About Image Sharing using MSDP cloud
Image sharing provides a self-describing storage solution over MSDP cloud. MSDP with Image sharing in cloud is a self-describing storage server. MSDP without Image sharing in cloud is not a self-describing storage server.
Image Sharing lets you use an easy and visualized way to manage and provision images in cloud object storage and even the ability to convert backed up VMs as AWS instances or Azure VHD in certain scenarios.
Before you begin, see the following topics:
In a situation where MSDP cloud backed up the deduplicated data to cloud, but the NetBackup catalog was available only on the on-premises NetBackup server. There, the data cannot be restored from the cloud without the on-premises NetBackup server.
Image sharing in 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 an all-in-one NetBackup in the cloud on demand called the cloud recovery host, and recover the backup images from 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.
Before you install NetBackup, create an instance based on RHEL 7.3 or later in cloud. You can also set up a computer based on 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 host name to the server's FQDN.
In Azure virtual machine, you must change the internal hostname, which is created automatically for you and cannot get internal hostname from 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.NetBackup should be an all-in-one setup.
Refer to the NetBackup Installation Guide for more information.
You can access NetBackup Web UI to use image sharing. For more information, refer to the Create a cloud recovery host 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>]
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_>]
Other S3 compatible cloud provider (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 third-party S3 cloud provider that is used.
-pt hitachicp: Specify cloud provider type as hitachicp (HCP LAN)
-t PureDisk_hitachicp_rawd: Specify storage server type as PureDisk_hitachicp_rawd
-sh <s3_hostname>: Specify HCP storage server host name
-sp <s3_http_port>: Specify HCP storage server HTTP port (Default is 80)
-sps <s3_https_port>: Specify 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 connection to <s3_hostname>. Otherwise, it uses <s3_https_port>.
You can access NetBackup Web UI to use image sharing. For more information, refer to the Using image sharing from the NetBackup Web UI 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 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> Import multiple images: nbimageshare --batchimport <image_list_file_path> Note: The format of the image_list_file_path is same as the output of "list images". The multiple images number must be equal to or less than 64. 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 <client> <policy> <backupID>
|
When KMS encryption is enabled, you can share the images in the cloud storage to the cloud recovery host 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 key group name is in contentrouter.cfg under [KMSOptions]
(Example KMSKeyGroupName=amazon.com:test1)
NetBackup master 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 host (cloud side):
Copy the exported key to the cloud recovery host
Config 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 host using NetBackup Web UI or with ims_system_config.py
On-premises KMS key changes:
In case of KMS key changes for the given group for on-premises storage server after the cloud recovery host is set up, you must export the key file from on-premises KMS server and import that key file on the cloud recovery host.
On-premises NetBackup master 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 host:
/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 external KMS server, then make sure that the same KMS server is configured on the cloud recovery host 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 host on a specific port.
We recommend that you launch a cloud recovery host in the cloud on demand and don't 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 host). If KMS encryption is enabled in on-premise side after image sharing server is configured, the encrypted image cannot be import 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.
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 instance 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 lists for the latest information on the supported policy types.
After the image sharing is configured, the storage server is in a read-only mode.
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.
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 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 host.
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 host. 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 host.
If you want to convert a VM image again, you must delete the VHD from Azure blob.