A DTS dedicated cluster gives your Data Transmission Service (DTS) tasks a private pool of CPU and memory, isolated from other users. Use a dedicated cluster when your workload is large enough that resource contention with other tenants would affect task stability, or when a single task needs the full resources of a node.
Dedicated cluster vs. public cluster
| Dimension | Dedicated cluster | Public cluster |
|---|---|---|
| Resource isolation | Fully isolated from other users | Shared with other users |
| Billing | Subscription (cluster resources only); DTS instances inside the cluster are free | Pay per DTS instance |
| Supported task types | Data migration, data synchronization, change tracking | Data migration, data synchronization, change tracking |
| Max clusters per region | 5 | N/A |
| Resource allocation | Allocate the full cluster to a single task | Per-instance limits apply |
Prerequisites
Before you begin, ensure that you have (required if you want to create a DTS dedicated cluster as a RAM user):
The AliyunDTSDefaultRole system role assigned to your Alibaba Cloud account
The AliyunDTSFullAccess permission granted to the Resource Access Management (RAM) user
For setup instructions, see Authorize DTS to access Alibaba Cloud resources and Use a system policy to authorize a RAM user to manage DTS instances.
Limitations
Billing: Subscription only. You pay for cluster resources upfront; DTS instances created inside the cluster have no additional cost.
Supported task types: Data migration, data synchronization, and change tracking.
Supported regions: China (Hangzhou), China (Shanghai), China (Beijing), China (Zhangjiakou), China (Hohhot), China (Ulanqab), China (Shenzhen), China (Guangzhou), China (Chengdu), China (Hong Kong), Singapore, Japan (Tokyo), US (Virginia), and Germany (Frankfurt).
Data synchronization: The source database, destination database, and cluster must all be in the same region.
Cluster limit per region: 5. To increase this limit, contact Alibaba Cloud technical support.
Scaling: Node count and disk size can only increase, not decrease.
Resource threshold: If CPU, memory, or disk usage exceeds 80%, scale up immediately. Task or switchover failures caused by insufficient resources are not covered by the DTS Service Level Agreement (SLA).
Disk reservation: After creation, the system reserves 5% of disk space for defragmentation.
Billing
Choose your cluster size
Each node has a fixed specification of 15 DTS units (DUs). A DU represents the minimum hardware resources for one DTS task, so a single non-overcommitted node supports up to 15 concurrent tasks with isolated resources.
Use these guidelines to size your cluster:
| Factor | Guidance |
|---|---|
| Number of concurrent tasks | Divide your expected concurrent tasks by 15 to get the minimum node count. Add at least one extra node for failover headroom. |
| Volume of incremental data | High-throughput change data capture (CDC) workloads consume more memory per task. If tasks process large volumes of row changes per second, allocate more nodes rather than enabling overcommitment. |
| Task scheduling pattern | If tasks run in non-overlapping time windows, overcommitment can reduce cost. If tasks run concurrently at peak hours, use dedicated (non-overcommitted) capacity. |
| Disk per node | Estimate based on the total data volume that DTS buffers during sync. Valid range: 1,024–30,720 GB per node, in increments of 512 GB. |
Overcommitment: Setting the Overcommit Ratio to 200% doubles the logical DUs on each node from 15 to 30, letting you configure up to 30 tasks per node. Because the underlying hardware resources remain the same, each task gets roughly half the resources compared to non-overcommitted mode. Enable overcommitment only when incremental data volume is low or tasks run at different times. For details, see Set an overcommit ratio for a DTS dedicated cluster.
Disaster recovery: We recommend that you create two or more DTS dedicated clusters to implement disaster recovery. Within each cluster, create at least two nodes so the cluster can absorb a node failure. Keep available resources (CPU, memory, disk, and DUs) above 1/N of the cluster total, where N is the number of nodes. For example, with three nodes, keep more than one-third of resources free at all times. To learn how to migrate tasks between clusters for maintenance or recovery, see Migrate a DTS instance between a dedicated cluster and a public cluster.
Create a dedicated cluster
Go to the Dedicated Cluster page.
In the top navigation bar, select the target region from the drop-down list to the right of Workbench.
Click Create Dedicated Cluster below Cluster Task.
Configure the parameters:
Parameter Description Billing Method Select Subscription. Cluster Type of DTS Instance Select Dedicated Cluster. Region Select the region for the cluster. The region cannot be changed after creation. Dedicated Cluster Fixed to Data Migration, Data Synchronization, and Change Tracking. Specification of Dedicated Cluster Nodes Fixed to 15 DTS Units per node. See Choose your cluster size for capacity planning. Storage Space of Dedicated Cluster Node Disk size per node, in GB. Valid range: 1,024–30,720 GB, in increments of 512 GB. To increase disk space after creation, see Increase the storage space of nodes in a DTS dedicated cluster. Destination Region Defaults to the same value as Region. Cannot be changed. Resource Group Optional. For details, see What is Resource Management? Order Time Subscription duration. Valid values: 1–9 months, 1 year, 2 years, or 3 years. To enable auto-renewal, select Auto-renewal and make sure your account has a sufficient balance before the renewal date. Nodes in a Dedicated Cluster Number of nodes. Default: 2. Valid range: 2–100. To add nodes after creation, see Scale out a DTS dedicated cluster. Click Buy Now.
On the Confirm Order page, review the configuration, read and accept the Terms of Service, then click Pay.
On the Purchase page, complete the payment.
You can also click Add to Cart to purchase the cluster together with other services. See Pay for services in the cart.
Verify the cluster
After payment, verify that the cluster was created:
Click Console to go to the DTS console.
In the left navigation pane, click Dedicated Cluster.
In the top navigation bar, select the region you chose during creation.
Check whether the DTS dedicated cluster that you just created is in the cluster list.
If the DTS dedicated cluster is displayed, the cluster is created. If the DTS dedicated cluster is in the Running state, the cluster is created and initialized.
If the DTS dedicated cluster is not displayed, wait 1–5 minutes and refresh the page.
Troubleshooting
If the cluster still does not appear after refreshing, check the following:
| Possible cause | Solution |
|---|---|
| Wrong region selected | In the top navigation bar, select the region you specified when creating the cluster. |
| Cluster creation failed due to insufficient resources | Creation may fail if the region has insufficient capacity. In this case, the payment is refunded. Go to the Expenses and Costs console, then choose Cost Management > Orders to check the Order Management page for the refund status. |
If neither issue applies, contact Alibaba Cloud technical support.
Pay for services in the cart
After configuring a subscription cluster, click Add to Cart to purchase it together with other services:
On the right side of the page, click the Cart icon.
In the Cart panel, select the services to purchase and click Proceed to Checkout.
On the Cart page, confirm the correct services are selected, then click Buy Now.
On the Confirm Order page, read and accept the Terms of Service, then click Pay.
On the Purchase page, complete the payment.