Associated Projects:
- https://github.com/seata/seata
- https://github.com/seata/seata-samples/tree/docker/springboot-dubbo-fescar
- https://github.com/seata/seata-docker
To deploy Seata Server using the Operator method, follow these steps:
-
Clone this repository:
git clone https://github.com/apache/incubator-seata-k8s.git
-
Deploy Controller, CRD, RBAC, and other resources to the Kubernetes cluster:
make deploy kubectl get deployment -n seata-k8s-controller-manager # check if exists -
You can now deploy your CR to the cluster. An example can be found here seata-server-cluster.yaml:
apiVersion: operator.seata.apache.org/v1alpha1 kind: SeataServer metadata: name: seata-server namespace: default spec: serviceName: seata-server-cluster replicas: 3 image: seataio/seata-server:latest store: resources: requests: storage: 5Gi
For the example above, if everything is correct, the controller will deploy 3 StatefulSet resources and a Headless Service to the cluster. You can access the Seata Server cluster in the cluster through
seata-server-0.seata-server-cluster.default.svc.
For CRD details, you can visit operator.seata.apache.org_seataservers.yaml. Here are some important configurations:
-
serviceName: Used to define the name of the Headless Service deployed by the controller. This will affect how you access the server cluster. In the example above, you can access the Seata Server cluster throughseata-server-0.seata-server-cluster.default.svc. -
replicas: Defines the number of Seata Server replicas. Adjusting this field achieves scaling without the need for additional HTTP requests to change the Seata raft cluster list. -
image: Defines the Seata Server image name. -
ports: Three ports need to be set under theportsproperty:consolePort,servicePort, andraftPort, with default values of 7091, 8091, and 9091, respectively. -
resources: Used to define container resource requirements. -
store.resources: Used to define mounted storage resource requirements. -
env: Environment variables passed to the container. You can use this field to define Seata Server configuration. For example:apiVersion: operator.seata.apache.org/v1alpha1 kind: SeataServer metadata: name: seata-server namespace: default spec: image: seataio/seata-server:latest store: resources: requests: storage: 5Gi env: - name: console.user.username value: seata - name: console.user.password valueFrom: secretKeyRef: name: seata key: password --- apiVersion: v1 kind: Secret metadata: name: seata type: Opaque data: password: seata
To debug this operator locally, we suggest you use a test k8s environment like minikube.
-
Method 1. Modify code and build the controller image:
Assume you are using minikube for testing,
eval $(minikube docker-env) make docker-build deploy
-
Method 2. Locally debug without building images
You need to use telepresence to proxy traffic to the k8s cluster, see telepresence tutorial to install its cli tool and traffic manager. After installing telepresence, you can connect to minikube by following commands:
telepresence connect # Check if traffic manager connected telepresence statusBy executing above commands, you can use in-cluster DNS resolution and proxy your requests to the cluster. And then you can use IDE to run or debug locally:
# Make sure generate proper resources first make manifests generate fmt vet go run . # Or you can use IDE to run locally instead
Due to certain reasons, Seata Docker images currently do not support external container calls. Therefore, the example projects should also be kept in link mode with the Seata image inside the container.
# Start Seata deployment (nacos,seata,mysql)
kubectl create -f deploy/seata-deploy.yaml
# Start Seata service (nacos,seata,mysql)
kubectl create -f deploy/seata-service.yaml
# Get a NodePort IP (kubectl get service)
# Modify the IP in examples/examples-deploy for DNS addressing
# Connect to MySQL and import table structure
# Start example deployment (samples-account,samples-storage)
kubectl create -f example/example-deploy.yaml
# Start example service (samples-account,samples-storage)
kubectl create -f example/example-service.yaml
# Start order deployment (samples-order)
kubectl create -f example/example-deploy.yaml
# Start order service (samples-order)
kubectl create -f example/example-service.yaml
# Start business deployment (samples-dubbo-business-call)
kubectl create -f example/business-deploy.yaml
# Start business deployment (samples-dubbo-service-call)
kubectl create -f example/business-service.yamlOpen the Nacos console in your browser [http://localhost:8848/nacos/] to check if all instances are registered successfully.
# Account service - Deduct amount
curl -H "Content-Type: application/json" -X POST --data "{\"id\":1,\"userId\":\"1\",\"amount\":100}" cluster-ip:8102/account/dec_account
# Storage service - Deduct stock
curl -H "Content-Type: application/json" -X POST --data "{\"commodityCode\":\"C201901140001\",\"count\":100}" cluster-ip:8100/storage/dec_storage
# Order service - Add order and deduct amount
curl -H "Content-Type: application/json" -X POST --data "{\"userId\":\"1\",\"commodityCode\":\"C201901140001\",\"orderCount\":10,\"orderAmount\":100}" cluster-ip:8101/order/create_order
# Business service - Client Seata version too low
curl -H "Content-Type: application/json" -X POST --data "{\"userId\":\"1\",\"commodityCode\":\"C201901140001\",\"count\":10,\"amount\":100}" cluster-ip:8104/business/dubbo/buy
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent