Overview of Scaling Up Storage Space

As more data is added, your GaussDB instance may start to run out of space. This section describes how to scale up the storage of your instance. Table 1 describes the scale-up methods provided by GaussDB.

**Table 1** Scale-up methods
Scale-up Method	Supported Instance Type	Scope	Description
Manually Scaling Up Storage Space for an Instance	Distributed Centralized	All shards	Shard-level scale-up is performed based on the selected storage space. All shards are scaled up at a time. The added storage space must be a multiple of 4 GB x Number of shards.
Manually Scaling Up Storage Space of Specified Shards	Distributed	Specified shards All shards	Shard-level scale-up is performed based on the selected storage space. One or more shards are scaled up at a time. The added storage space must be a multiple of 4 GB. This method is only available for distributed instances using independent deployment.
Configuring Storage Autoscaling for an Instance	Distributed Centralized	Specified shards All shards	Shard-level scale-up is performed based on the user-defined scaling increment (specified by the Scale Up By parameter). When the available storage of a DN component drops to or below the specified threshold, an automatic scale-up task is triggered to scale up the shard where the DN component with insufficient available storage is located. Storage can be scaled by percentage or a fixed amount. Percentage: The storage is scaled up by a percentage you specify. The added storage space is the storage space of the shard multiplied by the scaling increment. Fixed amount: The storage is scaled up by a fixed amount you specify. The added storage space is a fixed value. Autoscaling is unavailable for distributed instances using combined deployment.

Introduction Video

Manually Scaling Up Storage Space for an Instance

Take an independently deployed distributed instance as an example. This instance has three shards, three replicas, and three CNs, which are the default configurations, with a total storage of 120 GB before the scale-up. This means that each shard is allocated 40 GB of capacity (each DN within the shard is provisioned with 40 GB). When you increase the storage capacity by 12 GB, the total storage after the scale-up reaches 132 GB. Each shard gets an extra 4 GB, and every DN within that shard also receives a 4 GB increase. This brings the capacity of each shard to 44 GB, with each DN provisioned accordingly.

Figure 1 Scaling up storage space
Click to enlarge

Manually Scaling Up Storage Space of Specified Shards

Take an independently deployed distributed instance as an example. This instance has three shards, three replicas, and three CNs, which are the default configurations, with a total storage of 120 GB before the scale-up. This means that each shard is allocated 40 GB of capacity (each DN within the shard is provisioned with 40 GB). When you apply scale-up specifically to shard 1 and shard 2 with increments of 4 GB and 8 GB, respectively, the total storage increases to 132 GB. Shard 1 grows from 40 GB to 44 GB, with each DN in that shard provisioned at 44 GB. Shard 2 expands from 40 GB to 48 GB, and every DN within it is provisioned at 48 GB.

Figure 2 Scaling up storage space
Click to enlarge

Configuring Storage Autoscaling for an Instance

By percentage
Take the independently deployed distributed instance with three shards, three replicas, and three CNs as an example. Its total storage space before the scale-up is 120 GB, with each shard holding 40 GB. In the autoscaling policy, both the Trigger If Available Storage Drops To and Scale Up By parameters are set to 20%. When the available storage of shard 1 drops to 20% or below, the system automatically triggers a scale-up. This adds 8 GB to shard 1 (calculated as 40 GB × 20%). After the scale-up, the storage of shard 1 grows to 48 GB, with each DN within the shard provisioned at 48 GB. The instance's total storage then reaches 128 GB.

Figure 3 Scaling up storage space
By fixed amount
Take the independently deployed distributed instance with three shards, three replicas, and three CNs as an example. Its total storage space is 120 GB before scale-up, and each shard has a storage space of 40 GB. In the autoscaling policy, the Trigger If Available Storage Drops To parameter is set to 20%, and the Scale Up By parameter is set to 40 GB. When the available storage of the DN 1_1 and DN 3_3 components drops to 20% or below, they system automatically triggers a scale-up for the shards containing DN 1_1 and DN 3_3. Each of these two shards receives a 40 GB storage increase. After the scale-up, the total storage reaches 200 GB. Shard 1 and shard 3 each increase by 40 GB, growing from 40 GB to 80 GB, with every DN in these shards provisioned at 80 GB.

Figure 4 Scaling up storage space