English | 中文

TapData One-Deploy is the official all-scenario automated deployment tool for TapData, covering the full delivery pipeline from single-node development to enterprise-grade cloud-native production.

Build Once, Run Anywhere

# Download
curl -L https://resource.tapdata.net/deploy/tap-deploy -o tap-deploy && chmod +x tap-deploy

# Init → Plan → Apply
./tap-deploy init      # Interactive configuration wizard
./tap-deploy plan      # Preview deployment
./tap-deploy apply     # Execute deployment

After deploying in Lite mode, open http://localhost:13030 in your browser to access the console (default account admin@admin.com, password admin).

Command reference:

tap-deploy uses a template rendering mechanism for configuration management:

1. init phase: Collect parameters interactively → Save to .tap-deploy.env → Render templates into final config files
2. apply phase: Read configuration → Execute deployment (Docker Compose / Terraform + Helm / Helm)

init interactive wizard
  ├─ Parameters → .tap-deploy.env (config store)
  ├─ Templates + .tap-deploy.env → docker-compose.yaml / values.yaml / application.yml / terraform.tfvars
  └─ Terraform init (Cloud mode)

apply deployment execution
  ├─ Lite:    docker compose up -d
  ├─ Cloud:   terraform apply → push images → helm install
  └─ On-Prem: helm install (optional offline image loading)

Option 1: Re-run the init wizard

The simplest way to change any parameter. The wizard regenerates all config files.

Option 2: Edit .tap-deploy.env directly

After manually editing, re-render templates or edit the generated files directly. Suitable for batch changes or automation.

Option 3: Edit generated files directly

Files like docker-compose.yaml, values.yaml, and application.yml are standard-format files that can be edited for deep customization. Use this for advanced parameters not covered by the init wizard (e.g., resource quotas, Ingress annotations, MongoDB storage engine settings).

Note: Re-running init overwrites generated config files and manual edits will be lost. Consider recording customizations in .tap-deploy.env or managing them through separate Helm values override files.

Single-node deployment with 5-minute quick start, ideal for development, testing, and POC.

Prerequisites: Docker 20.10+ with Docker Compose plugin, current user has Docker execution permissions, public internet access.

The following are defined in .tap-deploy.env but not prompted by the init wizard. Edit manually to adjust:

Edit docker-compose.yaml directly to adjust:

• Container resource limits (default: 4 CPU / 12G memory)
• Volume mount paths
• Health check policies
• Environment variables (FRONTEND_WORKER_COUNT, API_WORKER_COUNT, etc.)

Automated cloud infrastructure provisioning + Kubernetes deployment, designed for production.

Prerequisites: Terraform 1.7+, Helm 3.14+, kubectl, Docker, cloud provider AK/SK permissions.

terraform apply → Create VPC / CCE(EKS) / SWR(ECR) / NAT
       ↓
Push images to cloud registry (SWR / ECR)
       ↓
Install MongoDB Community Operator
       ↓
helm install → Deploy TapData (MongoDB + server + engine + apiserver)

Common Parameters

Huawei Cloud Specific

AWS Specific

In Cloud mode, registry information is automatically obtained after terraform apply — no manual configuration needed:

• Huawei Cloud: SWR internal pull URL + external push URL + temporary AK/SK authentication
• AWS: ECR internal pull URL + external push URL + ECR authentication

Automatically written to .tap-deploy.env and rendered into values.yaml during the apply phase.

Each cloud provider uses a dedicated Ingress config file, automatically merged during apply:

To customize Ingress (domain, TLS certificate, bandwidth, etc.), edit the corresponding values-ingress-*.yaml file.

Edit values.yaml directly to adjust:

• Replica counts (tapdata-server.replicaCount, etc., default: 2 each)
• Resource quotas (CPU/memory limits & requests)
• MongoDB replica set members, storage class, storage size
• Service type (ClusterIP / NodePort)
• API Server enable/disable

Edit terraform.tfvars to adjust:

• VPC CIDR
• Key pair name
• Registry organization/repository name

# Huawei Cloud
kubectl describe ingress -n tapdata -l app.kubernetes.io/name=tapdata | grep 'kubernetes.io/elb.ip'

# AWS
kubectl get ingress -n tapdata

For customer data centers and air-gapped environments, deployed on existing Kubernetes clusters.

Prerequisites: Helm 3.14+, kubectl, access to a K8s cluster. Offline mode requires Docker and a pre-built image bundle.

When the K8s cluster can access the public internet (or a private registry):

./tap-deploy init    # Select On-Prem → Non-offline
./tap-deploy apply

Build image bundle on an online machine:

./tap-deploy bundle --output images-bundle.tar.gz

Deploy in the offline environment:

# Load images
docker load -i images-bundle.tar.gz

# Push to internal registry
docker tag tapdata8/tapdata:latest <harbor>/tapdata:latest
docker push <harbor>/tapdata:latest

# Deploy
./tap-deploy init    # Select On-Prem → Offline, enter internal registry address
./tap-deploy apply

Same as Cloud mode — edit values.yaml directly to adjust replica counts, resource quotas, MongoDB parameters, etc.

• Huawei Cloud: Ultra-High I/O EVS (SSD)
• AWS: gp3 / io2 EBS
• Self-hosted: SSD recommended to reduce CDC latency

• MongoDB 3-node replica set with automatic failover
• Multi-replica application + load balancing
• Zero-downtime rolling updates, 30s graceful shutdown

Docker 20.10+ · Kubernetes 1.24+ · Helm 3.14+ · Terraform 1.7+ · MongoDB 6.0+ · AMD64 / ARM64

• init requires downloading config templates; apply requires pulling images. Prepare offline bundles in advance for air-gapped environments
• Use ultra-high I/O storage in production to reduce CDC latency
• All Connectors are ARM-compiled and verified
• macOS: ensure terminal has disk access permissions
• 30-day free trial License included by default; contact TapData team for extensions

GitHub · Issues