TDengine Multi-Tier Storage Configuration Guide

Juno Qiu

July 5, 2026 /

Multi-tier storage overview

Multi-tier storage places recent hot data on high-speed media and cold data on low-cost media.

Core benefits include lower storage costs, higher write performance, easier maintenance and management, and SQL transparency.

How it works

Data flows through three tiers: hot data resides on SSD or high-speed storage for frequent access, warm data moves to HDD for medium-frequency access, and cold data is archived to low-cost storage.

Configuration

Creating storage tiers

Enable multi-tier storage on the database:

ALTER DATABASE demo SET MULTI_TIER 1;

Configure mount points for each tier:

ALTER MOUNT 'tier0' PATH '/ssd/data';
ALTER MOUNT 'tier1' PATH '/hdd/data';

Configuration parameters

The following parameters control multi-tier storage behavior:

  • multiTierEnable yes
  • tier0Path /ssd/data
  • tier1Path /hdd/data

Performance characteristics

Write performance

Tier 0 (SSD) achieves 300 million measurements per second with a parallel write speed of 2 GB/s.

Query transparency

Regardless of where data physically resides, a single SQL query retrieves all matching records. For example:

SELECT * FROM demo.meters WHERE ts > '2023-01-01' AND ts < '2024-01-01';

Cost reduction

Storage cost comparison

The following table compares storage options across cost and performance dimensions:

Storage tierCost (per TB/month)Write performanceAccess frequencyUse case
SSD100 yuanHighestHighRecent hot data (0 to 7 days)
HDD20 yuanMediumMediumWarm data (7 to 30 days)
Low-cost storage5 yuanLowerLowCold archived data (30+ days)

Tiering strategies

A recommended tiering policy:

  • Last 7 days: SSD
  • 7 to 30 days: HDD
  • 30+ days: low-cost storage

Maintenance

Automatic data migration

Data migration between tiers runs automatically in the background. No manual intervention is required, and the process is transparent to applications.

Flexible expansion

Adding storage nodes is straightforward. The system supports online expansion without disrupting running workloads.

Best practices

Configuration

CREATE DATABASE demo (KEEP 365, CACHE 256);
ALTER DATABASE demo DATA_RETENTION_POLICY 'tiered';

Monitoring

SELECT * FROM INFORMATION_SCHEMA.INS_MOUNTS;

Summary

Multi-tier storage reduces storage costs, automates data distribution across tiers, provides transparent queries, and supports online expansion.