Commvault Grid uses Commvault File System (CVFS) to protect data against disk and node failures. CVFS provides built-in resiliency through replication and erasure coding technologies that help ensure data availability while optimizing storage efficiency.
The resiliency mode configured for a Commvault Grid cluster determines the level of fault tolerance, the data protection method used, and the minimum number of nodes required in the deployment.
You can choose Commvault Grid Resiliency Mode if you are deploying the software on Vault OS 9 Operating System.
Data Protection Technologies
Commvault Grid uses the following data protection technologies to protect data and metadata.
Replication
Commvault Grid uses replication to protect critical cluster metadata.
Metadata is replicated across multiple nodes to maintain cluster availability and ensure that the system can continue operating during node failures. Metadata replication works together with erasure coding, which protects user backup data.
| Replication Model | Replica Count | Fault Tolerance | Minimum Nodes |
|---|---|---|---|
| RF3 | 3 replicas | 1 node failure | 3 |
| RF5 | 5 replicas | 2 node failures | 6 |
Erasure Coding
Erasure Coding is the default and preferred data protection method. It provides greater storage efficiency than replication by dividing data into d data fragments and p parity fragments, enabling recovery from up to p failures.
| EC Configuration | Fragment Layout | Minimum Nodes |
|---|---|---|
| EC 4:2 | 4 data + 2 parity | 3 |
| EC 8:4 | 8 data + 4 parity | 12 |
Resiliency Modes
Commvault Grid supports three resiliency modes.
| Resiliency Mode | Metadata Protection | Data Protection | Fault Tolerance |
|---|---|---|---|
| Standard | RF3 | EC 4:2 | Any two disk failures or one node failure |
| Medium | RF5 | EC 4:2 | Any two disk failures or up to two node failures |
| High | RF5 | EC 8:4 | Any four disk failures or up to two node failures |
Standard Resiliency
Standard resiliency is designed for smaller Commvault Grid deployments.
Characteristics
- Requires a minimum of three nodes.
- Uses EC 4:2 protection.
- Protects against any two disk failures or one node failure.
Note
During temporary EC failure-domain violations, write operations may temporarily use RF3 secondary protection when erasure coding (EC) write requirements cannot be met. After normal conditions are restored, the data is automatically self-healed back to the configured EC layout.
Benefits
- Efficient use of storage capacity.
- Suitable for small and medium-sized deployments.
- Provides a balance between protection and storage efficiency.
- Uses RF3 metadata protection to provide optimal performance compared to higher resiliency modes.
Considerations
- Provides lower node-failure tolerance than Medium and High resiliency modes.
Medium Resiliency
Medium resiliency is designed for deployments that require increased node-level protection.
Characteristics
- Requires a minimum of six nodes.
- Uses EC 4:2 protection.
- Protects against any two disk failures or up to two node failures.
Benefits
- Improved availability compared to Standard resiliency.
- Additional protection against node failures.
Considerations
- Requires a larger cluster footprint than Standard resiliency.
- Medium resiliency introduces additional performance overhead to provide higher metadata resiliency than Standard resiliency.
High Resiliency
High resiliency is designed for large-scale deployments with the highest availability requirements.
Characteristics
- For clusters with 12 or more nodes, High Resiliency is the only supported resiliency mode.
- Uses EC 8:4 protection.
- Protects against any four disk failures or up to two node failures.
Benefits
- Provides the highest level of fault tolerance available in Commvault Grid.
- Supports large-scale enterprise deployments.
- Offers increased protection against simultaneous component failures.
Considerations
- Designed for environments that prioritize maximum resiliency and availability.
Default Resiliency Mode Selection
The default resiliency mode is determined by the number of nodes in the cluster during the initial deployment.
| Cluster Size | Default Resiliency Mode |
|---|---|
| 3 to 5 nodes | Standard |
| 6 to 11 nodes | Standard |
| 12 or more nodes | High |
For clusters containing 6 to 11 nodes, Medium resiliency can be selected during deployment.
Cluster Expansion
You can expand a Commvault Grid cluster by adding nodes as capacity or performance requirements increase.
Adding nodes increases the available compute and storage resources in the cluster. However, expanding the cluster does not change the resiliency mode that was selected during the initial deployment.
Choosing a Resiliency Mode
Use the following guidance when selecting a resiliency mode.
| Requirement | Recommended Mode |
|---|---|
| Small deployment with efficient storage utilization | Standard |
| Increased protection against node failures | Medium |
| Maximum fault tolerance and availability | High |