VMware requirement for a 1:1 relationship between vCenter and NSX.
vSphere datastore clusters
Single-site/single vCenter and dual- site/single vCenter topologies
Enabling additional BaaS-enabled vCenter endpoints
These topologies require independent Avamar instances for each vCenter endpoint to enable BaaS services.
This topology can support:
Two vCenters per tenant with the ability to execute STaaS and BaaS services. More than two vCenters per tenant as long as the third and subsequent vCenter
endpoints only require IaaS services.
Enabling additional STaaS and BaaS-enabled vCenter endpoints
Additional STaaS and BaaS enabled vCenter endpoints require additional tenants and Avamar instances similar to the single vCenter topologies.
The following configurations are permitted for each Federation Enterprise Hybrid Cloud instance:
Local only (single-site/single vCenter) Local plus CA combined
Uses the CA dual-site/single vCenter topology and provides local-only and CA functionality via distinct Workload Pods
Local plus DR combined
Uses the DR dual-site/dual vCenter topology and provides local-only and DR functionality via distinct Workload Pods
Note: Federation Enterprise Hybrid Cloud 3.1 does not support both DR and CA functionality on the same Federation Enterprise Hybrid Cloud instance.
Both the distributed and collapsed management models can support all of the following topologies:
Single-site/single vCenter
Standard dual-site/single vCenter CA dual-site/single vCenter Standard dual-site/dual vCenter DR dual-site/dual vCenter Dual-site/dual vCenter topologies Combining topologies Management model and topology combinations
Single-site Federation Enterprise Hybrid Cloud deployments can be upgraded to CA dual- site/single vCenter topology by adopting either an online or offline upgrade approach with the following considerations.
Considerations
The topology upgrade is an EMC professional services engagement and provides three basic methods of conversion based on the original storage design
NFS to VPLEX Distributed VMFS (Online via Storage vMotion)
Standard VMFS to VPLEX Distributed VMFS (Offline via VPLEX encapsulation) VPLEX Local VMFS to VPLEX Distributed VMFS (Online via VPLEX Local to VPLEX
Metro conversion)
If NFS is in use for management platform storage, then new VPLEX storage is required.
In non-BaaS environments, local workloads can be migrated to new CA clusters using storage vMotion if required.
Note: Federation Enterprise Hybrid Cloud 3.1 does not currently provide an automated mechanism to achieve this. Contact EMC Professional Services to assist in this process. Existing local-only workload clusters may remain as local-only clusters or be converted
to CA-enabled clusters after the topology upgrade. Note: EMC Professional Services should execute this process
In BaaS environments, virtual machines requiring CA protection should remain on the original cluster and the cluster should be converted to a CA-enabled cluster.
This is due to the need to carefully manage the relationships of vSphere clusters, Avamar grids, Avamar proxies, and vCenter folder structure to preserve the ability to restore backups taken prior to the topology upgrade.
After the topology upgrade, new clusters can be provisioned to provide CA or local- only functionality for new tenant virtual machines.
Single-site Federation Enterprise Hybrid Cloud deployments can be upgraded to DR dual- site/dual vCenter topology by adopting with the following considerations:
Considerations
Additional Core and NEI Pod infrastructure and components need to be deployed on the second site.
Additional Automation Pod infrastructure needs to be deployed on the second site to become the target for the Automation Pod failover.
EMC RecoverPoint needs to be installed and configured and all Automation Pod LUNs replicated to the second site.
Note: If NFS volumes were used for Automation Pod storage then new FC-based block datastores should be provided, and the Automation Pod components migrated to the new storage using Storage vMotion.
Prior to the upgrade, the Automation Pod components must be deployed on a distinct network segment from the Core and NEI Pods.
Single site to continuous availability upgrade Single-site to disaster recovery upgrade
A Microsoft SQL Server instance and a vCenter Single Sign-On role must be deployed to a server in the Automation Pod during the initial deployment.
Migration of previously existing virtual machines from local to DR clusters is not currently supported with default functionality.
Note: If there is a requirement to DR-enabled pre-existing tenant workloads, contact EMC Services teams to provide this as custom functionality.
For environments that require existing virtual machines to be imported into the Federation Enterprise Hybrid Cloud, the bulk import feature of vRealize Automation enables the import of one of more virtual machines.
This functionality is available only to vRealize Automation users who have Fabric Administrator and Business Group Manager privileges. The Bulk Import feature imports virtual machines intact with defining data such as reservation, storage path, blueprint, owner, and any custom properties.
Note: While the Federation Enterprise Hybrid Cloud supports the import of existing workloads, it does not currently offer out-of-the-box functionality to apply data protection services such as backup or DR to the newly imported virtual machines. If additional Federation Enterprise Hybrid Cloud services for imported workloads are required, contact EMC Services teams.
Load balancers cannot be deployed as part of a protected multimachine blueprint. However, you can manually edit the upstream Edge to include load-balancing features for a newly deployed multimachine blueprint.
Failover state operations
Provisioning of virtual machines to a protected DR cluster is permitted at any time, as long as that site is operational. If you provision a virtual machine while the recovery site is unavailable due to vCenter Site Recovery Manager disaster recovery failover, you need to run the DR Remediation catalog item to bring it into protected status when the recovery site is back online.
During STaaS provisioning of a protected datastore, Federation Enterprise Hybrid Cloud workflows issue a DR auto-protect attempt for the new datastore with vCenter Site Recovery Manager. If both sites are operational when the request is issued, this should be successful. If, however, one site is offline (vCenter Site Recovery Manager Disaster Recovery Failover) when the request is made, the datastore will be provisioned, but you must run the DR
Remediation catalog item to bring it into a protected status.
Note: The DR Remediation catalog item can be run at any time to ensure that all DR items are protected correctly.
While replication is at the datastore level, the unit of failover for in a DR configuration is a DR-enabled cluster. It is not possible to failover a subset of virtual machines on a single DR- protected cluster. This is because all networks supporting these virtual machines are converged to the recovery site during a failover.
Importing from non-Federation Enterprise Hybrid Cloud environments Multimachine blueprints vRealize Automation Failover granularity
There is also a limit of 64 consistency groups per RecoverPoint appliance and 128 consistency groups per RecoverPoint cluster. Therefore, the number of nodes deployed in the RecoverPoint cluster should be sized to allow appropriate headroom for surviving appliances to take over the workload of failed appliance.
The Federation Enterprise Hybrid Cloud supports RecoverPoint CL-based licensing only. It does not support RecoverPoint SE or RecoverPoint EX, as these versions are not currently supported by EMC ViPR.
Protection maximums
Table 7 shows the maximums that apply for SRM-protected resources. SRM protection maximums
Table 7.
Total number of Maximum
Virtual machines configured for protection using array-based replication 5,000
Virtual machines per protection group 500
Protection groups 250
Recovery plans 250
Protection groups per recovery plan 250
Virtual machines per recovery plan 2,000
Replicated datastores (using array-based replication) 255 Recovery maximums
Table 8 shows the maximums that apply for SRM recovery plans. SRM protection maximums
Table 8.
Total number of Maximum
Concurrently executing recovery plans 10
Concurrently recovering virtual machines using array-based replication 2,000 Table 9 indicates the storage maximums in a Federation Enterprise Hybrid Cloud DR environment, when all other maximums are taken into account.
Implied Federation Enterprise Hybrid Cloud storage maximums Table 9.
Total number of Maximum
DR enabled datastores per RecoverPoint Consistency Group 1 DR enabled datastores per RecoverPoint Cluster 128
DR enabled datastores per EHC environment 250
To ensure maximum protection for DR-enabled vSphere clusters, the Federation Enterprise Hybrid Cloud STaaS workflows create each LUN in its own RecoverPoint consistency group. This ensures that ongoing STaaS provisioning operations have no effect on either the synchronized state of existing LUNs or the history of restore points for those LUNs maintained by EMC RecoverPoint.
Because there is a limit of 128 consistency groups per EMC RecoverPoint cluster, there is therefore a limit of 128 Federation Enterprise Hybrid Cloud STaaS provisioned LUNs per RecoverPoint cluster limitations RecoverPoint licensing VMware Site Recovery Manager limitations Implied Federation Enterprise Hybrid Cloud storage maximums
RecoverPoint cluster. To extend the scalability further, additional EMC RecoverPoint clusters are required.
Each new datastore is added to its own SRM protection group. As there is a limit of 250 protection groups per SRM installation, this limits the total number of datastores in a DR environment to 250, irrespective of the number of RecoverPoint clusters deployed.
Supports VNX and VMAX only.
Supports NSX only.
Only supports the assignment of blueprint virtual machines to a security group. Does not support the assignment of blueprints to security policies or security tags.
As vRealize Automation endpoints are visible to all vRealize Automation IaaS administrators, resource isolation in the truest sense is not possible. However, use of locked blueprints and storage reservation policies can be used to ensure that certain types of workload (such as those whose licensing is based on CPU count) can be restricted to only a subset of the Workload Pods available in the environment. This includes the ability to control those licensing requirements across tenants by ensuring that all relevant deployments are on the same set of compute resources.
All endpoints configured across the vRealize Automation instance by an IaaS administrator are available to be added to fabric groups, and therefore consumed by any business group across any of the vRealize Automation tenants.
Provisioning to vCenter endpoints, however, can still only be done through the tenant configured as part of the Federation Enterprise Hybrid Cloud foundation installation in that tenant and its vCenter Orchestrator server.
The Federation recommends that applications provisioned using vRealize Automation Application Services each have their own business group by application type to enable administrative separation of blueprint creation and manipulation.
The Federation Enterprise Hybrid Cloud supports physical Avamar infrastructure only. It does not support Avamar Virtual Edition
Federation Enterprise Hybrid Cloud 3.1 supports a maximum of 15 Avamar replication pairs (30 individual physical instances).
For information about qualified components and versions required for the initial release of the Federation Enterprise Hybrid Cloud 3.1 solution, refer to the Federation Enterprise Hybrid Cloud 3.1: Reference Architecture Guide. For up-to-date supported version information, refer to the EMC Simple Support Matrix: EMC Hybrid Cloud 3.1: elabnavigator.emc.com. Storage support Network support NSX security support Resource isolation Resource sharing Application tenant integration Supported Avamar platforms Scale out limits
Federation
Enterprise Hybrid Cloud software resources
For all Federation Enterprise Hybrid Cloud sizing operations, refer to the EMC Mainstay Sizing tool: mainstayadvisor.com/go/emc.
Federation
Enterprise Hybrid Cloud sizing
This chapter presents the following topic:
The Federation Enterprise Hybrid Cloud solution provides on-demand access and control of infrastructure resources and security while enabling customers to maximize asset use. Specifically, the solution integrates all the key functionality that customers demand of a hybrid cloud and provides a framework and foundation for adding other services. This solution provides the following features and functionality:
Continuous availability Disaster recovery Data protection
Automation and self-service provisioning Multitenancy and secure separation Workload-optimized storage Elasticity and service assurance Monitoring
Metering and chargeback
The solution uses the best of EMC and VMware products and services to empower customers to accelerate the implementation and adoption of hybrid cloud while still enabling customer choice for the compute and networking infrastructure within the data center.
This chapter presents the following topic:
These documents are available on EMC.com. Access to Online Support depends on your login credentials. If you do not have access to a document, contact your Federation representative.
Federation Enterprise Hybrid Cloud 3.1: Foundation Infrastructure Reference Architecture Guide
Federation Enterprise Hybrid Cloud 3.1: Operations Solution Guide
Federation Enterprise Hybrid Cloud 3.1: Security Management Solution Guide Federation Enterprise Hybrid Cloud 3.1: Hadoop Applications Solution Guide