Metallic enhancement for backing up on-prem workloads to your own Azure storage or Metallic Recovery Reserve in Azure

  • 14 September 2023
  • 4 replies
  • 532 views

Userlevel 2
Badge +5

What is changing in this enhancement?
Metallic has optimized the backup configuration of your on-prem workloads by hosting the deduplication cache on backup gateway i.e., near the workload, to significantly reduce network calls to the Metallic control plane to retrieve deduplication information and thereby improve performance of backup operations.

Important Note This change is only applicable to new on-prem workloads onboarded in Metallic after Aug 25. Existing workload already protected in Metallic will continue to operate as before.

Which workloads does this enhancement apply to?
This enhancement is applicable only to on-prem workloads protected using your own Azure storage (BYOS) or Metallic Recovery Reserve in Azure. 

How does this change impact my on-prem workloads onboarded in Metallic Prior to Aug 25?
Existing workload already protected in Metallic will continue to operate as before.  

How does this change impact workloads currently in the process of onboarding?
If you started onboarding on-prem workloads by creating server plans using your own Azure storage (BYOS) or Metallic Recovery Reserve in Azure prior to Aug. 25, then you will not be able to select those plans for onboarding new on-prem workloads. 

To resolve this - You must create new storage using your own Azure Storage (BYOS) or Metallic Recovery Reserve in Azure and use this newly created storage to configure a new server plan to protect new or additional on-prem workloads.

Where can I see previously created server plans?
To review previously created plans, refer to Server Plans.

What is the benefit of this enhancement?
This change improves the deduplicated data transfer operation when backing up on-premises workloads using your existing backup gateway to your own Azure storage (BYOS) or Metallic Recovery Reserve in Azure. The enhancement boosts throughput and increases the resilience of backups against networking errors.

How do I take advantage of this enhancement?
The best practice is to use separate server plans for cloud resources and on-prem resources. Note that existing configurations will continue to operate as before.

To protect new or additional on-prem workloads after Aug 25 - You must create a new storage using your own Azure storage (BYOS) or Metallic Recovery Reserve in Azure and use this new storage to create a new server plan for your on-prem workloads.

To protect new or additional Azure workloads after Aug 25 - You must create a new storage using your own Azure storage (BYOS) or Metallic Recovery Reserve in Azure and use this new storage to create a new server plan for your Azure workloads. 


4 replies

Userlevel 3
Badge +5

Hello Mohit,

  • can you describe the changes in more technical way? 
  • What are the limitations preventing form applying new, more resilient configuration to old on-prem plans? 
Userlevel 7
Badge +19

I'm also curious to learn more about this enhancement and of course is this also coming to Commvault. In addition can you easily see which plans and/or libraries are new and which ones are "old”?

Userlevel 2
Badge +5

Before the August 25th metallic release, the on-premises workloads writing to Azure storage (BYOS) or Metallic Recovery Reserve in Azure transferred deduplicated data to the storage layer. The deduplication feature enabled lookups to the control plane, impacting the overall time to backup an on-premises workload. The reliability of networks affects the backup reliability, causing multiple restarts.

 

To resolve this, we have made changes to automatically host the cache on the backup gateway near the workload when you configure a backup for a new workload. The new configuration improves network latency and removes the dependency on reliable networks for lookups. This new enhancement has a significant gain to the backup throughput. 

Userlevel 3
Badge +5

Before the August 25th metallic release, the on-premises workloads writing to Azure storage (BYOS) or Metallic Recovery Reserve in Azure transferred deduplicated data to the storage layer. The deduplication feature enabled lookups to the control plane, impacting the overall time to backup an on-premises workload. The reliability of networks affects the backup reliability, causing multiple restarts.

 

To resolve this, we have made changes to automatically host the cache on the backup gateway near the workload when you configure a backup for a new workload. The new configuration improves network latency and removes the dependency on reliable networks for lookups. This new enhancement has a significant gain to the backup throughput. 

Got it, is it possible to add that mechanism for existing configuration, e.g. to allow MediaAgents to populate cache overtime, together with running new backups etc. and achieve the same level of improved network performance? 

Reply