ApsaraDB for HBase supports two capacity expansion methods: adding nodes (scale out) to increase compute and storage capacity, or expanding data disk capacity (scale up) to increase storage only. Both operations are available in the Actions column of the ApsaraDB for HBase console.
Before you scale
Before triggering either operation, verify the following:
-
Your client has retry enabled. By default, the retry feature is enabled. If you configured a short timeout on your client, request errors may occur during rebalancing — confirm retry is active before scaling.
-
If possible, schedule the operation during off-peak hours to minimize user-facing impact.
Add nodes
When you add nodes, each new node inherits the same configuration as the existing nodes — CPU cores, memory, disk type, and disk capacity. Adding nodes also adds disks, which increases the total storage capacity of the instance.
Limits:
| Limit | Value |
|---|---|
| Maximum nodes added per operation | 5 |
| Maximum nodes per instance | 100 |
| To exceed the 100-node limit | Submit a ticket |
Impact of adding nodes
After new nodes are added, ApsaraDB for HBase runs the server load balancing service to redistribute data partitions across nodes.
Scale-out operations have controllable impacts under normal conditions. To minimize disruption, perform scale-out during off-peak hours.
During partition transfers:
-
Workloads on a partition being transferred are interrupted briefly.
-
The interruption is within 200 milliseconds under normal conditions.
-
The client retry feature is enabled by default and automatically retries interrupted operations.
If your client has a short timeout configured, request errors may occur during the rebalancing period. Verify that retry is enabled on your client before scaling.
Increase disk capacity
To increase storage without adding nodes, expand the data disks attached to the instance using the Expand Disk button in the Actions column.
Limits:
| Limit | Value |
|---|---|
| Minimum increase per disk | 40 GB |
| Maximum capacity per disk | 2,000 GB |
Impact of expanding disk capacity
After the disk expansion completes, the system performs a rolling restart on the RegionServers and DataNodes. The impact on the instance is minor.