This document describes how to create a TDSQL Boundless instance via the console.
Prerequisites
A Tencent Cloud account has been registered with identity verification completed.
Operation Steps
2. On the purchase page, select relevant configurations.
Basic Configuration
Engine: Select TDSQL Boundless.
Billing Mode: Yearly/monthly subscription and pay-as-you-go are supported. For stable long-term business demands, it is recommended to choose yearly/monthly subscription. For scenarios with significant instantaneous fluctuations in business volume, it is recommended to choose pay-as-you-go.
Region: Select the region where you need to deploy TDSQL Boundless. It is recommended to choose the same region as your CVM. Cloud products in different regions cannot communicate over the private network and the region cannot be changed after purchase.
Network: Supports VPC environments. You can select the network and subnet for the instance. If existing networks are unsuitable, you may Create VPC or Create Subnet. Note:
Subnet is a logical network space within a VPC. You can create subnets in different AZs under the same VPC. Subnets in different AZs under the same VPC can interconnect via the private network by default.
After the network is selected, the subnet IP address defaults to that of the same AZ as the selected instance. You may also manually choose subnet IP addresses within other AZs in the same region. Business connections follow proximity-based routing logic without adding network latency.
It is recommended to choose the same VPC as your CVM in the same region; otherwise, the CVM and database will be unable to connect via the private network. The default setting is Default-VPC (default).
Specs Configuration
Compatibility Mode: Select MySQL-compatible mode and HBase-compatible mode based on business needs.
Instance Type: are divided into two types: cluster edition and basic edition.
Cluster Edition: Composed of multiple nodes, it provides high-performance and highly available database services in the form of a multi-replica Raft cluster, suitable for enterprise production environments.
Basic Edition: Composed of a single node, it provides complete database features at a relatively low cost but does not include high availability, suitable for individual users.
Warning:
Basic Edition Basic Edition Enterprise users must not use basic edition instances in production environments, as this may expose business operations to risks of service interruptions due to potential database downtime.
Note:
Basic Edition instances can be upgraded to Cluster Edition via the console after creation; however, Cluster Edition instances cannot be downgraded to Basic Edition after creation.
Deployment Mode: Users can choose the number of AZs to deploy. Single-AZ: All resources are in the same AZ, providing the lowest network latency; Two AZs: Applications run across two AZs, which can withstand a single AZ failure; Three AZs: Resources are distributed across three AZs, allowing the service to continue even if one AZ fails and reducing recovery time.
AZ: For instances deployed in multiple AZs, users can set a primary AZ. After the primary AZ is set, the weight of TDSQL Boundless nodes in that AZ will be increased to 99%, meaning that 99% of requests will be routed to that node for processing. Users can choose not to set a primary AZ, or they can cancel / adjust the primary AZ configuration via the console.
After the primary AZ is set, the configuration adjustment feature becomes unavailable unless the user temporarily cancels the primary AZ setting. This is because once the primary AZ is configured, TDSQL cannot proactively switch the Leader to different AZs as needed to ensure business continuity during configuration changes. If users need to perform this operation, they must first acknowledge the potential impact on business performance from canceling the primary AZ setting, then remove the setting before proceeding with configuration adjustments.
Number of Fully Functional Replicas: Select the number of replicas capable of read and write operations that participate in voting for data consistency.
Note:
Setting the weight of TDSQL Boundless nodes in the primary AZ to 100% is currently not supported.
Performance Specification: defaults to Enhanced, equipped with new-generation CPU cores, providing powerful computing performance and ample computing resources, featuring excellent data read/write capabilities, designed specifically for large-scale web applications, high-concurrency services, and other performance-sensitive businesses.
Storage Class: includes Cloud Disk and Local Disk.
Cloud Disk: Cloud Disk offers higher availability and data durability, supporting larger single-node capacity and rapid backup.
Local Disk: Local Disk offers better throughput and IOPS, but has a longer fault recovery time compared with Cloud Disk.
Hybrid Node Disk Capacity: Select the storage space required for each node.
Note:
To ensure stable operation of the database, the system has a built-in storage space protection mechanism. It is crucial to closely monitor storage space usage. The rules of this mechanism are as follows:
Read-only mode trigger: When the storage space utilization of any node exceeds the threshold set by the tdstore_enter_readonly_threshold parameter (default: 99%), that node will automatically enter read-only mode. At this time, the node can query data normally, but only data cleanup operations (such as DROP/TRUNCATE) are allowed; other write operations (such as INSERT/UPDATE/DELETE) will be rejected. Data read and write operations on other nodes that have not triggered the threshold are not affected. In a 3-replica scenario, if two or more replicas enter read-only mode, the entire instance will enter read-only mode.
This utilization threshold is calculated based on the data storage capacity of a single node. The system reserves 8% of space for its own operation. Therefore, from the user's perspective, when "Available Space" is not yet fully exhausted, the protection mechanism will be triggered as long as "Total Usage" reaches the threshold.
Example: Assume the total data storage capacity is 500GB, and the system reserves 8% of the space (40GB). Then:
User's actual available space: 500GB - 40GB = 460GB
Threshold for triggering read-only mode: 500GB × 99% = 495GB
At this point, the node's user data storage has used 495GB - 40GB = 455GB, and the node will automatically switch to read-only mode.
Recovery from read-only mode: When you reduce space utilization below the threshold set by the tdstore_leave_readonly_threshold parameter (default: 95%) through methods such as scaling out or data cleanup, the node will automatically recover to read-write mode.
Example: Assume the total data storage capacity is 500GB, and the system reserves 8% of the space (40GB). When user space usage drops to 500GB × 95% - 40GB = 435GB, the node will automatically recover to read-write mode.
Hybrid Nodes: Select the required number of peer nodes.
Initialization configurations
Character Set: Applicable to MySQL compatibility mode. Supports UTF8, LATIN1, GBK, UTF8MB4, and GB18030 character sets. The default character set encoding format is UTF8MB4.
Table Name Case Sensitivity: Applicable to MySQL compatibility mode. Specifies whether table names are case-sensitive. Note that in MySQL compatibility mode, the case sensitivity setting for table names cannot be changed after configuration. Please configure with caution.
Custom Port: The access port for the database, which defaults to 3306.
Others
Tag: Add tags to instances for categorized management of instance resources.
Instance Name: You can choose to name it after creation or immediately.
3. After the above configurations are confirmed, read and select the Terms of Service, then click Buy Now. After the purchase is completed, return to the Instance List. The instance will display as "Delivering" (approximately 3-5 minutes, please wait patiently). Once the instance status changes to Running, normal operations can be performed.