Deploy the StarRocks Operator and use the Operator to deploy the HA StarRocks cluster

This commit is contained in:
Nicolas 2025-08-20 18:50:45 +08:00
parent e7ec6a4258
commit 44f8a10431
4 changed files with 457 additions and 0 deletions

View File

@ -0,0 +1,128 @@
# StarRocks 高可用集群部署
## 概述
本项目使用 StarRocks Kubernetes Operator 直接部署高可用 StarRocks 集群。
## 组件架构
- **Frontend (FE)**: 3个副本负责元数据管理和查询协调
- **Backend (BE)**: 3个副本负责数据存储和计算
- **Compute Node (CN)**: 2个副本提供弹性计算能力
## 文件说明
### 1. ha-starrocks-cluster.yaml
高可用 StarRocks 集群的主要配置文件,包含:
- FE、BE、CN 组件的副本数和镜像配置
- Pod 反亲和性配置,确保高可用
- 服务配置和环境变量
### 2. starrocks-configmaps.yaml
各组件的配置文件:
- **FE配置**: 日志、端口、JVM参数等
- **BE配置**: 存储、端口、日志级别等
- **CN配置**: 计算节点端口和日志配置
### 3. ha-starrocks-values.yaml
Helm values 配置文件(用于参考)
### 4. values.yaml
Helm Chart 的默认值配置(用于现有集群)
### 5. vpa.yaml
垂直 Pod 自动扩缩容配置
## 部署步骤
### 1. 应用配置
```bash
# 首先应用 ConfigMap 配置
kubectl apply -f starrocks-configmaps.yaml
# 然后应用 StarRocks 集群配置
kubectl apply -f ha-starrocks-cluster.yaml
```
### 2. 检查部署状态
```bash
# 检查 StarRocks 集群状态
kubectl get starrockscluster -n freeleaps-data-platform
# 检查 Pod 状态
kubectl get pods -n freeleaps-data-platform -l app=starrocks
# 检查服务状态
kubectl get svc -n freeleaps-data-platform -l app=starrocks
```
### 3. 查看日志
```bash
# 查看 FE 日志
kubectl logs -n freeleaps-data-platform -l app.kubernetes.io/component=fe
# 查看 BE 日志
kubectl logs -n freeleaps-data-platform -l app.kubernetes.io/component=be
# 查看 CN 日志
kubectl logs -n freeleaps-data-platform -l app.kubernetes.io/component=cn
```
## 高可用特性
- **FE高可用**: 3个FE节点支持自动故障转移
- **BE高可用**: 3个BE节点数据多副本存储
- **CN弹性**: 支持水平扩展和自动扩缩容
- **Pod反亲和性**: 确保组件分布在不同节点上
- **自动故障恢复**: Operator自动管理集群状态
## 网络配置
- **内部服务**: 使用ClusterIP类型
- **端口映射**:
- FE: 8030(HTTP), 9010(编辑日志), 9020(RPC), 9030(查询)
- BE: 8040(HTTP), 9050(心跳), 8060(BRPC), 9060(BE端口)
- CN: 8040(HTTP), 9050(心跳), 8060(BRPC), 9060(Thrift)
## 存储配置
- **FE元数据**: 使用现有存储配置
- **BE数据**: 使用现有存储配置
- **CN缓存**: 使用现有存储配置
- **数据持久化**: 支持PVC自动创建和绑定
## 监控和运维
- **健康检查**: 内置就绪和存活探针
- **日志收集**: 结构化日志输出
- **资源监控**: 支持Prometheus指标收集
- **自动扩缩容**: VPA支持垂直资源调整
- **Operator管理**: 自动集群生命周期管理
## 升级和回滚
```bash
# 升级集群配置
kubectl apply -f ha-starrocks-cluster.yaml
# 回滚到上一个版本
kubectl rollout undo deployment/starrocks-operator -n freeleaps-data-platform
# 查看集群状态
kubectl describe starrockscluster freeleaps-starrocks -n freeleaps-data-platform
```
## 故障排除
```bash
# 检查集群状态
kubectl describe starrockscluster freeleaps-starrocks -n freeleaps-data-platform
# 检查 Pod 事件
kubectl get events -n freeleaps-data-platform --sort-by='.lastTimestamp'
# 检查 Operator 日志
kubectl logs -n freeleaps-data-platform -l app=starrocks-operator
# 检查存储状态
kubectl get pvc -n freeleaps-data-platform
kubectl get pv
```
## 注意事项
1. **现有集群**: 此配置将升级现有的 `freeleaps-starrocks` 集群
2. **数据安全**: 升级过程中数据不会丢失
3. **滚动更新**: Operator 会自动进行滚动更新,确保服务连续性
4. **资源需求**: 确保集群有足够的资源支持增加的副本数

View File

@ -0,0 +1,105 @@
apiVersion: starrocks.com/v1
kind: StarRocksCluster
metadata:
name: freeleaps-starrocks
namespace: freeleaps-data-platform
labels:
app: starrocks
component: analytics
cluster-type: ha
spec:
starRocksFeSpec:
replicas: 3
image: starrocks/fe-ubuntu:3.3-latest
service:
type: ClusterIP
feEnvVars:
- name: TZ
value: UTC
configMapInfo:
configMapName: ha-starrocks-fe-config
resolveKey: fe.conf
affinity:
podAntiAffinity:
requiredDuringSchedulingIgnoredDuringExecution:
- labelSelector:
matchExpressions:
- key: app.kubernetes.io/component
operator: In
values:
- fe
topologyKey: "kubernetes.io/hostname"
tolerations:
- key: "node-role.kubernetes.io/control-plane"
operator: "Exists"
effect: "NoSchedule"
- key: "node-role.kubernetes.io/devops"
operator: "Exists"
effect: "NoSchedule"
- key: "node-role.kubernetes.io/disk-pressure"
operator: "Exists"
effect: "NoSchedule"
starRocksBeSpec:
replicas: 3
image: starrocks/be-ubuntu:3.3-latest
service:
type: ClusterIP
beEnvVars:
- name: TZ
value: UTC
configMapInfo:
configMapName: ha-starrocks-be-config
resolveKey: be.conf
affinity:
podAntiAffinity:
requiredDuringSchedulingIgnoredDuringExecution:
- labelSelector:
matchExpressions:
- key: app.kubernetes.io/component
operator: In
values:
- be
topologyKey: "kubernetes.io/hostname"
tolerations:
- key: "node-role.kubernetes.io/control-plane"
operator: "Exists"
effect: "NoSchedule"
- key: "node-role.kubernetes.io/devops"
operator: "Exists"
effect: "NoSchedule"
- key: "node-role.kubernetes.io/disk-pressure"
operator: "Exists"
effect: "NoSchedule"
starRocksCnSpec:
replicas: 2
image: starrocks/cn-ubuntu:3.3-latest
service:
type: ClusterIP
cnEnvVars:
- name: TZ
value: UTC
configMapInfo:
configMapName: ha-starrocks-cn-config
resolveKey: cn.conf
affinity:
podAntiAffinity:
requiredDuringSchedulingIgnoredDuringExecution:
- labelSelector:
matchExpressions:
- key: app.kubernetes.io/component
operator: In
values:
- cn
topologyKey: "kubernetes.io/hostname"
tolerations:
- key: "node-role.kubernetes.io/control-plane"
operator: "Exists"
effect: "NoSchedule"
- key: "node-role.kubernetes.io/devops"
operator: "Exists"
effect: "NoSchedule"
- key: "node-role.kubernetes.io/disk-pressure"
operator: "Exists"
effect: "NoSchedule"

View File

@ -0,0 +1,181 @@
# 高可用 StarRocks 集群配置
# 基于现有的 values.yaml配置高可用特性
# 设置时区
timeZone: UTC
# 部署 StarRocks 集群
starrocksCluster:
# 集群名称
name: "freeleaps-starrocks"
# 命名空间
namespace: "freeleaps-data-platform"
# 启用 BE 和 CN 组件
enabledBe: true
enabledCn: true
# 全局组件配置
componentValues:
# 镜像版本
image:
tag: "3.3-latest"
# 时区
timeZone: UTC
# FE 组件配置
starrocksFESpec:
# 副本数 - 高可用需要至少3个
replicas: 3
# 镜像
image:
repository: starrocks/fe-ubuntu
tag: "3.3-latest"
# 资源配置
resources:
requests:
cpu: 2
memory: 4Gi
limits:
cpu: 4
memory: 8Gi
# 存储配置
storageSpec:
name: "fe"
storageClassName: "azure-disk-std-ssd-lrs"
storageSize: 20Gi
logStorageClassName: "azure-disk-std-ssd-lrs"
logStorageSize: 10Gi
# 服务配置
service:
type: "ClusterIP"
# 环境变量
feEnvVars:
- name: TZ
value: UTC
# 配置文件
config: |
LOG_DIR = ${STARROCKS_HOME}/log
DATE = "$(date +%Y%m%d-%H%M%S)"
JAVA_OPTS="-Dlog4j2.formatMsgNoLookups=true -Xmx4g -XX:+UseG1GC -Xlog:gc*:${LOG_DIR}/fe.gc.log.$DATE:time"
http_port = 8030
rpc_port = 9020
query_port = 9030
edit_log_port = 9010
mysql_service_nio_enabled = true
sys_log_level = INFO
# Pod 反亲和性 - 确保高可用
affinity:
podAntiAffinity:
requiredDuringSchedulingIgnoredDuringExecution:
- labelSelector:
matchExpressions:
- key: app.kubernetes.io/component
operator: In
values:
- fe
topologyKey: "kubernetes.io/hostname"
# BE 组件配置
starrocksBeSpec:
# 副本数 - 高可用需要至少3个
replicas: 3
# 镜像
image:
repository: starrocks/be-ubuntu
tag: "3.3-latest"
# 资源配置
resources:
requests:
cpu: 4
memory: 8Gi
limits:
cpu: 8
memory: 16Gi
# 存储配置
storageSpec:
name: "be"
storageClassName: "azure-disk-std-ssd-lrs"
storageSize: 50Gi
logStorageClassName: "azure-disk-std-ssd-lrs"
logStorageSize: 20Gi
# 服务配置
service:
type: "ClusterIP"
# 环境变量
beEnvVars:
- name: TZ
value: UTC
# 配置文件
config: |
be_port = 9060
webserver_port = 8040
heartbeat_service_port = 9050
brpc_port = 8060
sys_log_level = INFO
default_rowset_type = beta
# Pod 反亲和性 - 确保高可用
affinity:
podAntiAffinity:
requiredDuringSchedulingIgnoredDuringExecution:
- labelSelector:
matchExpressions:
- key: app.kubernetes.io/component
operator: In
values:
- be
topologyKey: "kubernetes.io/hostname"
# CN 组件配置
starrocksCnSpec:
# 副本数 - 弹性计算节点
replicas: 2
# 镜像
image:
repository: starrocks/cn-ubuntu
tag: "3.3-latest"
# 资源配置
resources:
requests:
cpu: 2
memory: 4Gi
limits:
cpu: 4
memory: 8Gi
# 存储配置
storageSpec:
name: "cn"
storageClassName: "azure-disk-std-ssd-lrs"
storageSize: 20Gi
logStorageClassName: "azure-disk-std-ssd-lrs"
logStorageSize: 10Gi
# 服务配置
service:
type: "ClusterIP"
# 环境变量
cnEnvVars:
- name: TZ
value: UTC
# 配置文件
config: |
sys_log_level = INFO
thrift_port = 9060
webserver_port = 8040
heartbeat_service_port = 9050
brpc_port = 8060
# Pod 反亲和性 - 确保高可用
affinity:
podAntiAffinity:
requiredDuringSchedulingIgnoredDuringExecution:
- labelSelector:
matchExpressions:
- key: app.kubernetes.io/component
operator: In
values:
- cn
topologyKey: "kubernetes.io/hostname"
# 监控配置
metrics:
serviceMonitor:
enabled: true
interval: 15s

View File

@ -0,0 +1,43 @@
apiVersion: v1
kind: ConfigMap
metadata:
name: ha-starrocks-fe-config
namespace: freeleaps-data-platform
data:
fe.conf: |
LOG_DIR = ${STARROCKS_HOME}/log
DATE = "$(date +%Y%m%d-%H%M%S)"
JAVA_OPTS="-Dlog4j2.formatMsgNoLookups=true -Xmx4g -XX:+UseG1GC -Xlog:gc*:${LOG_DIR}/fe.gc.log.$DATE:time"
http_port = 8030
rpc_port = 9020
query_port = 9030
edit_log_port = 9010
mysql_service_nio_enabled = true
sys_log_level = INFO
---
apiVersion: v1
kind: ConfigMap
metadata:
name: ha-starrocks-be-config
namespace: freeleaps-data-platform
data:
be.conf: |
be_port = 9060
webserver_port = 8040
heartbeat_service_port = 9050
brpc_port = 8060
sys_log_level = INFO
default_rowset_type = beta
---
apiVersion: v1
kind: ConfigMap
metadata:
name: ha-starrocks-cn-config
namespace: freeleaps-data-platform
data:
cn.conf: |
sys_log_level = INFO
thrift_port = 9060
webserver_port = 8040
heartbeat_service_port = 9050
brpc_port = 8060