When using this version of Arcfra Data Protection Adapter, note the following limitations.
CPU configuration after virtual machine recovery
After an ACOS virtual machine is restored, its CPU configuration might not be consistent with that of the previous virtual machine. If the previous virtual machine is configured with a multi-socket, multi-core CPU topology, the CPU number might be restored based only on the number of CPU sockets, resulting in fewer vCPUs than the previous virtual machine. In this case, after recovery is complete, check the CPU configuration of the virtual machine and manually adjust it as needed.
Veeam 12.x file-level recovery for Windows virtual machines
In Veeam 12.x, when file-level recovery is performed for a Windows virtual machine hosted on ACOS, the restore point might fail to be identified as Windows, thus preventing assigning a Windows mount server. In this case, try mounting it by using a Linux mount server. Note that mounting with a Linux mount server is not the standard practice for Windows file-level recovery. Its availability depends on the file system type, disk layout, encryption status, and the recognition capability of the Linux mount server.
Disk bus conversion during cross-platform recovery
When an ACOS virtual machine backup is restored to a VMware environment, or when a VMware virtual machine backup is restored to an ACOS cluster, the disk bus type of the recovered virtual machine will be converted to SCSI. In this case, after recovery is complete, check disk recognition and system boot status in the guest operating system.
Repeatedly rebuilding a virtual machine with the same name within a short period of time
After you use Veeam to complete recovery by rebuilding a virtual machine, if you perform another rebuild recovery within a short period of time and choose to overwrite the target virtual machine with the same name, the restore job might fail. In this case, try again later. If the job still fails, check the virtual machine with the same name on the target platform. After confirming that the virtual machine does not need to be retained, delete it and perform the restore job again.
Access node deployment
Access nodes can be deployed only on Linux virtual machines, and cannot be deployed on Windows virtual machines.
File-level recovery
The current version supports only full-VM recovery and does not support file-level recovery for Windows virtual machines, while it supports both full-VM recovery and file-level recovery for Linux virtual machines.
Single-cluster VM group
When configuring a VM group, the virtual machines in the same VM group must reside in the same ACOS (AVE) cluster, and the access nodes configured for the VM group must reside in the same cluster as the virtual machines in that VM group. Otherwise, the backup job might fail.
Network configuration of restored virtual machine NICs
When a virtual machine is restored, the restored NICs will be preferentially associated with the networks recorded for the previous virtual machine. If the corresponding networks do not exist in the target cluster, the NICs will be associated with the default network of the target cluster. Therefore, during cross-cluster recovery, all NICs of the recovered virtual machine might be associated with the default network of the target cluster. In this case, after recovery is complete, check and adjust the NIC configuration of the virtual machine as needed.
Number of disks mounted to an access node
When a backup or restore job is performed, the system will temporarily mount virtual machine disks to the selected access node. In the current version, the total number of SCSI disk slots occupied by the same access node concurrently cannot exceed 30. This number includes disks of the access node that use the SCSI bus, and disks of backup virtual machines and restored virtual machines that are running and using the access node concurrently. Therefore, plan access nodes based on the number of virtual machine disks and job concurrency. If necessary, reduce concurrency, run jobs at different times, or add access nodes.
Configuration and storage policy after virtual machine recovery
The current version supports restoring only the basic compute specifications and disk data of a virtual machine. The disk order after recovery might differ from that of the previous virtual machine, and restored disks will be created according to the default storage policy of the target cluster. The disk provisioning method selected in Commvault Restore Wizard, such as thin provisioning, thick provisioning, or same as the previous virtual machine, is not currently used when creating restored disks.
