About Image Sharing using MSDP cloud
Use image sharing to share the images from your on-premises NetBackup server to another NetBackup server. The NetBackup server that is configured for image sharing is called Cloud Recovery Server (CRS). Image sharing also provides the ability to convert backed up VMs as AWS instances or Azure VHD in certain scenarios.
MSDP with image sharing is a self-describing storage server. When you configure image sharing, NetBackup stores all the data and metadata that is required to recover the images in the cloud.
Note:
The Cloud Recovery Server version must be the same or later than the on-premises NetBackup version.
The following table describes the image sharing feature workflow.
Table: Image sharing workflow
Task | Description |
|---|---|
Prepare a cloud recovery server. | You must have a virtual machine in your cloud environment and have NetBackup installed on it. You can deploy the virtual machine using one of the following ways.
|
Configure the NetBackup KMS server. | If KMS encryption is enabled, perform the following tasks. |
Configure image sharing on the cloud recovery server. | The NetBackup virtual machine in the cloud that is configured for image sharing is called a cloud recovery server. Perform the following steps to configure the image sharing: |
Prepare images for import for high-latency storage. | This step is applicable only for high-latency storage such as Amazon Glacier, Amazon Glacier Deep Archive, and Azure Archive storage. |
Use the image sharing. | After you configure this NetBackup virtual machine for image sharing, you can import the images from your on-premises environment to the cloud and recover them when required. You can also convert VMs to VHD in Azure or AMI in AWS. |
Read additional information about image sharing. |
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.
For Veritas Alta Recovery Vault, 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 Veritas Alta Recovery Vault account on the image sharing server.
The temporary bucket or blob container name format is vrtsonvert-<timestamp>/VRTSConvert-<timestamp>.
For Veritas Alta Recovery Vault Amazon, MSDP-C credentials with AWS account with IAM and EC2 related permissions must be created before the VM conversion. For Veritas Alta Recovery Vault Azure, MSDP-C credentials with Azure general-purpose storage accounts must be created before the VM conversion.
For Veritas Alta Recovery Vault, only the image import function uses Veritas Alta Recovery Vault 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 Veritas Alta Recovery Vault.
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/hostsfile:"External IP" "Server's FQDN"
"Internal IP" "Server's FQDN"
For a computer, add the following items in the
/etc/hostsfile:"IP address" "Server's FQDN"
(Optional) For an instance, change the search domain order in the
/etc/resolv.conffile 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
You can access NetBackup web UI to use image sharing. For more information, refer to the Create a Media Server Deduplication Pool (MSDP) storage server for image sharing topic in the NetBackup Web UI Administrator's Guide.
After installing NetBackup, you can run the ims_system_config.py script to configure image sharing.
The path to access the command is: /usr/openv/pdde/pdag/scripts/.
Amazon Web Service cloud provider:
ims_system_config.py -t PureDisk -k <AWS_access_key> -s <AWS_secret_access_key> -b <name_S3_bucket> -bs <bucket_sub_name> [-r <bucket_region>] [-p <mount_point>] [-sc <storage class>]
If you have configured IAM role in the EC2 instance, use the following command:
ims_system_config.py -t PureDisk -k dummy -s dummy <bucket_name> -bs <bucket_sub_name> [-r <bucket_region>] [-p <mount_point>]
Microsoft Azure cloud provider:
ims_system_config.py -cp 2 -k <key_id> -s <secret_key> -b <container_name> -bs <bucket_sub_name> [-p <_mount_point_>] [-sc <storage tier>]
Other S3 compatible cloud providers (For example, Hitachi HCP):
If Cloud Instance has been existed in NetBackup, use the following command:
ims_system_config.py -cp 3 -t PureDisk -k <key_id> -s <secret_key> -b <bucket_name> -bs <bucket_sub_name> -c <Cloud_instance_name> [-p <mount_point>]
Or use the following command:
ims_system_config.py -cp 3 -t PureDisk -k <key_id> -s <secret_key> -b <bucket_name> -pt <cloud_provider_type> -sh <s3_hostname> -sp <s3_http_port> -sps <s3_https_port> -ssl <ssl_usage> [-p <mount_point>]
Example for HCP provider:
ims_system_config.py -cp 3 -t PureDisk -k xxx -s xxx -b emma -bs subtest -pt hitachicp -sh yyy.veritas.com -sp 80 -sps 443 -ssl 0
Description: (Specify the following options to use HCP cloud)
-cp 3: Specify the third-party S3 cloud provider that is used.
-pt hitachicp: Specify the cloud provider type as hitachicp (HCP LAN)
-t PureDisk: Specify the storage server type as PureDisk.
-sh <s3_hostname>: Specify an HCP storage server host name.
-sp <s3_http_port>: Specify an HCP storage server HTTP port (Default is 80).
-sps <s3_https_port>: Specify an HCP storage server HTTP port (Default is 443).
-ssl <ssl_usage>: Specify whether to use SSL. (0- Disable SSL. 1- Enable SSL. Default is 1.) If SSL is disabled, it uses <s3_http_port> to make a connection to <s3_hostname>. Otherwise, it uses <s3_https_port>.
Note:
Configuring image sharing using MSDP cloud with the ims_system_config.py script is not supported for SUSE Linux Enterprise. Use NetBackup web UI to configure image sharing using MSDP cloud for SUSE Linux Enterprise.
If you are using image sharing for high-latency storage such as Amazon Glacier, Amazon Glacier Deep Archive, and Azure Archive storage, you must prepare the images before you import them to the cloud storage. For the source version 11.0 or later, MSDP automatically prepares images for import in the background.
To prepare the images for import to the high-latency storage
- If the source version is earlier than 11.0 or the source server is down and did not finish pre-import operations, run the following command to prepare the images for import:
tiermover --start --lsu <LSU name> --client <client> --policy <policy> [--backupid <backup ID>] [--retrieval Bulk|Standard|Expedited] [--verbose] [--debug]
- If the source version is 11.0 or later, MSDP automatically prepares images for import in the background. Wait for images to be ready for the import. Run the following command to check the status:
tiermover --status --lsu <LSU name> [--client <client>] [--policy <policy>] [--backupid <backup ID>] [--sobins] [--active] [--lsulist] [--debug] [--verbose]
Note:
Import fails if the images are not ready for the import.
You can access NetBackup web UI to use image sharing. For more information, refer to the Share images from an on-premises location to the cloud topic in the NetBackup Web UI Administrator's Guide.
You can use the nbimageshare command to configure image sharing.
Run the nbimageshare command to list and import the virtual machine and standard images and then recover the virtual machines.
The path to access the command is: /usr/openv/netbackup/bin/admincmd/
For more information about the nbimageshare command, refer to the NetBackup Commands Reference Guide.
The following table lists the steps for image sharing and the command options:
Table: Steps for image sharing and the command options
Step | Command |
|---|---|
|
Log on to NetBackup. |
nbimageshare --login <username> <password> nbimageshare --login -interact |
|
List all the backup images that are in the cloud. |
nbimageshare --listimage <LSU name> <MSDP image sharing server> Note: In the list of images, the increment schedule type might be differential incremental or cumulative incremental. |
|
Import the backup images to NetBackup. |
Import a single image: nbimageshare --singleimport <client> <policy> <backupID> <LSU name> <MSDP image sharing server> Import multiple images: --batch-import <image list file path> <LSU name> <MSDP image sharing server> Note: The format of the image_list_file_path is same as the output of "list images". You can import the multiple images. For every 100 images, a new import job is created. You can import an already imported image. This action does not affect the NetBackup image catalog. |
|
Recover the VM as an AWS EC2 AMI or VHD in Azure. |
nbimageshare --recovervm <LSU name> <MSDP image sharing server>
|
When KMS encryption is enabled, you can share the images in the cloud storage to the cloud recovery server with manual KMS key transfer.
On-premises side:
Storage server: Find the key group name for the given Storage server.
Find contentrouter.cfg in /etc/pdregistry.cfg
Find the key group name is in contentrouter.cfg under [KMSOptions]
(Example KMSKeyGroupName=amazon.com:test1)
NetBackup primary server: Exports the key group with a passphrase to a file:
/usr/openv/netbackup/bin/admincmd/nbkmsutil -export -key_groups <key-group-name> -path <key file path>
cloud recovery server (cloud side):
Copy the exported key to the cloud recovery server.
Configure the KMS server.
/usr/openv/netbackup/bin/nbkms -createemptydb /usr/openv/netbackup/bin/nbkms /usr/openv/netbackup/bin/nbkmscmd -discovernbkms -autodiscover
Import keys to KMS service.
/usr/openv/netbackup/bin/admincmd/nbkmsutil -import -path <key file path> -preserve_kgname
Configure the cloud recovery server using NetBackup web UI or with ims_system_config.py
On-premises KMS key changes:
In the case of KMS key changes for the given group for on-premises storage server after the cloud recovery server is set up, you must export the key file from the on-premises KMS server and import that key file on the cloud recovery server.
On-premises NetBackup primary server:
Exports the key group with a passphrase to a file:
/usr/openv/netbackup/bin/admincmd/nbkmsutil -export -key_groups <key-group-name> -path <key file path>
Cloud recovery server:
/usr/openv/netbackup/bin/admincmd/nbkmsutil -deletekg -kgname <key-group-name> -force
/usr/openv/netbackup/bin/admincmd/nbkmsutil -import -path <key file path> -preserve_kgname
If an on-premises storage server is configured to use keys from an external KMS server, then make sure that the same KMS server is configured on the cloud recovery server before running ims_system_config.py. To know more about configuring an external KMS server in NetBackup, refer to NetBackup Security and Encryption Guide.
Make sure that the external KMS server is reachable from the cloud recovery server on a specific port.
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.
Cloud LSU requires free disk space. When you configure image sharing server using the ims_system_config.py script, ensure that you have enough disk space in the default mount point or storage, or you can use -p parameter of ims_system_config.py to specify a different mount point to meet the requirement of free disk spaces.
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.confto 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"
Workaround:
You can change the maximum policy size limit for the vmimport role.
You can list and delete the existing policies using the following commands:
aws iam list-role-policies --role-name vmimport aws iam delete-role-policy --role-name vmimport --policy-name <bucketname> -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.
For more information, refer to the following articles:
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.