Version 10 of Veeam Backup & Replication isn’t too far away and we are currently at the end of a second private BETA for our customers and partners. There has been a fair bit of content released around v10 functionality and features from our Veeam Vanguard’s over the past couple of weeks and as we move closer to GA, as part of the lead up, I am doing a series on some of the cool new enhancements that are coming as part of the release. These will be quick short takes that give a glimpse into what’s coming as part of the v10 release.
Mounting Object Storage Repository for Streaming Disaster Recovery
The Cloud Tier was introduced in Veeam Backup & Replication 9.5 Update 4 and focused on the offloading of data from local repositories to Object Storage repositories. Essentially looking to reduce the cost and overheads of ever growing local primary repositories. Due to the smarts we built into the feature, the use cases for Cloud Tier started to expand beyond the offloading of data and looked at recovery options.
Because we hold a replicated copy of the VBK metadata as well the actual backup data that is indexed as blocks in Object Storage we have the ability to leverage the data sitting there for recovery purposes. I’ve already shown this a number of times this year, and presented on the recovery and resiliency of the Cloud Tier at Cloud Field Day 5.
With v10, we have made this process even easier by introducing a Mount function that will enable users to import backup restore points for recovery purposes in the case of disaster. This can even be done with the Community Edition which means that the Cloud Tier now becomes a mechanism for recovery from any device to almost any platform.
Quickly going over how this works, the first step is to recreate the Object Storage Repository with the same settings as the one existed in the original location.
At this point we can leverage the new v10 feature that allows you to Import the backup data contained on the Object Storage Repository by right clicking on the repository and selecting Import Backups.
This will store the available restore points in the Backup & Replication database and have them appear under Imported Backups in the console.
It’s important and cool to note, that as this stage we haven’t downloaded the metadata shells that constitute the de-hydrated VBK. One of the extra smart things we have built into this feature is that the metadata and VBK shells are only downloaded once a restore operation has been started, meaning quicker setup and more specific re-syncing of the metadata shells.
On that note, all existing restore operations are available at this point.
Once a restore operation is triggered, only then is the required metadata downloaded and reconstructed into the required shell chain to a temp directory. The example below shows the shells of a full and an incremental triggered by an Instant VM Recovery (IVMR) Operation.
The data required to perform the IVMR is streamed from the Object Storage Repository (Capacity Tier Extent).
Once restore operations have been completed you can go back to the Object Storage Repository, right click and select Detach
This unmounts the Object Storage and removed the restore points from the Imported Backup view and deletes the downloaded contents of the temp folder where the metadata shells where staged.
Wrap Up:
That was a quick look at one of my personal favourite new enhancements in v10. We have improved an operation that was being leveraged in 9.5 Update 4 due to the smarts built into the Cloud Tier for recovery operations and made it quicker and more efficient. This also allows users to effectively restore to any platform from any device that has Veeam Backup & Replication installed and has access to the Object Storage platform!
When put together with the new Copy mode being introduced into v10, we all of a sudden have a solution that can achieve very low RPO and RTO for disaster recovery… more to come on that aspect when v10 launches.
Stay tuned over the next few weeks as I go through some more hidden gems.
Disclaimer: The information and screen shots in this post is based on BETA code and may be subject to change come final GA.