Pure Storage FlashBlade NFS and SMB filebackup with Veeam Availability Suite V10 now possible

[ NOTE: machine translation with the help of DeepL translator without additional proofreading and spell checking ]


It's February 18, 2020 and Veeam Software releases its new software release with numerous new features - Veeam Availability Suite V10. Among other things, the long-awaited NAS functionality is now available. As of now, Veeam can now back up any common form of fileshares...an often-suffering kO criterion for Veeam projects is now a thing of the past. I'm especially pleased because the two ecosystems and user experience of Veeam and Pure Storage are very similar, and the Veeam Pure stack together brings many capabilities to the table when backing up and restoring Pure Storage file/object storage FlashBlade.


The key innovations are:


  • NAS backup for NFS V3+V4.1/SMB V1+V2+V3/Windows and Linux file servers on NON-NDMP technology.

-> FlashBlade supports: NFS V3+V4.1 and SMB V2.1


  • Veeam DataLabs: StagedRestore (backup scanning with built-in deletion of personal data) and Secure Restore (scanning of restores to prevent viruses/malware from being re-transmitted to production environment during recovery)

  • Cloud Tier: automatic transfer of backups to the cloud as a capacity tier for long-term retention.


Fortunately, I had the Availability Suite V10 in access for quite some time and could already test it excessively. In the following article I would like to show you an impression especially in the use case with FlashBlade.


How does the file share backup work in V10?


When Veeam Backup & Replication runs a new file share job, the following operations are performed. 1:


1. Veeam Backup & Replication instructs the file proxy (a new component in V10) to index folders on the file share and build a tree structure of files and folders.


2. The file proxy transfers the created tree structure to the cache repository (a new component in V10)


3. The cache repository [NEW!] stores the tree structure for files and folders. When the file proxy receives a new tree structure from the cache repository, it compares the new tree structure with the tree structure created for the previous restore point and the tree structure stored in the cache repository.

If any files or folders in the file share have changed from the last backup run, the cache repository instructs to reserve resources and start the file proxy reading this data from the source file share


4. The file proxy reads new data/files from the share


5. The file proxy creates data packets.

Data packets contain backup data, which are stored (each 64 MB in size) together with metadata files (metadata describes the names and versions of the backup files and the mapping of the data in the backup data)


7. The file proxy sends new backup data to the backup repository.


Licensing


With the switch to Veeam Universal Licensing (VUL), licensing for NAS shares was also announced in advance. In the weighting, 250GB are considered as one instance. Conversely, this means that a 1TB share consumes 4 instances.


For more information on licensing, please contact Veeam or directly the appropriate account team at Veeam.


Establishment


With Veeam V10, a new Veeam role/component has been introduced, the so-called File Proxy [NEW!] The file proxy takes over the data transfer for NAS backups during a backup, as well as during the restore.

The classic Veeam proxy does not transfer backup data to the backup storage when backing up shares. However, a combined role of classic Veeam proxy and file proxy is theoretically also possible.

A file proxy is the component that sits between the file share and other components of the backup infrastructure. The file proxy acts as a data mover from the source file share to the backup repository. It processes backup jobs and trades backup and restore traffic.


The use of file proxies is essentially similar to traditional backup proxies for VM backups.

To improve backup performance, you can use multiple file proxies to enable backup across multiple threads.


Visually, I have two shares (NFS and SMB) that I want to back up.

The NFS share "SPLUNK-***" with NFSv3 as data protocol. The share "veeambackup" is published with SMB.

For redundancy reasons, I have prepared two virtual machines, both of which take on a file proxy role. With this setup, we are also able to perform file backups when one machine is unavailable.

Both proxies were equipped with a Windows Server 2016 4vCPUs and 8GB RAM.

Next we need to announce/register the shares to be backed up to Veeam, this is done via Inventory and File Shares [NEW!] - this tab is new with V10.


Integration NFS share


First we add the NFS share.

It is important at this point that the file proxies are released with created restrictions in the export rules on the NFS server (here in the example on the NFS file system of the FlashBlade). Here the IP address with the access permissions (rw-read/write, r-read) of the file proxies would have to be stored. In my test environment there are no restrictions here.

When backing up, you have the option to use previously created storage snapshots as the backup source. To ignore the snapshot functionality, select "Backup directly from the file share". Veeam Backup & Replication will then ignore open/locked files and folders. When you create a backup job, you can configure notifications which lists files and folders that were skipped (due to locking) during the backup process.


However, you can also backup file shares from native storage snapshots as mentioned above, select "Backup from storage snapshot at the following path" and specify the path to the folder on the NFS file share where the snapshots are stored in the Server:/folder format. If you select this option, you can also use scripts, for example, to create a snapshot before the backup and remove it after the backup. You can define these scripts when creating a new backup job.


Select "Failover to direct backup if a snapshot is not available" if you want Veeam Backup & Replication to read data for backup directly from the file share when the storage snapshot is not available. If you do not select the option and the snapshot is not available, the file share backup job will terminate with an error.

When backing up shares, file backup uses a so-called cache repository: a cache repository is a location where Veeam Backup & Replication stores temporary metadata. Using this cached metadata during the backup process reduces the load on the file share and makes it possible to perform incremental backups from the file share very quickly and efficiently.

To minimize network load, place the cache repository between the file share and the file proxy, and at the same time as close as possible to the file share on the network: at best, they should be one hop apart.

In my example, I used the FlashBlade itself as the cache repository, since this system provides more than enough performance. Of course, the load-offload makes no sense at that point in this setup!

As you can read in the note, losing the cache repository is not a problem, as it can be restored at any time.


Use the Backup I/O control slider to set the power/effect ratio during backup. This setting is based on the number of parallel threads that can be used by all proxies configured for file share processing. Backup-I/O-control defines how fast all proxies can read source data. If your NAS device resources are limited, I recommend using the lower impact option. If your NAS device is powerful enough, select the "faster backup" option.


Integration SMB share


Mounting an SMB share is no different than the setup shown for NFS shares, except for entering access credentials and using VSS snapshots on Windows servers.

To back up files from Microsoft VSS snapshots, select Backup from Microsoft VSS Snapshot. If you select this option, make sure that the file share and file proxy used for the file backup job support SMB protocol version 3.0 or later.


Backup


NFS backup


The setup of a "file share backup" is done via already known ways. It is important to know that a "Virtual machine backup job" cannot be combined with a file backup. A new backup job must be created!

The previously created file shares are now available for selection. I make no restriction here and select the whole share. There is also the possibility to save single subfolders.

It is also possible to apply filters to folders, files and certain file formats. This would be possible via the "Advanced" button in the lower part of the dialog window.

In the next step we select our backup repository and define the retention period by days/months/years.

In case of File Backup Jobs, it is also possible for us to move the backup content to an Archive repository for additional days/months/years after the regular retention and for Long-Term Retention.

The Advanced Settings in the "Storage" tab are the most important advanced settings in my eyes, because you can set essential settings for saving access/security settings here. By default, only ACL settings on folder level are saved. If you maintain ACL settings on file level, the setting on "Files and folders" must be adjusted accordingly, otherwise unauthorized access may occur during a restore. However, this setting logically results in a longer backup time.


In the last tab of the Advanced Settings you can still store your pre/post scripts here if required.

Classic Veeam backup copy jobs cannot be created for file backups!

Here, (synchronous) additional copies are set up directly within the file backup job. Additional copies are started immediately after the actual "primary" backup job. A limitation of the additional copy, i.e. a definition of the copy time window, also takes place at this point. A different retention in the "secondary" can also be set/applied.

Our NFS backup job is set up.

A completed NFS backup job:

(the performance data are to be considered neutral within the scope - of the available system resources - of the test environment!)

SMB backup


The setup of SMB backup jobs compared to NFS backup jobs does not differ.

SAME AS WITH NFS: The previously created file shares are now available for selection. I make no restriction here and select the whole share. There is also the possibility to backup single subfolders.

SAME AS WITH NFS: It is also possible to apply filters to folders, files and certain file formats. This would be possible via the "Advanced" button in the lower part of the dialog window.

SAME AS WITH NFS: In the next step we select our backup repository and define the retention period by days/months/years.

For file backup jobs, it is also possible for us to move backup content to an archive repository after regular retention for additional days/months/years for long term retention.

SAME AS WITH NFS: The Advanced Settings in the "Storage" tab are in my eyes the most important advanced settings, because you can set essential settings to backup access/security settings here. By default only ACL settings on folder level are backed up. If you maintain ACL settings on file level, the setting on "Files and folders" must be adjusted accordingly, otherwise unauthorized access may occur during a restore. However, this setting logically results in a longer backup time.


In the last tab of the Advanced Settings, you can still store your pre/post scripts here if required.


Classic Veeam backup copy jobs cannot be created for file backups!

Here, (synchronous) additional copies are set up directly within the file backup job. Additional copies are started immediately after the actual "primary" backup job. A limitation of the additional copy, i.e. a definition of the copy time window, also takes place at this point. A different retention in the "secondary" can also be set/applied.


A completed SMB backup job: