Enterprise Edition and Basic Edition consolidate the capabilities of Data Warehouse Edition (reserved mode and elastic mode) and Data Lakehouse Edition into a single, unified offering. Both editions use a storage-compute coupled architecture with reserved resources for consistent query performance, and support on-demand scaling when workloads change.
Edition changes are irreversible. After the change is complete, you cannot restore the cluster to its original edition.
Prerequisites
Before you begin, make sure that:
Your Alibaba Cloud account has no unpaid renewal orders
The cluster is in the Running state
The cluster's minor version is 3.1.10.6 or later — to view or update the minor version, log on to the AnalyticDB for MySQL console, open the Cluster information page, and go to the Configuration information section. For instructions, see Update the minor version of a cluster
(Data Lakehouse Edition clusters only) The cluster has no resource groups whose configurations are being changed, or resource groups that are being created, modified, started, or stopped
(Data Lakehouse Edition clusters only) The cluster has no Interactive Resource Group scaling plans or EIU scaling plans
Potential impacts
Review the impacts for your source edition before starting the change.
| Impact | Data Lakehouse Edition | Data Warehouse Edition |
|---|---|---|
| Reversibility | Cannot revert to Data Lakehouse Edition. | Cannot revert to Data Warehouse Edition. |
| Specification restriction | No restriction. | Clusters with S8 or S16 specifications cannot be changed to Enterprise Edition or Basic Edition. |
| Read/write availability | Control operations and INSERT OVERWRITE are blocked. Other read/write operations continue normally. | The source cluster becomes read-only for 5–10 minutes during switchover. |
| `SUBMIT JOB` statements | Blocked during scaling. | Blocked if continuously executed during the edition change — this stalls the switchover. |
| Transient connections | May occur during the edition change. Perform the change during off-peak hours, or configure your application to automatically reconnect. | May occur near the end of the edition change. Perform the change during off-peak hours, or configure your application to automatically reconnect. |
| Asynchronous job statuses | Not applicable. | Asynchronous job statuses differ between editions. Map job statuses after the edition change is complete. See Asynchronously submit an import job for status mappings. |
| Backup sets | Not applicable. | Existing backup sets immediately become invalid and cannot be used for cluster cloning. The backup cycle configuration remains unchanged. |
| New features after change | — | After changing to Enterprise Edition, you can create job resource groups and XUANWU_V2 tables billed on a pay-as-you-go basis. For billing details, see Editions and Pricing. |
Scaling behavior during a Data Lakehouse Edition change
If you modify the reserved resource node configurations during the edition change, scaling occurs on the cluster:
Data in the cluster is migrated and redistributed. Migration time is proportional to data volume.
Downgrading specifications may require up to dozens of hours for data migration. Plan accordingly if the cluster holds a large amount of data.
When scaling down, make sure each reserved resource node does not hold an excessive number of shards. Select a node count that keeps the total reserved resources comparable to the original. For more information, see error messages.
Migration process for a Data Warehouse Edition change
The edition change follows these steps, in order:
New cluster resources are provisioned.
Full data migration runs from the source cluster to the new cluster.
Incremental data migration keeps the two clusters in sync.
Workloads switch over from the source cluster to the new cluster. The source cluster enters read-only mode for 5–10 minutes during this step.
The edition change is complete. The cluster status changes to Running.
Migration time varies based on cluster specifications and data volume.
To control when the switchover occurs, connect to the source cluster and run the following statement before starting the edition change. The example below sets the switchover window to 23:00–23:30.
SET adb_config REPLICATION_SWITCH_TIME_RANGE=`23:00, 23:30`;For instructions on connecting to the cluster, see Connect to an AnalyticDB for MySQL cluster.
Change the cluster edition
(Optional) Submit a ticket to confirm that your cluster meets the edition change requirements.Submit a ticket
Log on to the AnalyticDB for MySQL console. In the upper-left corner, select a region. In the left-side navigation pane, click Clusters.
Open the configuration change page for your cluster.
If the edition change results in a downgrade, find the cluster and choose More > Scale Down in the Actions column.
If the edition change results in an upgrade, find the cluster and click Scale Up in the Actions column.
AnalyticDB for MySQL determines which page to show based on cost: if the new cluster costs more than the original, the Upgrade/Downgrade page appears after you click Scale Up; if it costs less, the Downgrade page appears after you choose More > Scale Down. Estimate costs in advance using Pricing.
On the Upgrade/Downgrade or Downgrade page, configure the following parameters.
If the reserved resources selected during the edition change don't fully meet your requirements, scale the cluster after the edition change is complete. See Scale an Enterprise Edition or Basic Edition cluster.
Parameter Description Configuration change type Select Cross-edition Configuration Change. Edition Select Enterprise Edition (Multi-replica) or Basic Edition (Single-replica). For details, see Editions. Data Lakehouse Edition clusters can only be changed to Enterprise Edition. Single-node specifications of reserved resources AnalyticDB for MySQL automatically recommends specifications aligned with the source cluster's performance. Adjust manually if needed. When scaling up, the recommended CPU and memory are not lower than the source cluster. When scaling down, the recommendation targets the best performance without exceeding the original cost. If reserved resources are set to 0 ACUs, only external tables are available; set reserved resources above 0 ACUs to create internal tables. If you set reserved resources to 0 ACUs, you can purchase reserved resources later by scaling up the cluster. For more information, see Scale an Enterprise Edition or Basic Edition cluster. For specification mappings between editions, see Create a cluster. Reserved resource nodes AnalyticDB for MySQL automatically recommends a node count aligned with the source cluster. Adjust manually if needed. Enterprise Edition clusters have 3 reserved resource nodes; Basic Edition clusters have 1. When scaling up, CPU and memory are not lower than the source cluster. When scaling down, the recommendation targets the best performance without exceeding the original cost. ESSD performance level The performance level of the Enterprise SSD (ESSD). Valid values: PL1, PL2, PL3. Read and accept the Terms of Service, then click Buy Now.
After the edition change is complete, the cluster status changes to Running.
What's next
If you changed from Data Warehouse Edition and your business uses asynchronous jobs, review the job status mappings in Asynchronously submit an import job.
To adjust cluster resources after the edition change, see Scale an Enterprise Edition or Basic Edition cluster.
To understand billing differences between editions, see Pricing.