Zerto Quick Start
Zerto Quick Start
Replication Quick
Start AWS
Environments
Rev01 U3
Jan 2020
ZVR-QSA-7.5
Zerto Virtual Replication Quick Start AWS Environments - Rev01 U3
ZVR-QSA-7.5
2
Zerto Quick Start AWS Environments
Zerto is an IT Resilience PlatformTM to provide business continuity (BC) and disaster recovery (DR) in a
virtual environment, enabling the replication of mission-critical applications and data as quickly as possible
and with minimal data loss. When devising a recovery plan, these two objectives, minimum time to recover
and maximum data to recover, are assigned target values: the recovery time objective (RTO) and the
recovery point objective (RPO). Zerto enables a virtual-aware recovery with low values for both the RTO
and RPO. In addition, Zerto enables protecting virtual machines for extended, longer term, recovery using
Long Term Retention.
This document provides a quick guide to setting up Zerto to recover virtual machines in Amazon Web
Services (AWS). The virtual machines can be protected by Zerto in either VMware vSphere, Microsoft
Hyper-V or Microsoft Azure.
See the following sections:
Introduction on page 4
Recommended Installation Best Practices on page 7
Installation on page 8
Registering the Zerto License on page 10
Pairing Sites to Enable Replicating From One Site to Another Site on page 11
Setting Up the Protected Site on page 12
Protecting Virtual Machines on page 13
Testing Disaster Recovery on page 27
3
Zerto Quick Start AWS Environments
Introduction
Zerto helps customers accelerate IT transformation by eliminating the risk and complexity of modernization
and cloud adoption. By replacing multiple legacy solutions with a single IT Resilience PlatformTM, Zerto is
changing the way disaster recovery, retention and cloud are managed. This is done by providing
enterprise-class disaster recovery and business continuity software for virtualized infrastructure and cloud
environments.
In on-premise environments, Zerto (ZVR) is installed with virtual machines to be protected and recovered.
In public cloud environments, Zerto Cloud Appliance (ZCA) is installed in the public cloud site that is to be
used for recovery.
The installation includes the following:
• Zerto Virtual Manager (ZVM): A Windows service that manages everything required for the replication
between the protection and recovery sites, except for the actual replication of data. The ZVM interacts
with the hypervisor management user interface, such as vCenter Server or Microsoft SCVMM, to get
the inventory of VMs, disks, networks, hosts, etc. and then the Zerto User Interface manages this
protection. The ZVM also monitors changes in the hypervisor environment and responds accordingly.
For example, a VMware vMotion operation, or Microsoft Live Migration of a protected VM from one
host to another is intercepted by the ZVM and the Zerto User Interface is updated accordingly.
• For the maximum number of virtual machines, either being protected or recovered to that site, see
Zerto Scale and Benchmarking Guidelines.
• Virtual Replication Appliance* (VRA): A virtual machine installed on each hypervisor hosting virtual
machines to be protected or recovered, to manage the replication of data from protected virtual
machines to the recovery site.
• For the maximum number of volumes, either being protected or recovered to that site, see Zerto
Scale and Benchmarking Guidelines.
Note: *In vSphere installations, OVF to enable installing Virtual Replication Appliances.
• Virtual Backup Appliance (VBA): A Windows service that manages File Level Recovery operations
within Zerto Virtual Replication.
• Zerto User Interface: Recovery using Zerto is managed in a browser or, in VMware vSphere Web
Client or Client console.
When Zerto is installed to work with an on-premise hypervisor it also comprises the following component:
• Data Streaming Service (DSS): Installed on the VRA machine, and runs in the same process as the
VRA. It is responsible for all the retention data path operations.
Introduction 4
Zerto Quick Start AWS Environments
Routable Networks
The instance on which the Zerto Cloud Appliance is installed must use a subnet that is accessible from all
Zerto Virtual Managers that may be connected to this instance.
Zerto Virtual Manager does not support NAT (Network Address Translation) firewalls.
Minimum Bandwidth
• The connectivity between sites must have the bandwidth capacity to handle the data to be replicated
between the sites. The minimum dedicated bandwidth must be at least 5 Mb/sec.
Zerto Cloud Appliance requires the following ports to be open in the AWS site firewall, set in the Amazon
security group:
Introduction 5
Zerto Quick Start AWS Environments
Port Description
443 Required between the ZVM and the AWS Cloud environment.
4006 Communication between the ZVM and local site VRAs and the site VBA.
4008 Communication between VRAs to pass data from protected virtual machines to a VRA on a
recovery site.
4009 Communication between the ZVM and local site VRAs to handle checkpoints.
7073 Internal port, used only on the ZVM VM. Used for communication with the service in charge of
collecting data for the Zerto Resource Planner.
Note: Unless you select the checkbox ‘Enable Support notification and product improvement
feedback’, data is not transmitted to Zerto Analytics.
9080* Communication between the ZVM, Zerto Powershell Cmdlets, and Zerto Diagnostic tool.
9669* Communication between ZVM and ZVM GUI and ZVM REST APIs, and the ZCM.
9779 Communication between ZVM and ZSSP (Zerto Self Service Portal).
9989 Communication between ZCM, and ZCM GUI and ZCM REST APIs.
*The default port provided during the ZVR installation which can be changed during the installation.
**When the same vCenter Server is used for both the protected and recovery sites, ZVR is installed on one site only and
this port can be ignored.
Introduction 6
Zerto Quick Start AWS Environments
%ProgramData%\Zerto\Data\zvm_db.mdf
Failure to do so may lead to the Zerto Virtual Replication folder being incorrectly identified as a threat
and in some circumstances corrupt the Zerto Virtual Replication folder.
Installation
The Zerto installation deploys the Zerto Cloud Appliance (ZCA) on the recovery site. A complete
installation includes installing Zerto on the protected site.
You can install Zerto using the defaults provided by Zerto or perform a custom install, in which you define
the ports that will be used by Zerto.
Note: You cannot install Zerto on the same machine where another version of Zerto has been
installed.
Installation 8
Zerto Quick Start AWS Environments
6. After the checks complete successfully, click NEXT and continue to the end of the installation.
7. You must exclude the following folders from antivirus scanning:
%ProgramData%\Zerto\Data\zvm_db.mdf
Failure to do so may lead to the Zerto Virtual Replication folder being incorrectly identified as a
threat and in some circumstances corrupt the Zerto Virtual Replication folder.
Installation 9
Zerto Quick Start AWS Environments
where zvm_IP is the IP address of the Zerto Virtual Manager for the AWS site. Ensure that port
9669 is open and set as an inbound rule in the security group of the instance where Zerto is
installed.
2. Log in using the user name and password of the instance on AWS on which you installed the Zerto
Cloud Appliance.
When you first access the Zerto User Interface, you must register your use of Zerto by entering the ZCA
license supplied by Zerto.
Note: The license is different from the license you use for your protected site.
After entering a valid license, the DASHBOARD tab is displayed with a summary of the site.
In order to protect virtual machines to AWS, you must first pair the protected site containing the virtual
machines that you want to protect with the AWS site on which you installed the Zerto Cloud Appliance.
This is described in Pairing Sites to Enable Replicating From One Site to Another Site on page 11.
To pair sites:
1. In the Zerto User Interface, in the SITES tab click PAIR.
The Add Site dialog is displayed.
• Zerto Import - zImport: an import method that does not have the same limitations as the AWS APIs. It
creates an AWS EC2 instance per protected VM volume, called zImporter, to convert the S3 objects
and write them to a zImport local disk. When all the data has been imported and its disk has been
attached to the recovered instance, the zImport instance is terminated.
Notes:
• zImporter is based on an official AWS Linux AMI (Amazon Machine Image), into which a
script is injected to perform the import.
• To ensure that the zImport instance cannot be accessed from the outside world, a security
group is created. During a recovery operation the zImport instance is connected to this
security group. All inbound traffic is blocked and only outbound traffic to access the script
online is allowed. The security group is deleted at the end of the recovery operation.
• Zerto can set default encryption on the S3 bucket so that all objects are encrypted when
they are stored in the bucket. To enable S3 encryption please contact support.
• The default zImporter instance type is c5.4xlarge and the AWS EC2 default maximum
instance quota is 10. If during the creation of zImport instances the maximum EC2 instance
quota is reached, the creation of the next and subsequent zImport instances will be
queued, increasing the RTO. If during recovery operations, the ZVM identifies a VPG with
the potential to exceed the EC2 instance quota, the user will receive an alert with advice to
contact AWS support to increase the service limits in order to improve RTO.
• Each zImporter VM is responsible for the import process of a single volume. Therefore,
it is recommended to contact AWS and increase the maximum instance quota of the
c5.4xlarge instance type to the maximum number of volumes you are planning to
failover to AWS at once.
• GPT formatted disks are supported for data volumes only, when using either of the zImport
methods.
• When using either of the zImport methods, each volume is created with EBS disk of type
io1 with maximum 1000 EBS Provision IOPS allocated. EBS disk type can be changed
post recovery without downtime, see the relevant for more information see the relevant
AWS documentation. The minimum disk size for io1 is 4GB.
• The default Max EBS Provision IOPS quota in a region across all io1 disks is 40000 EBS
Provision IOPS, meaning that with 1000 EBS Provision IOPS per volume, the maximum
possible number of volumes is 40. If the Max EBS Provision IOPS quota is reached, the
failover process will switch to using slower gp2 disks. An event will notify the user of this,
and recommend that the user contact AWS support to increase the Max EBS Provision
IOPS quota.
• Depending on the desired RTO during recovery operations, or when testing failover, the user can
select an import method per VPG or per virtual machine from the following options:
• Zerto Import for Data Volumes on page 14
• Zerto Import for All Volumes on page 15
• AWS Import on page 16
• Each machine that you intend to protect must have at least 250MB free space. This is because
AWS adds files to the recovered machines during failover, move, test failover, and clone operations.
• Protected boot volumes are recovered in EC2 as EBS disks with magnetic disk type. Virtual
machines with disks that are less than 1GB are recovered with disks of 1GB. Temporary disks may be
created based on the selected instance size.
• Temporary disks may be created based on the selected instance size.
• The maximum protected data volume size is 16TB, while the boot volume can be up to 1TB.
• The AWS ImportInstance API only supports single volume VMs. The boot volume of the protected
virtual machine should not be attached to any other volume to successfully boot. For more information,
see http://docs.aws.amazon.com/AWSEC2/latest/APIReference/API_ImportInstance.html
Note: Some VMs use the MBR partitioning scheme, which only supports up to 2047 GiB boot
volumes. If your instance does not boot with a boot volume that is 2TB or larger, the VM
you are using may be limited to a 2047 GiB boot volume size. Non-boot volumes do not
have this limitation. See AWS Documentation for more information:
http://docs.aws.amazon.com/AWSEC2/latest/UserGuide/EBSVolumeTypes.html
When recovering to AWS instance types listed below using Zerto Import for All Volumes import method,
Windows 2012, Windows 2012R2, Windows 2016 or Windows 2019 are supported.
• C3 • I2
• C4 • R3
• D2 • M4 (excluding M4.16xlarge)
• For Windows 2012R2, Windows 2016 and Windows 2019, the following drivers are downloaded on
the protected virtual machine. ZertoTools installs these drivers on the recovered virtual machine:
• Windows ENA (Elastic Network Adapter) Drivers
• NVMe driver
• For Windows 2008R2 and Windows 2012, Windows ENA (Elastic Network Adapter) drivers and
NVMe driver are downloaded on the protected virtual machine. ZertoTools installs these drivers on the
protected virtual machine and executes the below command:
This command removes computer-specific information for the drivers. When recovering to C5/M5, the
instance will boot on AWS and install all the drivers again.
Note: If these drivers are installed on a VM running Windows 2012, Windows 2012R2,
Windows 2016 or Windows 2019, the other AWS import methods will fail. To
overcome this, you must uninstall the drivers before using the other AWS import
methods.
Note: C5/M5 instance types are supported with the Zerto Import for All Volumes import
method only.
! Important:
When using this import method for Windows machines, ZertoTools for Windows needs to be
run on the protected Windows virtual machine in VMware before VPG creation. For more
information, see ZertoTools for Windows on page 17.
AWS Import
This method uses a combination of the AWS import-instance and import-volume APIs for the boot and
data volumes respectively. This was the only method supported until version 5.5. To use this method when
creating or editing a VPG, an Access Key ID and a Secret Access Key is required. Both fields can be set
in the Site Settings window, see Site Settings on page 1.
• Each machine that you intend to protect must have at least 250MB free space. This is because
AWS adds files to the recovered machines during failover, move, test failover, and clone operations.
• Protected boot volumes are recovered in EC2 as EBS disks with General Purpose SSD (gp2). Virtual
machines with disks that are less than 1GB are recovered with disks of 1GB. Additional volumes might
be created in the recovered instance, dependent on the instance type used for the recovery. These
volumes can be ignored.
• Protected volumes are recovered in EC2 as EBS disks with General Purpose SSD (gp2). Virtual
machines with disks that are less than 1GB are recovered with disks of 1GB. Additional volumes might
be created in the recovered instance, dependent on the instance type used for the recovery. These
volumes can be ignored. Temporary disks may be created based on the selected instance size.
• The maximum protected data volume and boot disk size is 1TB.
• The AWS ImportInstance API only supports single volume VMs. The boot volume of the protected
virtual machine should not be attached to any other volume to successfully boot. For more information,
see http://docs.aws.amazon.com/AWSEC2/latest/APIReference/API_ImportInstance.html
ZertoTools Requirements
• Windows 2019* (*Windows 2019 is supported only if using Zerto Import for All Volumes)
• Run ZertoTools from C: Windows OS drive only.
• .Net Framework 4.5 and up must be installed.
ZertoTools Limitations
• When using the Zerto Import for All Volumes import method, ZertoTools should not be installed on
machines with Windows that are Domain Controllers.
When failing over machines with Windows 2012, 2012R2 and 2016 that are Domain Controllers, the
Windows Citrix PV drivers need to be downloaded manually on the protected machines.
To download and install Windows PV drivers:
a. Go to https://www.xenproject.org/downloads/windows-pv-drivers/winpv-drivers-81/winpv-drivers-
820.html
b. Follow the instructions for downloading and installing all Windows PV drivers.
• To failover Windows 2008R2 with Domain Controller, contact Zerto Support.
• Failback for Windows machines with Domain Controller is not supported.
• When recovering to the AWS instance types listed below using Zerto Import for All Volumes import
method, Windows 2012, Windows 2012R2, Windows 2016 and Windows 2019 are supported.
• C3 • I2
• C4 • R3
• D2 • M4 (excluding M4.16xlarge)
• The following drivers are downloaded on the protected virtual machine. When recovering to these
AWS instance types, ZertoTools installs these drivers on the recovered virtual machine:
• AWS PV Drivers
• Windows ENA (Elastic Network Adapter) Driver
• When recovering to C5/M5 instances using Zerto Import for All Volumes import method, Windows
2008R2, Windows 2012, Windows 2012R2, Windows 2016 and Windows 2019 are supported.
• For Windows 2012R2, Windows 2016 and Windows 2019, the following drivers are downloaded
on the protected virtual machine. ZertoTools installs these drivers on the recovered virtual
machine:
• Windows ENA (Elastic Network Adapter) Drivers
• NVMe driver
For Windows 2008R2 and Windows 2012, Windows ENA (Elastic Network Adapter) drivers and
NVMe driver are downloaded on the protected virtual machine. ZertoTools installs these drivers on the
protected virtual machine and executes the below command:
This command removes computer-specific information for the drivers. When recovering to C5/M5, the
instance will boot on AWS and install all the drivers again.
Note: If these drivers are installed on a VM running Windows 2012, Windows 2012R2,
Windows 2016 and Windows 2019, the other AWS import methods will fail. To
overcome this, you must uninstall the drivers before using the other AWS import
methods.
Note: C5/M5 instance types are supported with the Zerto Import for All Volumes import
method only.
AWS PV Driver
When an instance is created in AWS or when performing Failover using the Zerto Import for Data
Volumes and AWS Import methods, by default, Amazon will install the latest version of the AWS PV
driver.
When using the Zerto Import for All Volumes import method, only the AWS PV driver version 7.4.6 is
supported when failing back from AWS to vSphere. ZertoTools installs AWS PV driver version 7.4.6
by default. You can install ZertoTools with the latest AWS PV driver manually but you will be required
to downgrade the AWS PV driver before performing Failback to vSphere.
Note: AWS PV driver is not installed when recovering to C5/M5 instance types.
Downgrade Script on page 21. (For instances using Zerto Import for Data Volumes and
AWS Import methods, the latest AWS PV driver will be installed by Amazon).
• -q (quiet mode): Install ZertoTools without any prompt. The default for prompt will be Yes.
This argument is relevant only when recovering to C5/M5 instance types with Windows
2008R2 and Windows 2012.
• -v: Verify that all required components are downloaded and your machine is ready for
failover.
• -u: Uninstall ZertoTools.
! Important:
4. Wait a few minutes and verify you get the following message: Process successfully finished.
! Important:
5. The script downloads the drivers and installs them upon failover on the AWS machine. If the
download of these drivers fails, manually download them from the following links:
• AWS PV driver version 7.4.6: https://s3.amazonaws.com/ec2-windows-drivers-
downloads/AWSPV/7.4.6/AWSPVDriver.zip
• AWS PV driver version <Latest>: https://s3.amazonaws.com/ec2-windows-drivers-
downloads/AWSPV/8.3.1/AWSPVDriver.zip
• Download ENA (Enhanced Network Adapter): https://s3.amazonaws.com/ec2-windows-
drivers-downloads/ENA/Latest/AwsEnaNetworkDriver.zip
• Download the NVMe driver: https://s3.amazonaws.com/ec2-windows-drivers-
downloads/NVMe/Latest/AWSNVMe.zip
Note: If you need to download the drivers manually, the zip file name of each driver
should not be changed.
ZertoTools will also backup the VMtools to ensure re-IP works upon Failback to the protected VMware
site.
When running ZertoTools, note the following:
• A folder named ZertoTools is created on C:/ProgramData folder.
This folder must not be deleted.
• Upon failover to AWS, the AWS PV driver update may force reboot of the recovered instances in
AWS.
Downgrade Script
If you decide to install the latest AWS PV driver, you will need to downgrade it before failing back to
vSphere.
! Important:
Your AWS machine may need to reboot a few times upon execution of the batch files.
5. After a few minutes, connect to the AWS instance and verify that the AWS PV driver is version
7.4.6.
If you installed ZertoTools after VPG creation, wait for the next checkpoint to be created before
performing failover.
2. Specify the name of the VPG and the priority of the VPG.
VPG Name: The VPG name must be unique.
Priority: Determine the priority for transferring data from the protected site to the recovery site
when there is limited bandwidth and more than one VPG is defined on the protected site. When
there are updates to virtual machines protected in VPGs with different priorities, first the updates
from the VPG with the highest priority are passed over the WAN. Medium priority VPGs will only be
able to use whatever bandwidth is left after the high priority VPGs have used it. This is also true
between medium and low priorities.
3. Click NEXT.
The VMs step is displayed.
4. Select the VMs that will be part of this VPG and click the right-pointing arrow to include these VMs
in the VPG.
• Zerto uses the SCSI protocol. Only virtual machines with disks that support this protocol can be
specified.
• When using the Search field, you can use the wildcards; * or ?
Virtual machines that are not yet protected are displayed in the list. A VPG can include virtual
machines that are not yet protected and virtual machines that are already protected.
5. You can view protected virtual machines in the Advanced (One-to-Many) section, by clicking
Select VMs.
The Select VMs dialog is displayed.
Note: Virtual machines can be protected in a maximum of three VPGs. These VPGs
cannot be recovered to the same site. Virtual machines protected in the maximum
number of VPGs are not displayed in the Select VMs dialog.
In on-premise environments, protecting virtual machines in several VPGs is enabled only if both
the protected site and the recovery site, as well as the VRAs installed on these sites, are of version
5.0 and higher.
6. To define the boot order of the virtual machines in the VPG, click DEFINE BOOT ORDER,
otherwise go to the next step.
When virtual machines in a VPG are started in the recovery site, by default these machines are not
started up in a particular order. If you want specific virtual machines to start before other machines,
you can specify a boot order. The virtual machines are defined in groups and the boot order applies
to the groups and not to individual virtual machines in the groups. You can specify a delay between
groups during startup.
Note: Up to 20 virtual machines may boot on a host simultaneously. Following the boot, a
15 second (default) delay occurs until the next boot batch.
Initially, virtual machines in the VPG are displayed together under the Default group. If you want
specific machines to start before other virtual machines, define new groups with one or more virtual
machines in each group.
Default group are started together and after 10 seconds the virtual machines in the Server
group are started. After 100 seconds the virtual machines in the Client group are started.
f. Click OK.
7. Click NEXT.
The REPLICATION step is displayed.
Note: If the protected site is paired with only one recovery site, the recovery step is
displayed with the Recovery Site field automatically filled in and defaults set for
the fields that are relevant for AWS.
8. Specify the recovery site and the values to use when replicating to this site.
Recovery Site: The site to which you want to recover the virtual machines.
As soon as you specify that the recovery site is on AWS, the display changes to show only fields
that are relevant for AWS.
9. The following settings can be changed later by editing the VPG definition. For your first VPG, leave
the default values and click NEXT.
After clicking NEXT, the RECOVERY step is displayed. Recovery details include the networks to
use for failover, move, and testing failover, and whether scripts should run as part of the recovery
process.
Note: Steps that do not require input are marked with a check mark. You can jump
directly to a step that has been marked with a check mark to edit the values for that
step. Every step must be marked with a check mark before you can click DONE to
create the VPG.
Note: For synchronization to work, the protected virtual machines must be powered on.
Once synchronized, the VRA on the recovery site includes a complete copy of every virtual machine in the
VPG. After synchronization, the virtual machines in the VPG are fully protected, meeting their SLA, and
the delta changes to these virtual machines are sent to the recovery site.
To verify that the disaster recovery that you have planned is the one that will be implemented, Zerto
recommends testing the recovery of the VPGs defined in the protected site to the recovery site.
• The updated checkpoint can be used to identify the point-in-time to restore the virtual machines in
the VPG during a failover.
Testing that recovery is accomplished successfully should be done periodically so that you can verify that a
failover will work. Zerto also recommends testing all the VPGs being recovered to the same cluster
together.
When configuring a VPG, specify the period between tests for that VPG in the Test Reminder field in
the REPLICATION step of the Create VPG wizard.
Note: You can initiate the failover test from either the protected site or recovery site.
To test failover:
Note: Selecting specific VMs in a VPG to failover is not supported when replicating from a
vCD site.
At the bottom, the selection details show the amount of data and the total number of virtual
machines selected.
The Direction arrow shows the direction of the process: from the protected site to the peer,
recovery, site.
3. Click NEXT.
The PARAMETERS step is displayed.
You can select the checkpoint to use for the recovery and see if a boot order and scripts are
defined for the VPG.
By default, the last checkpoint added to the journal is displayed. The checkpoints determine the
RPO and ensure crash consistency and write-fidelity when the virtual machines in a VPG are
recovered. These checkpoints are written every few seconds and you can recover to any of the
available checkpoints.
4. Click NEXT.
The FAILOVER TEST step is displayed. The topology shows the number of VPGs and virtual
machines being tested to failover to each recovery site.
If you did not define a private IP for a virtual machine in the VPG definition, during recovery AWS sets the
private IP from the defined subnet range.
The virtual machines in the virtual protection group are created in AWS. In the AWS console, the new
virtual machines appear with their original names and the suffix testing recovery.
While a test is running:
• The virtual machines in the VPGs continue to be protected.
• You can add checkpoints to the VPGs, and if necessary fail over the VPGs.
• You cannot move VPGs being tested.
• You cannot initiate a failover while a test is being initialized or closed.
Monitor the status of a failover test by doing the following:
• In the Zerto User Interface, click the VPGs tab. The Operation field in the GENERAL view displays
Testing Failover when a failover test is being performed.
• In the Zerto User Interface, click the VPGs tab, and then click the name of a VPG you are testing. A
dynamic tab is created displaying the specific VPG details including the status of the failover test.
You can also stop the test via the TASKS popup dialog in the status bar, or by selecting
MONITORING > TASKS.
The Stop Test dialog is displayed.
This checkpoint can be used to identify the point-in-time to use to restore the virtual machines in the VPG
during a failover.