Guide de déduplication NetBackup™

In a situation where MSDP cloud backed up the deduplicated data to cloud, the NetBackup catalog was available on the on-premises NetBackup server.

Image sharing in the cloud uploads the NetBackup catalog along with the backup images and lets you restore data from the cloud without the on-premises NetBackup server.

You can launch NetBackup on demand, which is called the cloud recovery server, and recover the backup images from the cloud.

Image sharing discovers the backup images that are stored in cloud storage through the REST APIs, command line, or web UI, recovers the NetBackup catalog, and restores the images.

You can use command line options or NetBackup web UI that have the function as REST APIs.

For the imported Standard, MS Windows, and Universal share backup images, you can instantly access them with NetBackup Instant Access APIs as the exported share is in a read-only mode. For the imported VMware images, you can instantly scan them with the VMware Malware Scan APIs as the exported share is in a read-only mode.

Se reporter à À propos de l'accès instantané pour le stockage d'objets .

For FortKnox, in the VM conversion procedure, a temporary bucket or blob container is created automatically. Region and the security options of the bucket are the same as the FortKnox account on the image sharing server.

The temporary bucket or blob container name format is vrtsonvert-<timestamp>/VRTSConvert-<timestamp>.

For FortKnox Amazon, MSDP-C credentials with AWS account with IAM and EC2 related permissions must be created before the VM conversion. For FortKnox Azure, MSDP-C credentials with Azure general-purpose storage accounts must be created before the VM conversion.

For FortKnox, only the image import function uses FortKnox credentials. Before running the VM image conversion, ensure that you create MSDP-C credentials with the access credentials of your Azure/AWS account. The VM image conversion needs the access credentials of your Azure/AWS account because the Recovery Vault storage does not support creating AMI or VHD. Additionally, Azure Service Principal and AWS IAM Roles Anywhere MSDP-C credentials are not supported for VM conversion with FortKnox.

Before you install NetBackup, create an instance based on SUSE Linux Enterprise or RHEL 7.3 or later. You can also set up a computer based on SUSE Linux Enterprise or RHEL 7.3 or later. The recommendation is that the instance has more than 64 GB of memory, 8 CPUs.

The HTTPS port 443 is enabled.

Change the host name to the server's FQDN.

In Azure virtual machine, you must change the internal host name, which is created automatically for you and you cannot get an internal host name from an IP address.

Add the following items in the /etc/hosts file:

"External IP" "Server's FQDN"

"Internal IP" "Server's FQDN"

For a computer, add the following items in the /etc/hosts file:

"IP address" "Server's FQDN"

(Optional) For an instance, change the search domain order in the /etc/resolv.conf file to search external domains before internal domains.

For a new image sharing server, ensure that NGINX is installed and running.

Install NGINX from Red Hat Software Collections. Refer to https://www.softwarecollections.org/en/scls/rhscl/rh-nginx114/ for instructions.

Because the package name depends on the NGINX version, run yum search rh-nginx to check if a new version is available. (For NetBackup 8.3, an EEB is required if NGINX is installed from Red Hat Software Collections.)

If you configure the storage server for image sharing before NGINX is installed and enabled, run the following command on the storage server after you install and enable NGINX:

/usr/openv/pdde/vpfs/bin/vpfs_config.sh --configure_byo

It is recommended that you launch a cloud recovery server on demand and do not upgrade it.

Do not use nbdevconfig to modify cloud LSU or add new cloud LSU in the image sharing server as it might cause an issue in the image sharing server (cloud recovery server). If KMS encryption is enabled in the on-premises side after image sharing server is configured, the encrypted image cannot be imported by this image sharing server.

After the image is imported in the image sharing server, the image catalog exists in the image sharing server. If the image is expired on the on-premises NetBackup domain, then restoring the image to the image sharing server fails even though the image catalog exists in the image sharing server.

The imported image expiration time is the time for which the imported image catalog exists in the image sharing server. If the image expires in the image sharing server, the image catalog in the image sharing server is removed but the image data in the cloud storage is not removed.

You can restore any image that you import in the image sharing server. Only VM images in AWS and Azure can be recovered because they can be converted into EC2 instances in AWS or VHD in Azure. VM images in other cloud storages cannot be converted, and can only be restored. You can recover only the VM images that are full backup images or accelerator-enabled incremental backup images.

Image sharing supports many policy types.

See the NetBackup compatibility list for the latest information on the supported policy types.

After the image sharing is configured, the storage server is in a read-only mode. Some MSDP commands are not supported.

For information on the VM recovery limitations in AWS, refer to the AWS VM import information in AWS help.

You can configure the maximum active jobs when the images are imported to cloud storage.

Modify the file path /usr/openv/var/global/wsl/config/web.conf to add the configuration item as imageshare.maxActiveJobLimit.

For example, imageshare.maxActiveJobLimit=16.

The default value is 16 and the configurable range is 1 to 100.

If the import request is made and the active job count exceeds the configured limit, the following message is displayed:

"Current active job count exceeded active job count limitation".

The images in cloud storage can be shared. If Amazon Glacier, Deep Archive or Azure Archive is enabled, you cannot use image sharing for the storage servers running versions earlier than 11.0.

If Amazon Glacier, Deep Archive or Azure Archive is enabled and the LSU was previously migrated from Cloud Catalyst, you cannot use image sharing.

Regarding the errors about role policy size limitation in AWS:

Errors that occur when the role policy size exceeds the maximum size is an AWS limitation. You can find the following error in a failed restore job:

"error occurred (LimitExceeded) when calling the PutRolePolicy operation:
Maximum policy size of 10240 bytes exceeded for role vmimport"

The recover operation with AWS provider includes the AWS import process. Therefore, a vmdk image cannot be recovered concurrently in two restore jobs at the same time.

In AWS, the image sharing feature can recover the virtual machines that satisfy the Amazon Web Services VM import prerequisites.

For more information about the prerequisites, refer to the following article:

https://docs.aws.amazon.com/vm-import/latest/userguide/vmie_prereqs.html

If you cannot obtain the administrator password to use an AWS EC2 instance that has a Windows OS, the following error is displayed:

Password is not available. This instance was launched from a custom AMI, or the default password has changed. A password cannot be retrieved for this instance. If you have forgotten your password, you can reset it using the Amazon EC2 configuration service. For more information, see Passwords for a Windows Server Instance.

This error occurs after the instance is launched from an AMI that is converted using image sharing.

You cannot cancel an import job on the cloud recovery server.

If there is data optimization done on the on-premises image, you might not be able to restore the image that you have imported on the cloud recovery server. You can expire this image, import it again on the image-sharing server, and then restore the image.

After the backup job, duplication job, or AIR import job completes, you can import the images on a cloud recovery server. The images that are created by User-Archive job cannot be imported.

AIR is not supported for the image sharing server.

If you want to convert a VM image again, you must delete the VHD from the Azure blob.