Kubernetes (K8s) 部署资源的完整配置OceanBase
- Kubernetes Deployment 配置（oceanbase-deployment.yaml）
# oceanbase-deployment.yaml
apiVersion: apps/v1
kind: Deployment
metadata:name: oceanbase-deployment
spec:replicas: 1selector:matchLabels:app: oceanbasetemplate:metadata:labels:app: oceanbasespec:nodeSelector: # 确保调度到有PV的节点kubernetes.io/hostname: node1 # 如果是固定服务器磁盘需要固定节点containers:- name: oceanbaseimage: registry.cn-hangzhou.aliyuncs.com/qiluo-images/oceanbase-ce:latestenv:- name: OB_MEMORY_LIMITvalue: "12G"- name: OB_SYSTEM_MEMORYvalue: "3G"- name: OB_CLUSTER_NAMEvalue: "OB_PROD_CLUSTER"- name: OB_ROOT_PASSWORDvalue: "6LuEpJaMudHd5yRc"volumeMounts:- mountPath: /root/ob/dataname: ob-data- mountPath: /root/ob/clogname: ob-clog- mountPath: /etc/oceanbasename: ob-configvolumes:- name: ob-datapersistentVolumeClaim:claimName: ob-data-pvc- name: ob-clogpersistentVolumeClaim:claimName: ob-clog-pvc- name: ob-configpersistentVolumeClaim:claimName: ob-config-pvc
- Kubernetes Service 配置（oceanbase-service.yaml）
apiVersion: v1
kind: Service
metadata:name: oceanbase-service
spec:selector:app: oceanbaseports:- name: sqlport: 2881targetPort: 2881nodePort: 32681 # NodePort方式暴露- name: metricsport: 2882targetPort: 2882- name: rpcport: 3881targetPort: 3881type: NodePort # 生产环境建议使用LoadBalancer
- PersistentVolumeClaim （oceanbase-pvc.yaml）
# oceanbase-pvc.yaml
apiVersion: v1
kind: PersistentVolumeClaim
metadata:name: ob-data-pvc
spec:accessModes:- ReadWriteOnceresources:requests:storage: 100GistorageClassName: local-storage
---
apiVersion: v1
kind: PersistentVolumeClaim
metadata:name: ob-clog-pvc
spec:accessModes:- ReadWriteOnceresources:requests:storage: 200GistorageClassName: local-storage
---
apiVersion: v1
kind: PersistentVolumeClaim
metadata:name: ob-config-pvc
spec:accessModes:- ReadWriteOnceresources:requests:storage: 20GistorageClassName: local-storage
4.手动创建 PV（适合无 StorageClass 的环境）oceanbase-pv.yaml
# oceanbase-pv.yaml
apiVersion: v1
kind: PersistentVolume
metadata:name: ob-data-pv
spec:capacity:storage: 100GivolumeMode: FilesystemaccessModes:- ReadWriteOncepersistentVolumeReclaimPolicy: RetainstorageClassName: local-storagelocal:path: /data/oceanbase/datanodeAffinity:required:nodeSelectorTerms:- matchExpressions:- key: kubernetes.io/hostnameoperator: Invalues:- node1 # 替换为实际节点名 # 如果是固定服务器磁盘需要固定节点
---
apiVersion: v1
kind: PersistentVolume
metadata:name: ob-clog-pv
spec:capacity:storage: 200GivolumeMode: FilesystemaccessModes:- ReadWriteOncepersistentVolumeReclaimPolicy: RetainstorageClassName: local-storagelocal:path: /data/oceanbase/clognodeAffinity:required:nodeSelectorTerms:- matchExpressions:- key: kubernetes.io/hostnameoperator: Invalues:- node1 # 如果是固定服务器磁盘需要固定节点
---
apiVersion: v1
kind: PersistentVolume
metadata:name: ob-config-pv
spec:capacity:storage: 20GivolumeMode: FilesystemaccessModes:- ReadWriteOncepersistentVolumeReclaimPolicy: RetainstorageClassName: local-storagelocal:path: /data/oceanbase/confignodeAffinity:required:nodeSelectorTerms:- matchExpressions:- key: kubernetes.io/hostnameoperator: Invalues:- node1 # 如果是固定服务器磁盘需要固定节点
- 部署步骤
mkdir /data/oceanbase/{data,clog,config} -p
chmod 777 /data/oceanbase/{data,clog,config} # 1. 创建PV和PVC
kubectl apply -f oceanbase-pv.yaml
kubectl apply -f oceanbase-pvc.yaml# 2. 验证存储
kubectl get pv
kubectl get pvc # 应全部显示Bound状态# 3. 部署OceanBase
kubectl apply -f oceanbase-deployment.yaml
kubectl apply -f oceanbase-service.yaml# 4. 检查状态
kubectl get pods -w
或者
kubectl get pods -n dev -o wide -A
查看日志
kubectl logs -f <pod-name> -c oceanbase --tail=1005.进入容器执行
docker exec -it <pod-name> /bin/bash
obclient -h127.0.0.1 -P2881 -uroot@sys -A
修改用户名
ALTER USER root IDENTIFIED BY '6LuEpJaMudHd5yRc';
此时此刻就连接成功啦
获取访问IP（根据实际环境调整）
kubectl get svc oceanbase
>**关键调整说明
存储配置：
必须使用 持久化存储（PV/PVC），推荐 SSD/NVMe
多节点部署时需要共享存储（如 Ceph RBD、NFS）
高可用改进：# 在Deployment中增加
```yaml
spec:affinity:podAntiAffinity:requiredDuringSchedulingIgnoredDuringExecution:- labelSelector:matchLabels:app: oceanbasetopologyKey: "kubernetes.io/hostname"
生产环境建议：
使用 StatefulSet 替代 Deployment（稳定网络标识）
通过 ConfigMap 管理配置文件：
volumes:
- name: ob-configconfigMap:name: oceanbase-config
网络优化：
若需跨节点通信，使用 Headless Service
监控端口建议通过 Ingress 或 ServiceMesh 暴露
故障排查命令
# 检查资源限制
kubectl describe pod <pod-name> | grep -A 10 "Limits"
# 进入容器调试
kubectl exec -it <pod-name> -- bash
obd cluster list
# 查看事件日志
kubectl get events --sort-by=.metadata.creationTimestamp
请根据您的实际 Kubernetes 环境（如存储类名称、网络插件等）调整配置。生产部署前务必在测试环境验证。