Adjust the compute and storage resources of your AnalyticDB for MySQL cluster when workload demands change — scale up before a heavy data load, or scale down during off-peak periods to reduce costs.
Prerequisites
Before you begin, make sure that:
Your Alibaba Cloud account has no unpaid renewal orders
The cluster is in the Running state
Before you scale
Impacts
Understand what happens during scaling before you start:
| Impact | Details |
|---|---|
| Asynchronous jobs blocked | You cannot run SUBMIT JOB statements during scaling. Schedule scaling around jobs that depend on this statement. |
| Transient connections | Near the end of scaling, brief connection interruptions may occur. Scale during off-peak hours, or configure your application to reconnect automatically. |
| Migration time | Scaling redistributes and migrates data. Migration time scales with data volume. Monitor progress on the Cluster List page. |
| Scale-in duration | Reducing cluster size can take several hours or more, depending on data volume. |
Disk level changes
Changing the disk level has additional constraints:
You cannot change any other configuration at the same time as the disk level.
When the amount of hot data reaches the minimum capacity of the target disk level, the change becomes eligible.
Downgrading from PL1, PL2, or PL3 to PL0: data is redistributed and migrated, and
SUBMIT JOBis blocked during the operation.Upgrading to a higher disk level: data is not redistributed or migrated, and your workload is unaffected.
Cross-specification scaling
Only AnalyticDB for MySQL clusters of V3.1.3.9 or later support cross-specification scaling.
NoteTo view and update the minor version of an AnalyticDB for MySQL cluster, log on to the AnalyticDB for MySQL console and go to the Configuration Information section of the Cluster Information page.
Cross-specification scaling involves requesting resources for the new specification, full data migration, incremental data migration, and a switchover from the source cluster to the destination cluster. The time required for the entire process depends on the cluster specifications and data volume.
During the switchover, the source cluster is read-only for 5 to 10 minutes.
After you connect to the source cluster, you can execute the following statement to specify a time window for the switchover. The following example sets the time window to 23:00 to 23:30:
SET adb_config REPLICATION_SWITCH_TIME_RANGE=`23:00, 23:30`;For more information about how to connect to a cluster, see Connect to a cluster.
When you change a C32 specification cluster from reserved mode to elastic mode, the number of cores per node is reduced to 8. A BUILD task uses 3 cores by default, which increases the average CPU utilization. If the increased CPU utilization does not affect your business, no action is required. If your business is affected, you can upgrade the cluster specifications or
Billing
Changing the disk level affects hot storage costs, calculated as:
Unit price of storage for the disk level × Amount of hot storage used × DurationFor pricing details, see Data Warehouse Edition pricing and Billing for specification changes.
Scale an elastic mode cluster
For elastic mode clusters, you can change computing resources and Elastic I/O Units (EIUs) independently or together.
Log on to the AnalyticDB for MySQL console. In the upper-left corner, select the region where your cluster is located.
In the left navigation pane, click Cluster List and select a product series. Find the target cluster, then click Scale Up or Scale Down in the Actions column.
On the Upgrade/Downgrade or Downgrade page, configure the parameters described in the following table.
Parameter Description Edition Fixed to Data Warehouse Edition. Mode Fixed to Elastic Mode. Series Fixed to Cluster Edition (New). Specifications Fixed to E32. Computing resources Resources used for data computation. Increasing computing resources improves query speed. For best performance, set CPU cores for computing resources to more than the CPU cores for EIUs. Clusters at 32 cores or more support resource groups, scheduled scaling, and hot and cold data separation. Elastic I/O Unit specifications The specification for a single EIU node. Each group of EIUs contains 3 nodes. EIUs measure the storage performance of an elastic mode cluster — use them to scale storage independently of compute. For details, see Scale out elastic I/O units (EIUs). Elastic I/O Units The number of EIUs. Available only for clusters with Computing resources of 32 Cores 128 GB or higher. After scaling out EIUs, the amount of hot data locked per EIU increases. Before scaling in, confirm that your hot data volume fits within the capacity of the target EIU count. If it does not, move some data to cold storage first. NoteThis configuration is supported only for clusters that are in elastic mode and have computing resources of 32 Cores 128 GB or higher.
Disk level The disk level for the cluster. Capacity ranges: PL0 and PL1 support 20 GiB–65,536 GiB; PL2 supports 461 GiB–65,536 GiB; PL3 supports 1,261 GiB–65,536 GiB. When changing the disk level, your hot data volume must meet the minimum capacity of the target level. Complete the purchase:
Subscription: Read and accept the Terms of Service, click Buy Now, then confirm the order and payment.
Pay-as-you-go: Read and accept the Terms of Service, then click Buy Now.
Scale a reserved mode cluster
For reserved mode clusters, you can change the number of node groups and storage space independently or together.
Log on to the AnalyticDB for MySQL console. In the upper-left corner, select the region where your cluster is located.
In the left navigation pane, click Cluster List and select a product series. Find the target cluster, then click Scale Up or Scale Down in the Actions column.
On the Upgrade/Downgrade or Downgrade page, configure the parameters described in the following table.
Parameter Description Edition Fixed to Data Warehouse Edition. Series Fixed to Cluster Edition. Specifications Select the product specifications. Number of node groups Select the number of node groups. Each node group contains 3 nodes (replicas) by default. Storage space Storage capacity for a single node group. Total storage = storage per node group × number of node groups. Disk level The disk level for the cluster. PL0 and PL1 both support 20 GiB–65,536 GiB. When changing the disk level, your hot data volume must meet the minimum capacity of the target level. Complete the purchase:
Subscription: Read and accept the Terms of Service, click Buy Now, then confirm the order and payment.
Pay-as-you-go: Read and accept the Terms of Service, then click Buy Now.