VMware Cloud Director Availability 4.3 Release Notes

VMware Cloud Director Availability™ 4.3 | 23 NOV 2021 | Build 18930473 (Product version: 4.3.0.3095245-9308521ec0) 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
• Upgrade
• Configuration Maximums
• Caveats and Limitations
• Supported Browsers
• Resolved Issues
• Known Issues

What's New

VMware Cloud Director Availability 4.3 now supports the following new functionality.

Advanced Retention Rules	Now providers can set more than one rule for retention of the instances. Up to five different retention rules can store instances with a granularity that is changing over time.  For example, the following four retention rules keep a total of 24 rotated instances. 10 instances, over distance 10 minutes unit - Retention period: 100 minutes. 10 instances, over distance 1 hours unit - Retention period: 10 hours. 2 instances, over distance 3 days unit - Retention period: 6 days. 2 instances, over distance 2 months unit - Retention period: 4 months.
DR and Migration Plans	Orchestrate and schedule the recovery and the migration operations across multiple virtual machines and vApps by using plans with sequences of delayable and pausable steps, containing execution settings and existing virtual machines and vApps replications references.
1-Minute Target Recovery Point Objective (RPO)	Now you can set the shortest RPO as 1 minute, by using the slider for RPO in Configure replication, in SLA profiles, or in replication policies. To meet shorter RPO, verify that you follow the recommendations for lowering the RPO violations occurrence and see https://docs.vmware.com/en/VMware-Cloud-Director-Availability/index.html. An I/O intensive workload protected with shorter RPO can cause RPO violations.
Multisite Cloud to Cloud Authentication	When multiple Availability cloud sites use a single VMware Cloud Director instance now you can use multisite authentication for performing regular replication operations when prompted to extend the session from Availability site A to Availability site B.
On-Premises Authentication by Using a Token	With vCenter Server 7.0 or later, a new authentication mechanism for the on-premises tenants allows performing disaster recovery operations in the vSphere Client Plug-In that require authentication to the cloud site. When the organization uses an external identity provider, for example, SAML, the on-premises tenants can now use that method for authentication.
Add Virtual Machines in Existing vApp Groups	Now, you can group one or several new virtual machines to an already existing vApp replication from on-premises to the cloud.
Placement Policies	Now the system providers and their tenants can select a placement organization VDC compute policy for a specific cluster or host for the recovered virtual machine.
Optimized Reverse	When performing a reverse task, now skips performing a full synchronization back to the original source workload by transferring only the deltas when there are no blocks changed in the original source and the original source disks are not modified in any way.
Backup and Restore	Now you can perform an in-place restore on existing appliances. Also, now you can perform a restore on a single cloud appliance.
Select Endpoint Address for Each Network Adapter	Now in a Cloud Replicator Appliance or in a Cloud Tunnel Appliance with multiple network interfaces, you can specify which of the network adapters to use for local site communication.
Replacing the Cloud Service Certificate Without Impacting Paired On-Premises	Now the paired on-premises appliances verify the entire SSL certificate chain of the Cloud Service. As a result, when replacing the certificate with a valid CA-signed certificate, the paired on-premises appliances do not need re-pairing with the cloud site.
Tunnel Connectivity	The following new sections now are available for verifying the statuses of the connections from the local Tunnel Service to the following destinations: Local components connectivity to all the remaining services on the cloud appliances in the local cloud site. Remote cloud sites connectivity to the remote Tunnel Service instances in all paired remote cloud sites with the local cloud site. On-Prem Incoming connectivity to all paired on-premises appliances with the local cloud site.
Improved Replicating Workloads	Now on start virtual machine replication when VMware Cloud Director Availability encounters a virtual machine that is already configured for replication, possibly by another replication solution, the stale replication is automatically unconfigured first and then it is configured for replication.

This VMware Cloud Director Availability release also includes important resolved issues and updates of third-party libraries that provide security fixes.

Upgrade

VMware Cloud Director Availability 4.3 supports an in-place upgrade directly from versions 4.1.1 and 4.2.1. To upgrade to version 4.3 from version 4.0.x, first, upgrade to version 4.2.1.

For more information, see Upgrading in the Cloud and Upgrading On-Premises.

Configuration Maximums

For the tested uptime, concurrency, and scale limits, see VMware Configuration Maximums.

Caveats and Limitations

Note: The VMware Cloud Director Availability vSphere Client Plug-In requires vSphere Client support. With earlier vSphere versions, to configure replications use the VMware Cloud Director Availability Tenant Portal. Alternatively, in VMware Cloud Director use the VMware Cloud Director Availability plug-in.

For interoperability between paired sites with mismatching VMware Cloud Director Availability versions, see Managing Connections Between Cloud Sites.

For interoperability between VMware Cloud Director Availability and other VMware products, see VMware Product Interoperability Matrices. 

Supported Browsers

VMware Cloud Director Availability 4.3 supports the following browsers:

• Google Chrome 96 and later
• Microsoft Edge Chromium 95 and later
• Mozilla Firefox 94 and later
• Safari 15 and later

Resolved Issues

• Changing the SLA profile of a replication applies retention policy but does not associate the updated SLA profile with the replication

  After changing the SLA profile of a replication, the retention policy of the new SLA profile applies but the updated SLA profile does not associate with the replication and the SLA profile column displays N/A.

• Tenants cannot start new outgoing replications by using multisite authentication to 4.2.x sites using VMware Cloud Director 10.2 or later

  Starting new outgoing protection or migration as a tenant when using multisite authentication to extend the session to destination sites using VMware Cloud Director Availability 4.2.x and VMware Cloud Director 10.2 or later, completes the task with an error message The provided credentials are not valid. This issue does not affect the service providers.

• After re-balancing replications, performing a test failover, and other management operations remain indefinitely at 1-2% progress

  To prevent the replication from being corrupted, performing management operations locks the replication before performing any changes to it.

  Re-balancing replications detaches the replication from one Replicator Service instance and attaches it to another instance. In case of an error, this replication lock does not get released.

  As a result, all subsequent management operations, for example, pause, resume, test failover, and others fail with the following error message. 

  Lock acquisition timed out for object: H4-bc6f965c-0be4-4325-9c87-a666ba418e44.

  To verify that this caused the issue, open an SSH connection to the Cloud Replication Management Appliance.

  psql -U h4postgres -d h4manager -c "select * from task where resource_id = 'H4-bc6f965c-0be4-4325-9c87-a666ba418e44' and state = 'RUNNING' and type = 'switchDestinationReplication';"

  Then, in /opt/vmware/h4/manager/log/manager.log or in one of the log archives verify that the following ERROR message is present.

  Re-attach of replication H4-bc6f965c-0be4-4325-9c87-a666ba418e44 to replicator 1ef53db8-fe7d-4cca-864e-012611273c25 failed.

• Updating the independent disk size loses its assigned storage policy and is assigned with the default storage policy of the datastore

  When updating the disk size of an independent disk, the disk gets recreated but loses its assigned storage policy. VMware Cloud Director then assigns the default storage policy of the datastore to the resized disk.

• Error when using the uplink IP address of the Edge Gateway when creating the server L2 VPN session

  Attempting to use the uplink IP address of the NSX-V Edge Gateway when creating the server L2 VPN session results in the following error message.

  IP Address is not within allowed range(s)

Known Issues

• Failing replication management operations due to VMware Cloud Director connectivity with vCenter Server

  Due to vCenter Server connectivity issues with VMware Cloud Director, any replication management operation in VMware Cloud Director Availability might fail with some of the following errors.

  • Unexpected vCloud Director error . Cannot import VM since it is already managed by vCloud Director.
  • Invalid or inaccessible datastore

  Workaround: Upgrade to vCenter Server 7.0 U3 and upgrade to VMware Cloud Director 10.3.1 or later.

• Replicating a powered off virtual machine results in red overall health

  After creating a replication of a virtual machine that is powered off, or after reversing a replication for which the recovered virtual machine is powered off, if no synchronization is triggered during the first RPO, the replication state changes to red overall health.

  Workaround:

  • Power on the virtual machine.
  • Alternatively, perform a manual synchronization. If the virtual machine is powered off after the synchronization completes, the state of the replication shows yellow overall health. 
• Upgrading Cloud Replicator Appliance instances with multiple network interfaces do not accept the replication traffic

  After upgrading to VMware Cloud Director Availability 4.3, the Cloud Replicator Appliance instances configured with multiple network interfaces do not accept the replication traffic.

  Workaround:

  • Set the endpoints in the Cloud Replicator Appliance instances to use the same network interface when communicating with the ESXi host and the Cloud Tunnel Appliance. 
  • Alternatively, upgrade to VMware Cloud Director Availability 4.3.0.1 or later.
• After upgrading from version 3.x, the appliances do not show their console

  After rebooting, VMware Cloud Director Availability appliances upgraded from version 3.x to version 4.3 do not show their blue console.

  Workaround: 

  • Open an SSH connection to each VMware Cloud Director Availability appliance, log in by using the root user credentials and then run the following commands.

  rm -rf /usr/lib/systemd/system/[e-mail].d

  systemctl enable [e-mail]

  systemctl start [e-mail]

  • Alternatively, before upgrading delete this folder. Then the upgrade can automatically enable the service so it successfully starts and shows the console of the appliance.
  • Alternatively, upgrade to VMware Cloud Director Availability 4.3.0.1 or later.
• Cannot browse and select an edge gateway for L2 stretch when any Edge uses an IPv6 address

  When one or more edge gateways are configured with an IPv6 address, in the VMware Cloud Director Availability management interface you cannot select an edge gateway for creating an L2 stretch.

  Workaround:

  • Use only IPv4 for all edge gateways.
  • Alternatively, upgrade to VMware Cloud Director Availability 4.3.1 or later.
• Active replications keep accumulating RPO violations without reporting errors and might end up as "not configured for replication"

  An active replication might show Virtual machine is not configured for replication. As a result, such replication gets in a stopped state until manually fixed.

  The replications can accumulate RPO violations due to monitoring issues between the Replicator Service and vSphere. This occurs after a false report that the replication is no longer configured at the vSphere side.

  In such cases while the event occurs, the Replicator Service log /opt/vmware/h4/replicator/log contains the following message:

  2022-02-13 20:39:17.162 ERROR - [a3306f0e-8948-4435-ab61-3f41187dc5b1] [hbr-poller3] c.v.h.r.m.source.SourceGroupMonitor : VM replication error [e-mail]: vSphere Replication operation error: Virtual machine is not configured for replication.{code}

  Workaround:

  • Reconfigure the affected replications, even without effectively changing the replication configuration. Reconfiguring the replications brings them back in a healthy state.
  • Alternatively, upgrade to VMware Cloud Director Availability 4.3.1 or later.
• Rebalancing replications between Replicator instances does not indicate any failed replications moves

  If a replication move fails, regardless between the source or the destination Replicator instance, the rebalance task would not indicate the failure.

  Workaround: Upgrade to VMware Cloud Director Availability 4.3.1 or later.

• Starting a replication might time out or take longer for larger vCenter Server inventories

  Starting a replication queries most of the source vCenter Server inventory. This operation might either time out or take a long time to complete on a vCenter Server with thousands of virtual machines.

  Workaround: Upgrade to VMware Cloud Director Availability 4.4.