Zerto Virtual Replication and Microsoft Hyper‑V Features

This section describes the interaction between Zerto Virtual Replication and commonly used Microsoft Hyper-V features such as live migration, dynamic optimization, and failover clustering.

If a host on which Hyper-V is running is restarted, a Zerto Virtual Replication Delta Sync is performed on all protected virtual machines on that host.

Microsoft dynamic disks use multiple hard disks, which can contain many dynamic volumes, to manage data. They enable you to create volumes that span multiple disks and to create fault-tolerant volumes. All volumes on dynamic disks are known as dynamic volumes. Dynamic disks offer greater flexibility for volume management because they use a database to track information about dynamic volumes on the disk. Dynamic disks and volumes rely on the Logical Disk Manager (LDM) and Virtual Disk Service (VDS) for management.

When recovering virtual machines in a VPG, VHDX disks are always recovered in the recovery site with dynamic disks. VHD disks are recovered in the recovery site by default with the same configuration as in the protected site.

Highly available virtual machines in Hyper-V reside on highly available storage such as a cluster shared volume (CSV) and can only handle disks residing on this type of storage. Therefore, when Zerto replicates a virtual machine, the type of virtual machine created on the recovery site is determined by the type of storage designated for it.

When high availability is enabled, a copy of each cached object or region is maintained on a separate cache host. The cache cluster manages these copies and supplies them to the application if the primary copies are not available.

High availability is automatically disabled by Zerto Virtual Replication while updating recovered virtual machines in the recovery site from the VRA journal. After promotion of data from the journal to the recovered virtual machine completes, high availability is automatically re-enabled.

A cluster is a group of tightly coupled hosts, called nodes, that work closely together so that in many respects they can be viewed as though they are a single computer. With a cluster, you define two or more physical machines that provide resources for the hosts that are assigned to that cluster. By using clusters, you can achieve high availability and load balancing of virtual machines.

When protecting a virtual machine on a host that is part of a cluster, Zerto recommends that you install a VRA on each host in the cluster, to ensure coverage if the virtual machine is migrated with Live Migration.

Setting up a failover cluster for Hyper-V hosts makes the virtual machines fault tolerant. The Failover Clustering feature requires you to manually specify the virtual machines that you wish to make fault tolerant. Fault tolerant virtual machines are treated as clustered roles. Virtual machines automatically restart when a node in a cluster unexpectedly failed.

Zerto Virtual Replication cannot protect virtual machines with pass-through disks. Also, a Hyper-V host with a pass-through disk is ignored by the Zerto Virtual Manager.

Performance and Resource Optimization (PRO) ensures that virtual machine hosts and virtual machines are operating in the most efficient possible manner. PRO generates recommendations for remedial actions based on alerts that the Operations Manager generates. You can configure PRO to implement the pre-defined corrective actions automatically or you can choose to perform these actions manually. During Dynamic Optimization, SCVMM migrates virtual machines within a host cluster to improve load balancing among hosts.

PRO is ignored by Zerto Virtual Replication. Dynamic Optimization is automatically disabled by Zerto Virtual Replication when updating recovered virtual machines in the recovery site from their journals. After promotion of data from the journal to the recovered virtual machine completes, Dynamic Optimization is automatically re-enabled.

Hyper-V can use multiple paths to a storage array, which is called multi-pathing. If all paths from a Hyper-V node to the storage array fail, an alternative route using the network LAN can be used to reach the storage. This is called Redirected IO.

Zerto Virtual Replication supports Multi Path IOs.

Availability sets keep virtual machines separate from each other so they do not run on the same host whenever possible. If you create an availability set in SCVMM for two different virtual machines, SCVMM attempts to keep those virtual machines on separate hosts and avoids placing them on the same host. Availability sets implement this by defining anti-affinity rules within a Hyper-V host cluster or in standalone hosts.

VRAs are defined with affinity rules to prevent their migration. Do not change these affinity rules.

During a recovery operation, when new virtual machines are created in the recovery site, migration of these virtual machines is prevented by Zerto-defined affinity rules. Once the recovery process is complete, the affinity rules are deleted and these virtual machines can be migrated.

If you use live migration to migrate a virtual machine, which is part of a VPG, from one host to another host, before moving the virtual machine make sure that:

■ The host to which you are moving the virtual machine has a Virtual Replication Appliance (VRA) installed on it, as described in the Zerto Virtual Replication Installation Guide.

■ The virtual machine is not a test virtual machine running on the recovery site during the performance of a failover test, as described in “Testing Recovery”, on page 186.

Do not migrate a Virtual Replication Appliance (VRA) from one host to another. Also, do not migrate a virtual machine on the recovery site that is being updated from the VRA journal until promotion of data to the virtual machine completes.

Microsoft quick storage migration enables you to move virtual machine storage from one location to another. The virtual disks of a running virtual machine can be migrated independently of their storage protocols or storage types. Downtime is minimal because quick storage migration takes a snapshot of the virtual machine and transfers data without requiring the virtual machine to be turned off.

Zerto Virtual Replication supports quick storage migration for protected virtual machine volumes and for journal volumes in the recovery site, but not for machine volumes in a VPG being promoted.

In SCVMM, you can configure maintenance mode for a virtual machine host anytime that you need to perform maintenance tasks on the physical host, such as applying security updates or replacing hardware on the physical host computer.

Before maintaining a host that is being used for recovery, edit VPGs that contain virtual machines that are replicating to this host and change their recovery host. Assuming that you only change the recovery host, the resulting update uses live migration and should be very quick.

Hyper-V Manager allows users designated as administrators of the management operating system to manage virtual machines on a host. Authorized users can perform management tasks on the host, such as starting and stopping VMs, importing and deploying VMs, and managing snapshots. By default, anyone who is a local administrator of the management operating system can use Hyper-V Manager on the host. In addition, a user can also use Hyper-V Manager to remotely manage Hyper-V on other servers in a domain to which the user has administrative access.

SCVMM may use run-as-account to perform operations on the Hyper-V host. Zerto Virtual Manager also uses the default run‑as‑account to perform operations on this host.

Hyper-V Replica asynchronously replicates Hyper-V virtual machines in a primary site to replica virtual machines in a secondary site.

Zerto Virtual Replication cannot protect virtual machines that are protected using Hyper-V Replica.

Zerto Virtual Manager must be restarted after SCVMM is upgraded.

SCVMM identifies and manages all its objects: VMs, hosts, networks, data stores, etc. using an internal database. Identifiers are allocated to hosts, virtual machines and other objects incrementally by SCVMM.

Each host has its own database and allocates its own IDs for the objects it holds (VM’s, Volumes, etc.). While the SCVMM database assigns IDs independently, it also receives data fields from the independent hosts’ databases.

When Zerto maps an environment within SCVMM for its configuration and operations, for example when creating a VPG, it acquires and maps the Identifiers of the available virtual machines, and refers to these Identifiers in its own environment management and replication operations.

If a host is removed from inventory and is then re-added, everything that was on it gets a new Identifier. As a result, a virtual machine, protected by Zerto that was running on that host, is allocated a new Identifier by the SCVMM. In order for Zerto to continue protecting virtual machines, Zerto Virtual Replication maintains an Identifier map containing the original Identifiers as well as unique persistent identifiers of the hosts and VMs. Zerto Virtual Manager constantly monitors the Hyper-V environment and maps the new identifiers to the same object identified by the original SCVMM identifier. This is done using the unique identifiers.

The mapping of Identifiers can take several minutes.

When system administration tasks require the importing of VPG definitions of a host that was removed and then re-added, it is advisable to wait several minutes from when the host was re-added before performing the import of the VPGs.