check-circle-line exclamation-circle-line close-line

vCloud Availability 3.0 | 11 APR 2019 | Build 13174385

Check for additions and updates to these release notes.

What's in the Release Notes

The release notes cover the following topics:

What's New

The VMware vCloud Availability 3.0 solution introduces a unified product architecture for disaster recovery and migration capabilities. With vCloud Availability 3.0 you can perform:

  • Migration and disaster recovery of virtual machines from an on-premise vCenter Server to a vCloud Director site
  • Migration and disaster recovery of virtual machines from a vCloud Director site to an on-premise vCenter Server
  • Migration of vApps and virtual machines between two Virtual Data Centers that belong to a single vCloud Director Organization
  • Replication and disaster recovery of vApps and virtual machines between vCloud Director sites

This release of vCloud Availability 3.0 provides the following capabilities:

  • Simplified installation
    • Appliance deployment with a web-based configuration user interface
    • A single configuration wizard for an on-premise setup and for pairing with a cloud site
  • Replication management and monitoring
    • Cross-site management for all replication types
    • Scheduling of initial sync between the source and the target replications
    • Co-existing with SDRS-enabled clustered datastore on the target site
    • Changing the replication storage policy on the target site
    • Configuring multiple replications that use seed vApps and virtual machines on the target site
  • Tenant security
    • Zero-config TLS encryption for data-in-motion with no incoming open ports required on the on-premise site
    • Control over the configuration of replications from the provider side
  • Administrative features for cloud providers
    • A single interface in the cloud for management and monitoring replications from multiple tenants.
    • Transferring replication ownership to the tenant
    • Enabling policies that apply to all replication types
    • Enabling or disabling network compression
    • Monitoring system health and the connection state of paired sites
    • Certificate management from the service management interface
  • Network configuration options on failover to the cloud
    • Connect each NIC to a separate network
    • Re-configure IP mode
    • Reset MAC address
  • Differentiation between Migrations and Protections
    • Simplified migration setup workflow
    • Separate usage reporting
  • Integrations
    • vCloud Availability Plug-in for vCloud Director
    • vCloud Availability Plug-in for vSphere Client

Upgrade

vCloud Availability 3.0 supports an in-place upgrade from vCloud Availability for Cloud-to-Cloud DR 1.5.X. For more information, see Upgrading vCloud Availability.

You can also migrate your protected workloads from vCloud Availability for vCloud Director 2.X. For more information, see Migrate from vCloud Availability for vCloud Director 2.0 to vCloud Availability DR 3.0.

Operational Limits

The following operational limits are discovered through rigorous testing of different disaster recovery scenarios. 

Scale Limits

  • 300 tenants with active protection paired to cloud
  • 20 vCloud Availability Replicator instances per cloud
  • 500 active protections per vCloud Availability Replicator instance. The maximum number of active protections depends on the RPO of the protections, the I/O of the protected virtual machines, the network connection bandwidth, and the storage speed.
  • 9500 active protections across tenants to a cloud
  • 15 TB protected virtual machine size (depends on the size of a single data store available in the cloud)

Concurrency Limits

  • 200 concurrent Protect, Test Failover, Failover, Reverse protect operations from on-premise to cloud
  • 200 concurrent Protect, Test Failover, Failover operations from cloud to cloud
  • 100 concurrent Reverse protect operations from cloud to cloud

Caveats and Limitations

Use matching vCloud Availability versions across sites. vCloud Availability 3.0.x is not compatible with vCloud Availability for Cloud-to-Cloud DR 1.5.x.

For more information about interoperability between vCloud Availability 3.0 and other VMware products, see VMware Product Interoperability Matrices.  

The vCloud Availability vSphere Client Plug-In requires a vSphere version that supports the vSphere Client. If you use an earlier version of vSphere, you cannot perform disaster recovery operations from the vSphere Client. If your vSphere does not support the vSphere Client, select the Allow Access from Cloud option in the on-premises vCloud Availability appliance, to perform all disaster recovery operations from the vCloud Availability Portal.
For more information about the configuration of the on-premises vCloud Availability appliance, see Configure the On-Premises vCloud Availability Appliance.

Supported Browsers

vCloud Availability 3.0 supports the following browsers:

  • Google Chrome 73 and later
  • Mozilla Firefox 66 and later
  • Microsoft Edge 42 and later
  • Safari 12 and later

Known Issues

  • You cannot re-protect a replicated VM from a secondary vCloud Availability instance

    If а vCloud Availability appliance is corrupted, previously configured replications cannot be re-protected from a secondary vCloud Availability appliance. When you perform a migrate or protect operation on a replicated VM, you receive the following error message and you cannot perform any further operations:

    The virtual machine has an active replication enabled.

    Workaround: To work around this issue, see http://kb.vmware.com/s/article/2106946.

  • After upgrade from vCloud Availability for Cloud-to-Cloud 1.5 to vCloud Availability 3.0, the vCloud Availability dashboard displays an error message

    After upgrade from vCloud Availability for Cloud-to-Cloud 1.5 to vCloud Availability 3.0, the vCloud Availability dashboard displays an error message:

    Certificate differs from the expected one.

    Workaround:

    • If you upgrade from vCloud Availability for Cloud-to-Cloud 1.5 with a tunneling appliance, update the Public API endpoint of the vCloud Availability 3.0 Tunnel appliance to 8048, and re-establish the trust between the two sites.
    • If you are upgrading from vCloud Availability for Cloud-to-Cloud 1.5 without tunneling appliance, configure the vCloud Availability 3.0 Tunnel appliance on both sites that you are pairing, and re-establish the trust between the two sites.
  • You cannot change the storage policy for outgoing replications using the vCloud Availability vSphere Client plug-in

    In the on-premise vCloud Availability vSphere Client plug-in, if you try to change the storage policy of an outgoing replication, a blank window displays, and you cannot proceed further.

    Workaround: To change the storage policy of an outgoing replication, use the vCloud Availability Portal.

  • If you configure a replication to the cloud, and then reverse the replication back to the on-premise site, the reversed VM loses the network configuration

    After you perform migration, test failover, and failover tasks to reverse a replication, the resulting on-premise VM does not have a network configuration.

    Workaround: Update the network configuration for the reversed VM manually.

  • A failover task fails for a VM with multiple disks that is replicated to-the-cloud

    During a failover, the VM disks are replicated to a single datastore. When you try to failover the VM to your cloud site, the task fails with the following error message:
    Invalid or inaccessible datastore.

    This issue can occur, if the summarized source VM disks size is greater than the free storage space on the destination datastore.

    Workaround: Before you run a failover task, ensure that the destination datastore has sufficient storage space to accommodate all source VM disks.

  • Your on-premise site is missing from the Replication Tasks drop-down menu in the vCloud Availability Portal

    You pair the on-premise site with the cloud site and configure replications between the two sites. When you go to the Replication Tasks list in the cloud vCloud Availability Portal, the on-premise site is not listed in the Replication Tasks drop-down menu.

    Workaround: To see the on-premise replications, use the date filter or the replication state filters in the Replication Tasks list.

  • An incoming replication using a seed VM succeeds but is not faster than a full sync

    You see that the incoming replication is successful but that the selected seed virtual machine is not removed from the vCenter Server inventory, indicating that the VM is replicated from scratch.

    Workaround: None.

  • After an initial deployment, intermittently the appliance host name is incorrect 

    When you deploy the vCloud Availability appliance by using an OVA file, intermittently the virtual machine host name is set to photon-machine.

    As a result, you are unable to configure the appliance. 

    Workaround: Before you start configuring the appliance, verify that the host name is correct.

    Log in to the vCloud Availability appliance by using your root user credentials and in the left pane click Configuration. Under Appliance settings expand the Network section and if the host name is not correct click Edit.

  • In vCenter Server 6.5 U2, applying the network settings for a replication displays an error message

    In the vSphere Web Client, when attempting to save the network settings for an existing replication, you receive a 501 Not Implemented error.

    Workaround: To configure the network settings for a replication, use the vCloud Availability Portal. 

  • After removing an on-premise site from vCloud Availability Replicator, in the vSphere Client the vCloud Availability vSphere Plug-In is still present

    After you remove your on-premise site from vCloud Availability Replicator, you see the vCloud Availability vSphere Client Plug-In present in the vSphere Client.

    Workaround: To remove the vCloud Availability vSphere Client Plug-In and clear the lookup service registration of vCloud Availability from the vCenter Server instance, in the vCloud Availability vSphere Client Plug-in, click the Close button in the upper-right corner and restart the vSphere Client.

    • If you use vCenter Server 6.5, run the following commands:

    service-control --stop vsphere-ui

    service-control --start vsphere-ui

    • If you use vCenter Server 6.7, run the following command:

    vmon-cli -r vsphere-ui

  • After a successful cloud to on-premises recovery, trying to unpair the cloud site from your on-premises site with no active replications you receive an error message

    If you have a cloud to on-premises replication, with at least one created instance, after the protected virtual machine is recovered at the on-premises site the replication is removed. When you unpair that cloud site from your on-premises site that has no active replications, you receive the following error message:  

    The manager 'XXX' can not be removed because there are replications that are managed by it.

    Workaround: None.

  • After rebooting a vCloud Availability appliance, intermittently the connectivity to a component may be lost when a link-local address is assigned to it

    vCloud Availability may intermittently assign a link-local address to the component endpoints after a reboot. For the affected vCloud Availability endpoints, there is no connectivity between components. When not using a link-local address, you can access the appliance through HTTP and SSH.

    Workaround: Disable the link-local addressing, by performing the following steps. 

    1. Log in to the vCloud Availability appliance by using a Secure Shell (SSH) client or by using the system console and authenticate with the root user credentials.

    2. Open the /etc/systemd/network/10-eth0.network file with a text editor.

    3. Under the [Network] section, add the following line and save the file. 

    LinkLocalAddressing=no

    4. Reboot the appliance.

    5. Verify that the endpoint does not use a link-local address.

    Repeat these steps for all vCloud Availability appliances.

  • vCloud Director users are not able to login to the vCloud Availability Tenant portal 

    In vCloud Director 9.1 or later, when you access the vCloud Availability portal, you can log in as the administrator, but users receive an Authentication failed error when using valid credentials.

    Workaround: Create an organization role with a unique name and grant permissions equivalent to those in the default Global Organizational Administrator role. Alternatively, you can upgrade to vCloud Availability 3.0.3.

  • vCloud Availability 3.0.x cannot operate if the service provider access to the legacy API endpoint is disabled in vCloud Director 10

    If you disable the service provider access to the /api/sessions API endpoint in vCloud Director 10, you cannot log in to vCloud Availability with service provider credentials or organization administrator vCloud Director credentials, you cannot protect or migrate vApps or virtual machines, and in vCloud Director you cannot access the vCloud Availability Plug-in.

    Workaround:

    1. Log in or SSH as root to the OS of any of the vCloud Director cells.

    2. Enable the provider access to the /api/sessions API endpoint, use the cell management tool and run the following command:

    ./cell-management-tool manage-config -n vcloud.api.legacy.nonprovideronly -v false