About automated disaster recovery in cloud using Cloud Catalyst
Automated disaster recovery (DR), provides a self-describing storage solution over Cloud Catalyst. Cloud Catalyst with automated DR in cloud is a self-describing storage server. Cloud Catalyst without automated DR in cloud is not a self-describing storage server.
Currently, NetBackup supports automated DR in cloud for VMware VM backups and Standard backups.
This topic contains the following sections:
Cloud Catalyst 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.
Automated DR 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 automated DR on demand and recover the backup images from cloud.
Automated DR discovers the backup images that are stored in AWS S3 through the REST APIs, recovers the NetBackup catalog, and restores the images.
You can use command line options from NetBackup that have the function as REST APIs.
Before you install NetBackup, create an instance based on RHEL 7.3 or later (up to RHEL 8.0) in AWS. The recommendation is that the instance has more than 64 GB of memory, 8 CPUs, and IPv4-only network.
The HTTPS port 443 is enabled.
Change host name to the server's FQDN.
Add the following items in the
/etc/hosts/file:"External IP" "Server's FQDN"
"Internal IP" "Server's FQDN"
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.
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 image sharing server. You can expire this image, import it again on the image-sharing server, and then restore the image.
After installing NetBackup, you can run the ims_system_config.py script to configure Automated DR.
Use the following command to run the ims_system_config.py script:
python /usr/openv/pdde/pdag/scripts/ims_system_config.py -k <AWS_access_key> -s <AWS_secret_access_key> -b <name_S3_bucket>
If you have configured IAM role in the EC2 instance, use the following command:
python /usr/openv/pdde/pdag/scripts/ims_system_config.py -k dummy -s dummy -b <name_S3_bucket>
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 automated DR and the command options:
Table: Steps for automated DR and the command options
| Step | Command |
|---|---|
|
Log on to NetBackup |
nbimageshare -login <username> <password> |
|
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 multiple images number must be equal to or less than 64. |
|
Recover the VM as an AWS EC2 instance |
nbimageshare -recovervm <client> <policy> <backupID>
|
After the image is imported to cloud, the image catalog still exists on the cloud. If the image is expired on the on-premises storage, then restoring the image on the cloud fails even though the image catalog exists on the cloud.
If the image expires on the cloud storage, the image catalog in the cloud is removed but the image data in the bucket is not removed.
You can only recover the VM images that full backup images or accelerator incremental backup images to cloud.
Automated DR supports VMware and Standard policy types in NetBackup 8.2 or later in the optimum deduplication and Auto Image Replication scenarios. In the optimum deduplication and Auto Image Replication scenarios, Cloud Catalyst, where the images are shared, must have a new installation of NetBackup 8.2.
After the automated DR is configured, the storage server is read-only in the DR mode.
For information on the VM recovery limitations, 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".
In optimized deduplication or AIR cascading scenarios, only the images in Cloud Catalyst that has optimized deduplication or has an AIR target can be shared.
If Cloud Catalyst is not set for optimized deduplication or is not an AIR target, you cannot use automated DR. If Amazon Glacier is enabled in Cloud Catalyst, you cannot use automated DR.
In these scenarios to disable automated DR:
Modify the
<install_directory>/etc/puredisk/spa.cfgfile and add the following configuration item:EnableIMandTIR=false
Regarding the errors about role policy size limitation:
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 includes AWS import process. Therefore, a vmdk image cannot be recovered concurrently in two restore jobs at the same time.
The Automated Disaster Recovery feature can recover 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 are unable to 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 automated disaster recovery.
For more information, refer to the following articles: