印刻数据库

外观

KingBaseES Docker/K8s部署 ​

随着容器技术和云原生架构的普及，将KingBaseES数据库部署到Docker和Kubernetes环境已成为越来越多企业的选择。本文将详细介绍KingBaseES在容器化环境中的部署方案。

Docker部署方案 ​

1. Docker镜像构建 ​

基础镜像选择 ​

推荐使用的基础镜像：

  • CentOS 7/8
  • RHEL 7/8
  • Ubuntu 20.04/22.04

Dockerfile编写 ​

dockerfile
# 基于CentOS 7构建KingBaseES V8 R7镜像
FROM centos:7

# 设置环境变量
ENV KINGBASE_VERSION=V8R7C7
ENV KINGBASE_HOME=/opt/kingbase/ES/V8R7C7
ENV PGDATA=/opt/kingbase/data
ENV PATH=$PATH:$KINGBASE_HOME/bin

# 安装依赖包
RUN yum -y update && \ 
    yum -y install epel-release && \ 
    yum -y install wget gcc gcc-c++ make perl readline-devel zlib-devel openssl-devel && \ 
    yum clean all

# 下载并安装KingBaseES
RUN wget -O /tmp/KingbaseES_V8R7C7_Lin64_install.iso http://your-repo-url/KingbaseES_V8R7C7_Lin64_install.iso && \ 
    mkdir -p /mnt/iso && \ 
    mount -o loop /tmp/KingbaseES_V8R7C7_Lin64_install.iso /mnt/iso && \ 
    cd /mnt/iso && \ 
    ./install.sh -i silent -DAPP_BASE_INSTALL_PATH=/opt/kingbase -DAPP_SHARE_DIR=/opt/kingbase/share -DDB_INIT=Y -DDB_PORT=54321 -DDB_USER=system -DDB_PASSWORD=Kingbase123 && \ 
    umount /mnt/iso && \ 
    rm -rf /tmp/KingbaseES_V8R7C7_Lin64_install.iso

# 配置KingBaseES
RUN sed -i 's/^listen_addresses.*/listen_addresses = "*"/' $PGDATA/kingbase.conf && \ 
    echo "host all all 0.0.0.0/0 trust" >> $PGDATA/pg_hba.conf

# 创建启动脚本
RUN echo '#!/bin/bash\n\nexec kdb5start -D $PGDATA -i' > /usr/local/bin/start-kingbase && \ 
    chmod +x /usr/local/bin/start-kingbase

# 暴露端口
EXPOSE 54321

# 设置启动命令
CMD ["start-kingbase"]

构建镜像 ​

bash
docker build -t kingbase:v8r7 .

2. 单容器部署 ​

基本部署 ​

bash
docker run -d \
  --name kingbase \
  -p 54321:54321 \
  -v kingbase-data:/opt/kingbase/data \
  kingbase:v8r7

挂载外部配置文件 ​

bash
docker run -d \
  --name kingbase \
  -p 54321:54321 \
  -v /host/path/kingbase.conf:/opt/kingbase/data/kingbase.conf \
  -v /host/path/pg_hba.conf:/opt/kingbase/data/pg_hba.conf \
  -v kingbase-data:/opt/kingbase/data \
  kingbase:v8r7

3. 主从复制部署 ​

主库部署 ​

bash
docker run -d \
  --name kingbase-master \
  -p 54321:54321 \
  -e REPLICATION_MODE=master \
  -e REPLICATION_USER=repl \
  -e REPLICATION_PASSWORD=repl123 \
  -v kingbase-master-data:/opt/kingbase/data \
  kingbase:v8r7

从库部署 ​

bash
docker run -d \
  --name kingbase-slave \
  -p 54322:54321 \
  -e REPLICATION_MODE=slave \
  -e REPLICATION_MASTER_HOST=kingbase-master \
  -e REPLICATION_MASTER_PORT=54321 \
  -e REPLICATION_USER=repl \
  -e REPLICATION_PASSWORD=repl123 \
  -v kingbase-slave-data:/opt/kingbase/data \
  --link kingbase-master \
  kingbase:v8r7

Kubernetes部署方案 ​

1. StatefulSet部署 ​

持久化存储配置 ​

yaml
# storage-class.yaml
apiVersion: storage.k8s.io/v1
kind: StorageClass
metadata:
  name: kingbase-storage-class
provisioner: kubernetes.io/aws-ebs  # 根据实际环境选择provisioner
parameters:
  type: gp2
  fsType: ext4
reclaimPolicy: Retain
allowVolumeExpansion: true
---
# persistent-volume-claim.yaml
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: kingbase-pvc
spec:
  storageClassName: kingbase-storage-class
  accessModes:
    - ReadWriteOnce
  resources:
    requests:
      storage: 50Gi

StatefulSet部署配置 ​

yaml
apiVersion: apps/v1
kind: StatefulSet
metadata:
  name: kingbase
spec:
  serviceName: kingbase
  replicas: 3
  selector:
    matchLabels:
      app: kingbase
  template:
    metadata:
      labels:
        app: kingbase
    spec:
      containers:
      - name: kingbase
        image: kingbase:v8r7
        ports:
        - containerPort: 54321
          name: kingbase
        env:
        - name: KINGBASE_HOME
          value: /opt/kingbase/ES/V8R7C7
        - name: PGDATA
          value: /opt/kingbase/data
        volumeMounts:
        - name: kingbase-data
          mountPath: /opt/kingbase/data
        - name: kingbase-config
          mountPath: /opt/kingbase/data/kingbase.conf
          subPath: kingbase.conf
        - name: kingbase-hba
          mountPath: /opt/kingbase/data/pg_hba.conf
          subPath: pg_hba.conf
      volumes:
      - name: kingbase-config
        configMap:
          name: kingbase-config
      - name: kingbase-hba
        configMap:
          name: kingbase-hba
  volumeClaimTemplates:
  - metadata:
      name: kingbase-data
    spec:
      storageClassName: kingbase-storage-class
      accessModes:
        - ReadWriteOnce
      resources:
        requests:
          storage: 50Gi
---
# service.yaml
apiVersion: v1
kind: Service
metadata:
  name: kingbase
spec:
  selector:
    app: kingbase
  ports:
  - port: 54321
    targetPort: 54321
  clusterIP: None
---
# configmap.yaml
apiVersion: v1
kind: ConfigMap
metadata:
  name: kingbase-config
data:
  kingbase.conf: |
    listen_addresses = "*"
    port = 54321
    max_connections = 1000
    shared_buffers = 2GB
    effective_cache_size = 4GB
    maintenance_work_mem = 512MB
    checkpoint_completion_target = 0.9
    wal_buffers = 16MB
    default_statistics_target = 100
    random_page_cost = 1.1
    effective_io_concurrency = 200
    work_mem = 2MB
    min_wal_size = 1GB
    max_wal_size = 4GB
    checkpoint_timeout = 30min
    wal_compression = on
---
apiVersion: v1
kind: ConfigMap
metadata:
  name: kingbase-hba
data:
  pg_hba.conf: |
    local all all trust
    host all all 127.0.0.1/32 trust
    host all all ::1/128 trust
    host all all 0.0.0.0/0 md5
    host replication all 0.0.0.0/0 md5

2. 主从复制配置 ​

主库配置 ​

yaml
# kingbase-master-configmap.yaml
apiVersion: v1
kind: ConfigMap
metadata:
  name: kingbase-master-config
data:
  kingbase.conf: |
    # 主库特有配置
    wal_level = replica
    max_wal_senders = 10
    wal_keep_size = 1GB
    hot_standby = on
    max_replication_slots = 10

从库配置 ​

yaml
# kingbase-slave-configmap.yaml
apiVersion: v1
kind: ConfigMap
metadata:
  name: kingbase-slave-config
data:
  kingbase.conf: |
    # 从库特有配置
    hot_standby = on
    max_standby_streaming_delay = 30s
    wal_receiver_status_interval = 10s
    hot_standby_feedback = on

3. 服务暴露 ​

NodePort服务 ​

yaml
apiVersion: v1
kind: Service
metadata:
  name: kingbase-nodeport
spec:
  selector:
    app: kingbase
  type: NodePort
  ports:
  - port: 54321
    targetPort: 54321
    nodePort: 30001

LoadBalancer服务 ​

yaml
apiVersion: v1
kind: Service
metadata:
  name: kingbase-loadbalancer
spec:
  selector:
    app: kingbase
  type: LoadBalancer
  ports:
  - port: 54321
    targetPort: 54321
  loadBalancerIP: 10.0.0.100  # 可选，指定固定IP

4. 监控和日志 ​

Prometheus监控 ​

yaml
# kingbase-exporter.yaml
apiVersion: apps/v1
kind: Deployment
metadata:
  name: kingbase-exporter
spec:
  replicas: 1
  selector:
    matchLabels:
      app: kingbase-exporter
  template:
    metadata:
      labels:
        app: kingbase-exporter
    spec:
      containers:
      - name: kingbase-exporter
        image: prometheuscommunity/postgres-exporter:v0.10.1
        env:
        - name: DATA_SOURCE_NAME
          value: "postgresql://system:Kingbase123@kingbase:54321/test?sslmode=disable"
        ports:
        - containerPort: 9187
---
apiVersion: v1
kind: Service
metadata:
  name: kingbase-exporter
spec:
  selector:
    app: kingbase-exporter
  ports:
  - port: 9187
    targetPort: 9187

日志收集 ​

yaml
# fluentd-configmap.yaml
apiVersion: v1
kind: ConfigMap
metadata:
  name: fluentd-config
data:
  fluent.conf: |
    <source>
      @type tail
      path /var/log/containers/kingbase-*.log
      pos_file /var/log/fluentd-kingbase.pos
      tag kingbase.*
      <parse>
        @type json
        time_format %Y-%m-%dT%H:%M:%S.%NZ
      </parse>
    </source>
    
    <match kingbase.*>
      @type elasticsearch
      host elasticsearch
      port 9200
      index_name kingbase-logs-%Y%m%d
      type_name log
    </match>

版本差异（V8 R6 vs V8 R7） ​

V8 R6 容器化支持 ​

  • 容器化支持相对有限
  • 安装过程需要较多手动干预
  • 配置文件结构较为复杂
  • 对Kubernetes的StatefulSet支持不够完善
  • 缺乏官方容器镜像

V8 R7 容器化支持 ​

  • 增强了容器化支持
  • 提供了更简洁的安装和配置方式
  • 优化了配置文件结构，更适合容器环境
  • 更好地支持Kubernetes StatefulSet
  • 提供了官方容器镜像（部分版本）
  • 支持环境变量配置
  • 增强了日志输出，更适合容器日志收集

最佳实践 ​

1. 持久化存储 ​

  • 使用Kubernetes StatefulSet + PersistentVolumeClaim管理持久化存储
  • 选择合适的StorageClass，推荐使用SSD存储
  • 合理设置存储容量，考虑未来数据增长
  • 启用VolumeSnapshot功能，方便备份和恢复

2. 资源管理 ​

  • 为容器设置合理的资源限制（CPU、内存）
  • 根据实际负载调整资源配置
  • 使用Horizontal Pod Autoscaler实现自动扩缩容
  • 为关键组件设置优先级和抢占策略

3. 网络配置 ​

  • 使用ClusterIP服务进行内部通信
  • 对外暴露使用NodePort或LoadBalancer服务
  • 启用网络策略，限制Pod间通信
  • 考虑使用Service Mesh（如Istio）增强网络安全性

4. 安全配置 ​

  • 不使用默认密码，通过环境变量或Secret管理密码
  • 启用SSL/TLS加密通信
  • 限制容器的特权模式
  • 使用最小权限原则配置容器
  • 定期更新容器镜像，修复安全漏洞

5. 监控和告警 ​

  • 部署Prometheus + Grafana监控KingBaseES性能
  • 配置日志收集，使用ELK或Loki栈
  • 设置合理的告警规则，及时发现问题
  • 监控存储使用率，避免磁盘空间不足

常见问题（FAQ） ​

Q1：Docker容器中如何访问KingBaseES？ ​

A1：

bash
# 本地访问
docker exec -it kingbase ksql -U system -d test

# 外部访问
docker run -it --rm kingbase:v8r7 ksql -h <host-ip> -p 54321 -U system -d test

Q2：Kubernetes中如何初始化KingBaseES集群？ ​

A2：可以使用InitContainer或Job来初始化数据库：

yaml
initContainers:
- name: init-kingbase
  image: kingbase:v8r7
  command: ["bash", "-c", "if [ ! -f /opt/kingbase/data/PG_VERSION ]; then initdb -D /opt/kingbase/data; fi"]
  volumeMounts:
  - name: kingbase-data
    mountPath: /opt/kingbase/data

Q3：如何处理容器化环境中的备份和恢复？ ​

A3：

  • 使用KingBaseES内置的备份工具（sys_rman）
  • 结合Kubernetes CronJob实现定时备份
  • 使用VolumeSnapshot进行存储级备份
  • 考虑使用云原生备份解决方案（如Velero）

Q4：容器化环境中如何进行版本升级？ ​

A4：

  • 采用滚动升级策略
  • 先升级从库，再升级主库
  • 升级前进行充分测试
  • 准备回滚方案

Q5：如何优化容器化KingBaseES的性能？ ​

A5：

  • 合理配置资源限制
  • 使用高性能存储
  • 优化KingBaseES配置参数
  • 调整容器网络设置
  • 启用JIT编译（V8 R7）

总结 ​

容器化部署KingBaseES可以带来诸多好处，包括简化部署流程、提高资源利用率、增强可扩展性和便于管理。在实际部署过程中，需要根据业务需求和环境特点选择合适的部署方案，并遵循最佳实践。同时，要注意不同版本KingBaseES在容器化支持方面的差异，V8 R7相比V8 R6具有更好的容器化支持和云原生特性。