VMware

VMware Disaster Recovery Service Release Notes

Published: 16 March 2015
Last updated: 15 January 2016

What's in the Release Notes

The release notes cover the following topics:

About the Disaster Recovery Service

Disaster Recovery is a Recovery-as-a-Service (RaaS) offering to protect virtual workloads managed by VMware vSphere that are either deployed in a private cloud or data center by replicating those workloads into a cloud.

The Disaster Recovery Service consists of the following VMware products and services:

Note: These release notes provide information for vSphere Replication 6.0 Disaster Recovery to Cloud and the vCloud Air - Disaster Recovery service released on 16 March 2015.

What's New

The following features are new for this release of vCloud Air - Disaster Recovery:

  • Increased Scalability: You can now configure for replication and manage a maximum of 1000 simultaneous incoming (from vCenter to vCloud Air - Disaster Recovery) virtual machine replications per VDC, up from the previous limit of 500 replications.

  • Support for vRealize Orchestrator Plug-In: You can use the vRealize Orchestrator plug-in for vSphere Replication, which allows you to simplify the management of vSphere Replication infrastructure by leveraging the robust workflow automation platform of vRealize Orchestrator. For features and known issues, see the vRealize Orchestrator Plug-In for vSphere Replication 6.0 Release Notes.

  • Multiple Point in Time Recovery: With multiple point in time (MPIT) recovery, you can pick the specific point in time that you want to have your data recovered successfully in the event of data corruption. You can recover back up to 24 previous replication points in time based on your recovery point objective (RPO) settings. For example, if you have set a 15 minutes RPO, the data changes will be replicated every 15 minutes. When you want to failover from your on-premises data center to the vCloud Air - Disaster Recovery data center, you can pick a specific time from which to recover the data. Consider that you start the replication process at 10 a.m. and your on-premises data center has an issue and you need to failover at 11 a.m. However, at that point in time the data may not be suitable, in which case you can recover from the replication that occurred at 10:15 a.m.

  • Native Failback Using vSphere Replication: You can replicate to and from the cloud by using vSphere Replication 6.0. In the event of a disaster you can failover all your virtual machines to the cloud. Native failback allows you to replicate your virtual machines back (reverse replication) to your on-premises data center after you have performed a failover.

Note: If you are using vCloud Air - Disaster Recovery released on 14 April 2014, depending on the vCenter Server and vSphere Replication versions you may or may not have access to the new features. In which case, you might need to use vCloud Connector to copy virtual machines from the vCloud Air - Disaster Recovery data center to your on-premises data center.

vCloud Air Supported Browsers

Connectivity Issue

The potential connectivity issue for this release of vCloud Air - Disaster Recovery that was documented in the VMware Knowledge Base (KB 2110988) has been resolved with the release of vSphere Replication 5.6.0.x and 5.8.0.x patches. For more information, see the corresponding KB articles.

Note: To download vSphere Replication 5.6.0.x or 5.8.0.x, go to Download VMware vCloud Air and click the Go to Downloads link corresponding to vSphere Replication 6.0. The Download VMware vSphere Replication 6.0 page is displayed. Choose the vSphere Replication version you want to download from the Select Version dropdown and then click Download Now.

Disaster Recovery Service Known Issues

The following known issues have been discovered through rigorous testing and will help you understand some behavior you might encounter in this release.

vSphere Replication Disaster Recovery to Cloud

  • Replication status changes to Error if you add or remove disks on the source virtual machine during test recovery

    Adding or removing disks to a replication source virtual machine during test recovery is not supported.

    Workaround: None.

  • vSphere Replication target sites are not visible in the vSphere Web Client after upgrading the vSphere Replication appliance

    After upgrading vSphere Replication to version 6.0, replication target sites are no longer visible in the vSphere Web Client user interface.

    Workaround: Clear the browser cache and log in to the vSphere Web Client interface again.

  • Unable to complete any operations on a replication to the cloud

    If the placeholder vApp of a configured replication to the cloud is removed, no operations can be completed on the replication and it will eventually go into Error state.

    Workaround: Stop the replication that indicates Error state and configure a new replication for the same source virtual machine.

  • Replications to cloud indicate Error state and you cannot reconfigure them

    vSphere Replication to Cloud is unable to reconfigure replications that failed during initial configuration. This problem occurs because replication details are missing from the cloud provider site.

    Workaround: Stop the replication by using the vSphere Replication UI in the vSphere Web Client, and try to configure a new replication to cloud.

  • Replication configuration task fails and you cannot create another replication for the same source virtual machine

    If you complete the Configure Replication wizard while the vSphere Replication service is not running, the following error message is displayed:
    VRM Server runtime error. Please check the documentation for any troubleshooting information.
    The detailed exception is: 'External service 'vSphereReplication' failed to respond in the specified timeout (10 SECONDS)'.

    No replication is created on the Outgoing Replications page for that virtual machine, but if you attempt to configure another replication for this virtual machine, the runtime exception details read Unable to create replication group on secondary site.
    If the vSphere Replication service stops during initial configuration, a placeholder vApp might be created on the target site before the initial configuration task fails. The placeholder vApp on the target site does not allow creating a new replication for the same source virtual machine.

    Workaround:

    1. Delete the placeholder vApp from the target site by using the vCloud Air web UI.
    2. Stop the replication on the source site.
    3. Try configuring another replication for the virtual machine

  • Replication configuration fails with 'Placeholder vApp creation failed' exception

    You cannot enable replications for virtual machines that have a CD/DVD device that is marked as Datastore ISO File and is in Connected or Not connected state.

    Workaround: Edit the virtual machine configuration and mark the CD/DVD device as Client Device. Do not select the Connected check box.

  • Concurrent export of more than 10 virtual machines through the vCloud Connector method causes the process bar to remain stuck at 0%

    If you trigger an export of more than 10 virtual machines (or 1 large virtual machine) concurrently from the vCloud Connector UI to an Offline Data Transfer node, the process bar remains at 0% until the process is complete.

    Workaround: None. Avoid exporting more than 10 virtual machines concurrently.

  • Replications from cloud turn into Error state

    If you use the vCloud Air Web UI to add a new disk to a virtual machine that serves as a replication source, vSphere Replication at your on-premises site automatically pauses the incoming replication for that machine, and moves the replication group into Error state.

    Workaround: Stop the replication from cloud that indicates Error state, and configure a new replication.

  • Hardware changes on the replication source virtual machine might not be automatically copied to the placeholder vApp in the cloud

    Changes to the protected virtual machine on the source site, such as changes to memory, CPU, networks, and so on, might not be replicated to the placeholder vApp in your cloud organization if you apply them while vSphere Replication is running a workflow, for example, a test recovery.

    Workaround: Edit the hardware of the replication source virtual machine again to trigger a full synchronization.

    1. In the vSphere Web Client inventory tree, right-click the source virtual machine.
    2. From the drop-down menu, select Edit Settings, and apply a change to the virtual hardware.
      Note: Opening and closing the Edit Setting dialog box is not enough. You must apply some change to the hardware.
    3. Click OK.

  • Disks are not automatically consolidated during recovery at the cloud site

    If you configure a replication to cloud that has the MPIT functionality enabled, and you recover the replicated virtual machine at the cloud site, its retained instances are not consolidated during the recovery. By design, replication instances are not consolidated to speed up the recovery process.
    The unconsolidated disks in the recovered virtual machine might cause performance problems as follows.

    • The recovered virtual machine runs slower than expected.
    • The recovered virtual machine requires more storage resources.

    Workaround: Use the vCloud Air interface to manually consolidate the recovered virtual machine disks.

  • Outgoing replications to cloud remain in Not Active state

    By default, when you power on the vSphere Replication appliance, a vSphere Installation Bundle (VIB) is installed on all supported ESXi hosts in the vCenter Server inventory where the appliance is deployed. The VIB creates a firewall rule, Replication-to-Cloud Traffic, that opens TCP ports 10000 to 10010 for outgoing traffic. However, the automatic installation of the VIB file might fail due to network issues in your environment. When the firewall rule is missing on the source ESXi hosts, outgoing replications to cloud remain in Not Active state.

    Workaround: Install the vSphere Replication VIB file on each ESXi instance that hosts a cloud replication source virtual machine.

    1. Temporarily disable the firewall on the ESXi host.
    2. Establish an SSH connection to the ESXi Server.
    3. Run the following command:
      $ esxcli software vib install -v https://VR_APPLIANCE_IP:8043/vib/vr2c-firewall.vib
    4. Enable the firewall on the ESXi host.
    5. Note: In case of a networking issue, the ESXi host may not be able to connect to vSphere Replication on port 8043.

  • Incoming replications from cloud remain in Not Active state

    If, in your vCloud Air inventory, you have a virtual machine that has disks of size not equal to whole Megabytes, for example 104.56 MB as opposed to 104 MB, and, at your on-premises site, you configure a replication from cloud or a reverse replication for that virtual machine, the replication appears in the Incoming Replications list, but remains in Not Active state.

    The problem is observed because the disk size that the cloud system reports is rounded down.

    Workaround:

    1. At your on-premises site, stop the replication from cloud that appears as Not Active.
    2. In vCloud Air, locate the virtual machine and open its properties.
      The disk size is rounded down. For example, the disk size for a 104.56 MB disk appears as 104 MB.
    3. Increase the disk size by one MB.
      For example, for a 104.56 MB disk change the size from 104 MB to 105 MB.

  • The network configuration of replication source virtual machine is mirrored when a replication is configured, but only partially updated after that

    By default, when you configure a virtual machine for replication to cloud, its NICs and MAC addresses are copied automatically to the target site as part of the provisioning of the placeholder virtual machine. If, after the provisioning of the placeholder virtual machine, you update the NICs on the replication source virtual machine, the NICs on the target site are updated. However, if you update the MAC address of a NIC on the source virtual machine, the address of the placeholder virtual machine on the target site is not updated. The MAC addresses of newly-added NICs are not copied to the target virtual machine either.

    Workaround: Use vCloud Air REST APIs.

  • Planned migration fails with error Synchronization monitoring has stopped

    If the synchronization phase of a planned migration task takes more than 3 minutes, the task fails with error Synchronization monitoring has stopped. The synchronization can take more than 3 minutes to complete due to slow VSS quiescing, due to large accumulated delta since the last synchronization, or if large delta has been accumulated during the guest shutdown.

    Workaround:

    • Manually power off or shut down the guest OS of the source virtual machine.
      1. After the failed planned migration task, wait for the sync status to change to OK.
      2. Manually power off or shutdown the guest OS of the source virtual machine.
      3. Try running the planned migration task again.
    • Increase the toleration period for replication synchronizations on the vSphere Replication Management Server (VRMS).
      1. Log in to the VRMS appliance as the root user and navigate to /opt/vmware/hms/conf/.
      2. Open the hms-configuration.xml file for editing and set the value of the hms-sync-replication-error-toleration-period property to a value that exceeds the actual duration of the synchronization phase.
        The value is set in milliseconds. For example, the value for 4 minutes is 4*60*1000=240000 milliseconds.
      3. Restart the VRMS service by running the command - /etc/init.d/hms restart.
      4. Try running the planned migration task again.

  • Planned migration or synchronization fails with error A replication error occurred at the vSphere Replication Server.

    If, during planned migration, the infrastructure (hosts, network, or storage) is under heavy load, running a planned migration might fail with the following error.
    A replication error occurred at the vSphere Replication Server for replication <group_name>. Details: 'Error for (datastoreUUID: "..."), (diskId: "..."), (hostId: "..."), (pathname: "..."), (flags: retriable): Class: NFC Code: 10; NFC error: The operation completed successfully; Set error flag: retriable; ...'
    Usually, this error is transient and the operation succeeds if you retry running it.

    Workaround: If the error occurs frequently in your environment, you can increase the toleration period for replication synchronizations on the vSphere Replication Management Server (VRMS).

    1. Log in to the VRMS appliance as the root user and navigate to /opt/vmware/hms/conf/.
    2. Open the hms-configuration.xml file for editing and set the value of the hms-sync-replication-error-toleration-period property to 300000.
    3. Try running the planned migration task again.

  • All operations on a seed vApp in vCloud Air are disabled

    If you configure a replication to cloud and select a vApp from the vCloud Air inventory to be used as a replication seed, all operations on the seed vApp are disabled.

    Workaround: None. Replication seeds cannot operate as virtual machines. A seed vApp can be used for only one replication.

  • You cannot disable the automatic mirroring of source virtual machine NICs and MAC addresses to the target site

    When you configure a virtual machine for replication to cloud, its NICs and MAC addresses are copied automatically as part of provisioning the placeholder virtual machine on the cloud site.
    If the test network is not isolated from the production network and the networks have common routing, a test recovery of the replicated virtual machine might result in duplicate MAC addresses in your virtual data center.

    Workaround: Use the vCloud Director UI to reset the MAC address of the replicated virtual machine. See Resetting the MAC Address of a Virtual Machine that is Replicated to Cloud (KB 2086292).

  • Exception Invalid datastore format '<Datastore Name>' appears when vSphere Replication runs a sync task

    If you rename a datastore on which replication source virtual machines are running, replication sync operations for these virtual machines fail with error
    VRM Server runtime error. Please check the documentation for any troubleshooting information. The detailed exception is: 'Invalid datastore format '<Datastore Name>'

    Workaround: Pause and resume the replications that use source virtual machines from the renamed datastore.

  • A timeout error appears while configuring a replication to the cloud to use a vApp as a seed

    Depending on the environment performance and connectivity, you might experience a timeout issue when you try to configure a vApp as a seed. The following error message appears on the Replication seed page of the Configure Replication wizard:
    The data service timed out because a back-end task took more than 120 seconds. See the vsphere-client-virgo.log for more details.

    Workaround: None. If the performance or connectivity issues in your environment persist, configure replications without using seeds.

vCloud Air - Disaster Recovery

  • Replication configuration of failedback virtual machines fails with error

    Configuration of replication for virtual machines which are recovered and failedback to the on-premises environment using the vCloud Connector copy fails with "VRM Server runtime error. Please check the documentation for any troubleshooting information. The detailed exception is: 'Placeholder vApp creation failed'." error.

    Workaround: Before configuring replication on failedback virtual machines, edit the virtual machine configuration through the vSphere web client as follows:

    1. Select the corresponding virtual machine.
    2. Right-click the virtual machine and select the "Edit Settings" option.
    3. Select the "vApp Options" tab.
    4. Deselect the "Enable vApp options" checkbox and confirm the step.

  • Ability to edit and delete a virtual machine replicated to the cloud while its status is "Placeholder"

    From the Virtual Machine Details section of the vCloud Air web UI, you can edit and delete a virtual machine that was replicated the cloud while its status is still "Placeholder".

    Workaround: None.
    Note: Do not perform any action on the virtual machine from the Virtual Machine Details section of the vCloud Air web UI. Instead first refer to its status under the Replication tab and Virtual Machine tab of the vCloud Air Dashboard > Virtual Data Center Details.

  • Replication status updates are not displayed

    Although the replication sync process keeps running in the background, the replication status window in the on-premises vCenter Server no longer reports updates about the replication status.

    Workaround: None. There are no issues with the replication (only the status window does not get updated).

  • The vApp in the cloud is not powered off after a recovery to the on-premises data center

    When you recover a replication from the cloud at the tenant site and, in the Recovery wizard, you choose to recover the virtual machine by using the Use latest available data option, vSphere Replication does not power off the source vApp in the cloud by default.
    This is because the Use latest available data option assumes that there is no connection to the replication source site.

    Workaround: You can connect to the cloud to manually power off the source vApp.

  • Virtual machine state does not get updated in the cloud virtual machine list on state change

    Even after the virtual machine state changes, a spinning wheel is displayed and the final state is not reflected until the virtual machine list is refreshed.

    Workaround: None.

  • The "Running Since" column is blank for powered-on virtual machines

    Even though virtual machines are powered-on and running, the "Running Since" column does not display any values.

    Workaround: None. The "Running Since" column is only relevant for virtual machines that have been successfully tested or recovered. The column name has been changed to "Recovered On".