Auto-scaling enables dynamic scaling of access nodes to meet changing workload demands. Based on a defined scaling profile, the system automatically deploys access nodes within a region only when required and decommissions them once operations are complete. This ensures optimal performance while minimizing unnecessary resource consumption.
Auto-scaling is particularly critical during large-scale backup, restore, and disaster recovery operations, where data volumes and workload demands can vary significantly. By allocating access nodes on demand and powering them down when no longer needed, the system prevents bottlenecks, reduces operational overhead, and optimizes compute costs.
Supported workloads
-
Amazon Elastic File System (EFS)
-
Amazon FSx for NetApp ONTAP
Scale manager
A scale manager is a Linux-based node that orchestrates auto scale operations for access nodes. It acts as the control layer between the CommServe and the customer’s AWS environment.
When a backup, content browse, or restore operation requires an access node, the CommServe sends a request to the scale manager. The scale manager uses the configured AWS STS AssumeRole credentials to call AWS APIs and launch a temporary EC2 instance from the Commvault Marketplace image. This instance functions as an auto-scaled access node.
The scale manager manages the lifecycle of these access nodes. It ensures that the nodes are initialized, registered with the CommServe, and available to execute workloads. After the operation completes, the scale manager terminates the EC2 instances based on inactivity or workload completion to avoid unnecessary resource usage.
The scale manager does not perform backup or data transfer operations. It handles orchestration tasks such as provisioning, scaling, and cleanup, ensuring that access nodes are available only when needed.
The scale manager can run multiple backup, content browse, or restore jobs on a dynamically launched access node when sufficient resources are available.
Supported AWS regions
Auto-scaling is supported for all commercial and GovCloud AWS Regions.
Amazon Machine Images (AMIs) used for auto-scaled access nodes
The software creates the auto-scaled access nodes by cloning an Amazon Machine Image (AMI). Commvault provides a version of the following AMIs (CIS hardened) for each platform release. To use these images, subscribe to them on AWS Marketplace. The AMIs are deployed for each AWS Region that has instances that reside in the AWS account.
- Commvault Cloud - Amazon Linux Access Node BYOL: Amazon Linux 2023
Amazon EC2 instance types used for auto-scaled access nodes
The following AWS instance types are supported:
-
c5.2xlarge
-
c6a.xlarge
-
c6i.2xlarge
-
c7i.xlarge
-
m5a.xlarge
-
m7a.xlarge
Deployment location of auto-scaled access nodes
The access nodes are created in zones or regions, depending on the region in which the cloud file share resides.
Network security requirements for autoscale (Amazon FSx for NetApp ONTAP only)
When autoscale deploys compute resources, the system automatically attaches the default network security group associated with your cloud account.
Important
The default network security group in your environment might allow a wide range of inbound and outbound traffic. If overly permissive rules are configured, compute resources created by autoscale might be exposed to external networks.
To reduce security risks, review the rules defined in the default network security group and restrict access to only the required ports and sources.
Required ports
Make sure that the following ports are allowed in the network security group attached to autoscale compute resources:
-
8400
-
8403
-
443 (HTTPS)
These ports are required for standard communication between auto-scaled components and the Commvault infrastructure.
Support for NetApp Data Protection Volumes (Amazon FSx for NetApp ONTAP only)
- To support NetApp Data Protection volumes with auto-scale, ensure that the volume export policy allows the subnet CIDR used for auto-scale. This enables dynamically created auto-scale access nodes to mount and access the volumes successfully.