RDS for MySQL local-disk instances support fast scaling, which scales up instances in-place without data migration or connection interruptions when the host has sufficient resources.
How it works
During a standard instance type change, if the current host lacks sufficient resources, the system migrates the instance to another host. This involves data migration and causes a connection interruption of about 30 seconds during switchover.
Fast scaling checks for idle resources on the current host to determine whether the instance can be scaled up in-place, without data migration or connection interruptions.
Comparison of the two scaling paths:
-
Fast scaling is available: The host has sufficient resources. The instance scales up in-place with no data migration, no connection interruption, and shorter completion time.
-
Fast scaling is not available: The host lacks sufficient resources. The instance is migrated to another host, which involves data migration and a connection interruption of about 30 seconds during switchover.
Prerequisites
-
The RDS for MySQL instance runs 5.7 or later. You can view or upgrade the major version in Change Configuration.
-
The instance uses local disk storage. Cloud disks are not supported.
-
The instance is in the Running state.
-
The scaling operation is one of the following:
-
Upgrade the instance type (for example, from 4-core 16 GB to 8-core 32 GB).
-
Increase or decrease the storage space.
-
Usage notes
-
Downgrade is not supported. Fast scaling applies only to upgrading an instance type or changing storage space. Downgrades use standard scaling, which may involve data migration and a connection interruption.
-
Not enabled by default. On the Change Configuration page, click Instant Upgrade Evaluation to check whether the host has sufficient resources. Otherwise, standard scaling is used.
-
Evaluation results are time-sensitive. After a successful evaluation, the system reserves resources for a limited time (about 5 minutes; check the console for the exact duration). You must place the order and complete payment within this window. If the reservation expires, you must re-evaluate.
-
Ensure your application supports automatic reconnection. If fast scaling is unavailable, the process involves data migration and a connection interruption of about 30 seconds.
Procedure
-
Log on to the ApsaraDB RDS console, click Instances in the left navigation pane, select a region in the top navigation bar, and click the target instance ID.
-
In the Basic Information section of the instance details page, click Change Configurations.
-
For subscription instances, in the Change Configurations dialog box, select Upgrade and click Next.
-
For pay-as-you-go instances, the configuration change page opens directly.
-
-
On the configuration change page, select the target instance type or storage space.
-
In the Instant Upgrade Evaluation and Configurations section, click Instant Upgrade Evaluation.
The system checks whether the current host has sufficient resources. One of the following results is returned:
-
Fast scaling is available: A blue message appears, for example: "Local resources are sufficient. To use fast scaling and avoid cross-host data migration, place your order and complete payment before YYYY.MM.DD HH:MM (UTC+8)." Complete payment before the stated deadline.
-
Fast scaling is not available: A yellow message indicates that the change involves data migration and a connection interruption of about 30 seconds. You can proceed with standard scaling or try again later.
-
-
Confirm the scaling settings, select the Terms of Service, and click Pay Now.
ImportantIf the evaluation shows fast scaling is available, complete payment before the deadline. Otherwise, the reserved resources may be released and the operation may revert to standard scaling with data migration and a connection interruption.
FAQ
Q: What should I do if the evaluation indicates that fast scaling is not available?
A: The host has insufficient idle resources for the target instance type. You can:
-
Proceed with standard scaling, which involves data migration and a connection interruption of about 30 seconds. Perform this during off-peak hours.
-
Re-evaluate later. Host resource availability may change over time.
Q: What happens if I do not complete payment before the reservation expires after fast scaling is available?
A: The reserved resources may be released, and the operation may revert to standard scaling with data migration and a connection interruption. Complete payment as soon as possible after a successful evaluation.
Q: Is there a cost difference between fast scaling and standard scaling?
A: No. Both methods cost the same. Billing is based on the target instance type and storage space. Fast scaling optimizes the process without incurring additional fees.