本文介绍如何在 Azure Kubernetes 服务（AKS）上配置和部署 Valkey 群集，包括创建 Valkey 群集 ConfigMap、主群集和辅助群集 Pod，以确保冗余和区域复制，以及确保高可用性的 Pod 中断预算（PDB）。
注意
本文包含对术语“主”和“从”的引用，Microsoft 不再使用这些术语。 在从 Valkey 软件中删除该术语后，我们会将其从本文中删除。
连接到 AKS 群集
使用
kubectl命令将az aks get-credentials配置为连接到 AKS 群集。az aks get-credentials --resource-group $MY_RESOURCE_GROUP_NAME --name $MY_CLUSTER_NAME --overwrite-existing --output table
创建命名空间
使用
kubectl create namespace命令为 Valkey 群集创建一个命名空间。kubectl create namespace valkey --dry-run=client --output yaml | kubectl apply -f -示例输出：
namespace/valkey created
创建机密
使用 OpenSSL 为 Valkey 群集生成随机密码，并使用
az keyvault secret set命令将其存储在 Azure 密钥保管库中。 设置策略以允许用户分配的标识使用az keyvault set-policy命令获取机密。SECRET=$(openssl rand -base64 32) echo requirepass $SECRET > /tmp/valkey-password-file.conf echo primaryauth $SECRET >> /tmp/valkey-password-file.conf az keyvault secret set --vault-name $MY_KEYVAULT_NAME --name valkey-password-file --file /tmp/valkey-password-file.conf --output none rm /tmp/valkey-password-file.conf az keyvault set-policy --name $MY_KEYVAULT_NAME --object-id $userAssignedObjectID --secret-permissions get --output table
使用
az aks show命令获取 Azure KeyVault 机密提供程序加载项创建的标识 ID 和对象 ID。export userAssignedIdentityID=$(az aks show --resource-group $MY_RESOURCE_GROUP_NAME --name $MY_CLUSTER_NAME --query addonProfiles.azureKeyvaultSecretsProvider.identity.clientId --output tsv) export userAssignedObjectID=$(az aks show --resource-group $MY_RESOURCE_GROUP_NAME --name $MY_CLUSTER_NAME --query addonProfiles.azureKeyvaultSecretsProvider.identity.objectId --output tsv)
创建一个
SecretProviderClass资源，以使用kubectl apply命令访问存储在密钥保管库中的 Valkey 密码。export TENANT_ID=$(az account show --query tenantId --output tsv) kubectl apply -f - <<EOF --- apiVersion: secrets-store.csi.x-k8s.io/v1 kind: SecretProviderClass metadata: name: valkey-password namespace: valkey spec: provider: azure parameters: usePodIdentity: "false" useVMManagedIdentity: "true" userAssignedIdentityID: "${userAssignedIdentityID}" keyvaultName: ${MY_KEYVAULT_NAME} # the name of the AKV instance objects: | array: - | objectName: valkey-password-file objectAlias: valkey-password-file.conf objectType: secret tenantId: "${TENANT_ID}" # the tenant ID of the AKV instance EOF
创建 Valkey 配置文件
创建用于存储 Valkey 配置文件的
ConfigMap资源。kubectl apply -f - <<EOF apiVersion: v1 kind: ConfigMap metadata: name: valkey-cluster namespace: valkey data: valkey.conf: |+ cluster-enabled yes cluster-node-timeout 15000 cluster-config-file /data/nodes.conf appendonly yes protected-mode yes dir /data port 6379 include /etc/valkey-password/valkey-password-file.conf EOF示例输出：
configmap/valkey-cluster created
创建 Valkey 主群集 Pod
使用
StatefulSet命令创建一个spec.affinity资源，其kubectl apply目标是将所有主数据库保留在区域 1 和区域 2 中（最好在不同的节点中）。kubectl apply -f - <<EOF --- apiVersion: apps/v1 kind: StatefulSet metadata: name: valkey-masters namespace: valkey spec: serviceName: "valkey-masters" replicas: 3 selector: matchLabels: app: valkey template: metadata: labels: app: valkey appCluster: valkey-masters spec: terminationGracePeriodSeconds: 20 affinity: nodeAffinity: requiredDuringSchedulingIgnoredDuringExecution: nodeSelectorTerms: - matchExpressions: - key: agentpool operator: In values: - valkey - key: topology.kubernetes.io/zone operator: In values: - ${MY_LOCATION}-1 - matchExpressions: - key: agentpool operator: In values: - valkey - key: topology.kubernetes.io/zone operator: In values: - ${MY_LOCATION}-2 podAntiAffinity: preferredDuringSchedulingIgnoredDuringExecution: - weight: 100 podAffinityTerm: labelSelector: matchExpressions: - key: app operator: In values: - valkey topologyKey: topology.kubernetes.io/zone - weight: 90 podAffinityTerm: labelSelector: matchExpressions: - key: app operator: In values: - valkey topologyKey: kubernetes.io/hostname containers: - name: role-master-checker image: "${MY_ACR_REGISTRY}.azurecr.cn/valkey:latest" command: - "/bin/bash" - "-c" args: [ "while true; do role=\$(valkey-cli --pass \$(cat /etc/valkey-password/valkey-password-file.conf | awk '{print \$2; exit}') role | awk '{print \$1; exit}'); if [ \"\$role\" = \"slave\" ]; then valkey-cli --pass \$(cat /etc/valkey-password/valkey-password-file.conf | awk '{print \$2; exit}') cluster failover; fi; sleep 30; done" ] volumeMounts: - name: valkey-password mountPath: /etc/valkey-password readOnly: true - name: valkey image: "${MY_ACR_REGISTRY}.azurecr.cn/valkey:latest" env: - name: VALKEY_PASSWORD_FILE value: "/etc/valkey-password/valkey-password-file.conf" - name: MY_POD_IP valueFrom: fieldRef: fieldPath: status.podIP command: - "valkey-server" args: - "/conf/valkey.conf" - "--cluster-announce-ip" - "\$(MY_POD_IP)" resources: requests: cpu: "100m" memory: "100Mi" ports: - name: valkey containerPort: 6379 protocol: "TCP" - name: cluster containerPort: 16379 protocol: "TCP" volumeMounts: - name: conf mountPath: /conf readOnly: false - name: data mountPath: /data readOnly: false - name: valkey-password mountPath: /etc/valkey-password readOnly: true volumes: - name: valkey-password csi: driver: secrets-store.csi.k8s.io readOnly: true volumeAttributes: secretProviderClass: valkey-password - name: conf configMap: name: valkey-cluster defaultMode: 0755 volumeClaimTemplates: - metadata: name: data spec: accessModes: [ "ReadWriteOnce" ] storageClassName: managed-csi-premium resources: requests: storage: 20Gi EOF示例输出：
statefulset.apps/valkey-masters created
创建 Valkey 集群容器组
使用
StatefulSet命令为 Valkey 辅助数据库创建第二个spec.affinity资源，其kubectl apply目标是将所有副本保留在区域 3 中（最好在不同的节点中）。kubectl apply -f - <<EOF --- apiVersion: apps/v1 kind: StatefulSet metadata: name: valkey-replicas namespace: valkey spec: serviceName: "valkey-replicas" replicas: 3 selector: matchLabels: app: valkey template: metadata: labels: app: valkey appCluster: valkey-replicas spec: terminationGracePeriodSeconds: 20 affinity: nodeAffinity: requiredDuringSchedulingIgnoredDuringExecution: nodeSelectorTerms: - matchExpressions: - key: agentpool operator: In values: - valkey - key: topology.kubernetes.io/zone operator: In values: - ${MY_LOCATION}-3 podAntiAffinity: preferredDuringSchedulingIgnoredDuringExecution: - weight: 90 podAffinityTerm: labelSelector: matchExpressions: - key: app operator: In values: - valkey topologyKey: kubernetes.io/hostname containers: - name: valkey image: "${MY_ACR_REGISTRY}.azurecr.cn/valkey:latest" env: - name: VALKEY_PASSWORD_FILE value: "/etc/valkey-password/valkey-password-file.conf" - name: MY_POD_IP valueFrom: fieldRef: fieldPath: status.podIP command: - "valkey-server" args: - "/conf/valkey.conf" - "--cluster-announce-ip" - "\$(MY_POD_IP)" resources: requests: cpu: "100m" memory: "100Mi" ports: - name: valkey containerPort: 6379 protocol: "TCP" - name: cluster containerPort: 16379 protocol: "TCP" volumeMounts: - name: conf mountPath: /conf readOnly: false - name: data mountPath: /data readOnly: false - name: valkey-password mountPath: /etc/valkey-password readOnly: true volumes: - name: valkey-password csi: driver: secrets-store.csi.k8s.io readOnly: true volumeAttributes: secretProviderClass: valkey-password - name: conf configMap: name: valkey-cluster defaultMode: 0755 volumeClaimTemplates: - metadata: name: data spec: accessModes: [ "ReadWriteOnce" ] storageClassName: managed-csi-premium resources: requests: storage: 20Gi EOF示例输出：
statefulset.apps/valkey-replicas created
验证 Pod 和节点分布
使用
master-N和replica-N命令验证kubectl get nodes和kubectl get pods在不同的节点和区域中运行。kubectl get pods -n valkey -o wide kubectl get node -o custom-columns=Name:.metadata.name,Zone:".metadata.labels.topology\.kubernetes\.io/zone"示例输出：
NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES valkey-masters-0 1/1 Running 0 2m55s 10.224.0.4 aks-valkey-18693609-vmss000004 <none> <none> valkey-masters-1 1/1 Running 0 2m31s 10.224.0.137 aks-valkey-18693609-vmss000000 <none> <none> valkey-masters-2 1/1 Running 0 2m7s 10.224.0.222 aks-valkey-18693609-vmss000001 <none> <none> valkey-replicas-0 1/1 Running 0 88s 10.224.0.237 aks-valkey-18693609-vmss000005 <none> <none> valkey-replicas-1 1/1 Running 0 70s 10.224.0.18 aks-valkey-18693609-vmss000002 <none> <none> valkey-replicas-2 1/1 Running 0 48s 10.224.0.242 aks-valkey-18693609-vmss000005 <none> <none> Name Zone aks-nodepool1-17621399-vmss000000 chinanorth3-1 aks-nodepool1-17621399-vmss000001 chinanorth3-2 aks-nodepool1-17621399-vmss000003 chinanorth3-3 aks-valkey-18693609-vmss000000 chinanorth3-1 aks-valkey-18693609-vmss000001 chinanorth3-2 aks-valkey-18693609-vmss000002 chinanorth3-3 aks-valkey-18693609-vmss000003 chinanorth3-1 aks-valkey-18693609-vmss000004 chinanorth3-2 aks-valkey-18693609-vmss000005 chinanorth3-3在继续下一步之前，等待所有 Pod 都运行。
创建无头服务
使用
Service命令创建三个无外设kubectl apply资源（第一个用于整个群集，第二个用于主项，第三个用于辅助项），用于获取 Valkey Pod 的 IP 地址。kubectl apply -f - <<EOF apiVersion: v1 kind: Service metadata: name: valkey-cluster namespace: valkey spec: clusterIP: None ports: - name: valkey-port port: 6379 protocol: TCP targetPort: 6379 selector: app: valkey sessionAffinity: None type: ClusterIP EOF kubectl apply -f - <<EOF apiVersion: v1 kind: Service metadata: name: valkey-masters namespace: valkey spec: clusterIP: None ports: - name: valkey-port port: 6379 protocol: TCP targetPort: 6379 selector: app: valkey appCluster: valkey-masters sessionAffinity: None type: ClusterIP EOF kubectl apply -f - <<EOF apiVersion: v1 kind: Service metadata: name: valkey-replicas namespace: valkey spec: clusterIP: None ports: - name: valkey-port port: 6379 protocol: TCP targetPort: 6379 selector: app: valkey appCluster: valkey-replicas sessionAffinity: None type: ClusterIP EOF示例输出：
service/valkey-cluster created service/valkey-masters created service/valkey-replicas created
创建 Pod 中断预算 （PDB）
创建 Pod 中断预算（PDB），以确保在自愿中断（例如升级或维护）期间最多只有一个 Pod 不可用。 这有助于维护 Kubernetes 群集中 Valkey 应用程序的稳定性和可用性。
kubectl apply -f - <<EOF apiVersion: policy/v1 kind: PodDisruptionBudget metadata: name: valkey namespace: valkey spec: maxUnavailable: 1 selector: matchLabels: app: valkey EOF示例输出：
poddisruptionbudget.policy/valkey created
运行 Valkey 群集
使用
kubectl exec命令将区域 1 和 2 中的 Valkey 主数据库添加到群集。kubectl exec -it -n valkey valkey-masters-0 -- valkey-cli --cluster create --cluster-yes --cluster-replicas 0 \ valkey-masters-0.valkey-masters.valkey.svc.cluster.local:6379 \ valkey-masters-1.valkey-masters.valkey.svc.cluster.local:6379 \ valkey-masters-2.valkey-masters.valkey.svc.cluster.local:6379 \ --pass ${SECRET}示例输出：
>>> Performing hash slots allocation on 3 nodes... Master[0] -> Slots 0 - 5460 Master[1] -> Slots 5461 - 10922 Master[2] -> Slots 10923 - 16383 M: ee6ac1d00d3f016b6f46c7ce11199bc1a7809a35 valkey-masters-0.valkey-masters.valkey.svc.cluster.local:6379 slots:[0-5460] (5461 slots) master M: fd1fb98db83976478e05edd3d2a02f9a13badd80 valkey-masters-1.valkey-masters.valkey.svc.cluster.local:6379 slots:[5461-10922] (5462 slots) master M: ea47bf57ae7080ef03164a4d48b662c7b4c8770e valkey-masters-2.valkey-masters.valkey.svc.cluster.local:6379 slots:[10923-16383] (5461 slots) master >>> Nodes configuration updated >>> Assign a different config epoch to each node >>> Sending CLUSTER MEET messages to join the cluster Waiting for the cluster to join ... >>> Performing Cluster Check (using node valkey-masters-0.valkey-masters.valkey.svc.cluster.local:6379) M: ee6ac1d00d3f016b6f46c7ce11199bc1a7809a35 valkey-masters-0.valkey-masters.valkey.svc.cluster.local:6379 slots:[0-5460] (5461 slots) master M: ea47bf57ae7080ef03164a4d48b662c7b4c8770e 10.224.0.176:6379 slots:[10923-16383] (5461 slots) master M: fd1fb98db83976478e05edd3d2a02f9a13badd80 10.224.0.247:6379 slots:[5461-10922] (5462 slots) master [OK] All nodes agree about slots configuration. >>> Check for open slots... >>> Check slots coverage... [OK] All 16384 slots covered.使用
kubectl exec命令将区域 3 中的 Valkey 副本添加到群集。kubectl exec -ti -n valkey valkey-masters-0 -- valkey-cli --cluster add-node \ valkey-replicas-0.valkey-replicas.valkey.svc.cluster.local:6379 \ valkey-masters-0.valkey-masters.valkey.svc.cluster.local:6379 --cluster-slave \ --pass ${SECRET} kubectl exec -ti -n valkey valkey-masters-0 -- valkey-cli --cluster add-node \ valkey-replicas-1.valkey-replicas.valkey.svc.cluster.local:6379 \ valkey-masters-1.valkey-masters.valkey.svc.cluster.local:6379 --cluster-slave \ --pass ${SECRET} kubectl exec -ti -n valkey valkey-masters-0 -- valkey-cli --cluster add-node \ valkey-replicas-2.valkey-replicas.valkey.svc.cluster.local:6379 \ valkey-masters-2.valkey-masters.valkey.svc.cluster.local:6379 --cluster-slave \ --pass ${SECRET}示例输出：
>>> Adding node valkey-replicas-0.valkey-replicas.valkey.svc.cluster.local:6379 to cluster valkey-masters-0.valkey-masters.valkey.svc.cluster.local:6379 >>> Performing Cluster Check (using node valkey-masters-0.valkey-masters.valkey.svc.cluster.local:6379) M: ee6ac1d00d3f016b6f46c7ce11199bc1a7809a35 valkey-masters-0.valkey-masters.valkey.svc.cluster.local:6379 slots:[0-5460] (5461 slots) master M: ea47bf57ae7080ef03164a4d48b662c7b4c8770e 10.224.0.176:6379 slots:[10923-16383] (5461 slots) master M: fd1fb98db83976478e05edd3d2a02f9a13badd80 10.224.0.247:6379 slots:[5461-10922] (5462 slots) master [OK] All nodes agree about slots configuration. >>> Check for open slots... >>> Check slots coverage... [OK] All 16384 slots covered. Automatically selected master valkey-masters-0.valkey-masters.valkey.svc.cluster.local:6379 >>> Send CLUSTER MEET to node valkey-replicas-0.valkey-replicas.valkey.svc.cluster.local:6379 to make it join the cluster. Waiting for the cluster to join >>> Configure node as replica of valkey-masters-0.valkey-masters.valkey.svc.cluster.local:6379. [OK] New node added correctly. >>> Adding node valkey-replicas-1.valkey-replicas.valkey.svc.cluster.local:6379 to cluster valkey-masters-1.valkey-masters.valkey.svc.cluster.local:6379 >>> Performing Cluster Check (using node valkey-masters-1.valkey-masters.valkey.svc.cluster.local:6379) M: fd1fb98db83976478e05edd3d2a02f9a13badd80 valkey-masters-1.valkey-masters.valkey.svc.cluster.local:6379 slots:[5461-10922] (5462 slots) master S: 0ebceb60cbcc31da9040159440a1f4856b992907 10.224.0.224:6379 slots: (0 slots) slave replicates ee6ac1d00d3f016b6f46c7ce11199bc1a7809a35 M: ea47bf57ae7080ef03164a4d48b662c7b4c8770e 10.224.0.176:6379 slots:[10923-16383] (5461 slots) master M: ee6ac1d00d3f016b6f46c7ce11199bc1a7809a35 10.224.0.14:6379 slots:[0-5460] (5461 slots) master 1 additional replica(s) [OK] All nodes agree about slots configuration. >>> Check for open slots... >>> Check slots coverage... [OK] All 16384 slots covered. Automatically selected master valkey-masters-1.valkey-masters.valkey.svc.cluster.local:6379 >>> Send CLUSTER MEET to node valkey-replicas-1.valkey-replicas.valkey.svc.cluster.local:6379 to make it join the cluster. Waiting for the cluster to join >>> Configure node as replica of valkey-masters-1.valkey-masters.valkey.svc.cluster.local:6379. [OK] New node added correctly. >>> Adding node valkey-replicas-2.valkey-replicas.valkey.svc.cluster.local:6379 to cluster valkey-masters-2.valkey-masters.valkey.svc.cluster.local:6379 >>> Performing Cluster Check (using node valkey-masters-2.valkey-masters.valkey.svc.cluster.local:6379) M: ea47bf57ae7080ef03164a4d48b662c7b4c8770e valkey-masters-2.valkey-masters.valkey.svc.cluster.local:6379 slots:[10923-16383] (5461 slots) master S: 0ebceb60cbcc31da9040159440a1f4856b992907 10.224.0.224:6379 slots: (0 slots) slave replicates ee6ac1d00d3f016b6f46c7ce11199bc1a7809a35 S: fa44edff683e2e01ee5c87233f9f3bc35c205dce 10.224.0.103:6379 slots: (0 slots) slave replicates fd1fb98db83976478e05edd3d2a02f9a13badd80 M: ee6ac1d00d3f016b6f46c7ce11199bc1a7809a35 10.224.0.14:6379 slots:[0-5460] (5461 slots) master 1 additional replica(s) M: fd1fb98db83976478e05edd3d2a02f9a13badd80 10.224.0.247:6379 slots:[5461-10922] (5462 slots) master 1 additional replica(s) [OK] All nodes agree about slots configuration. >>> Check for open slots... >>> Check slots coverage... [OK] All 16384 slots covered. Automatically selected master valkey-masters-2.valkey-masters.valkey.svc.cluster.local:6379 >>> Send CLUSTER MEET to node valkey-replicas-2.valkey-replicas.valkey.svc.cluster.local:6379 to make it join the cluster. Waiting for the cluster to join >>> Configure node as replica of valkey-masters-2.valkey-masters.valkey.svc.cluster.local:6379. [OK] New node added correctly.使用以下命令验证 Pod 的角色：
for x in $(seq 0 2); do echo "valkey-masters-$x"; kubectl exec -n valkey valkey-masters-$x -- valkey-cli --pass ${SECRET} role; echo; done for x in $(seq 0 2); do echo "valkey-replicas-$x"; kubectl exec -n valkey valkey-replicas-$x -- valkey-cli --pass ${SECRET} role; echo; done示例输出：
valkey-masters-0 master 84 10.224.0.224 6379 84 valkey-masters-1 master 84 10.224.0.103 6379 84 valkey-masters-2 master 70 10.224.0.200 6379 70 valkey-replicas-0 slave 10.224.0.14 6379 connected 98 valkey-replicas-1 slave 10.224.0.247 6379 connected 98 valkey-replicas-2 slave 10.224.0.176 6379 connected 84
后续步骤
若要详细了解如何在 Azure Kubernetes 服务 (AKS) 上部署开源软件，请参阅以下文章：
供稿人
Microsoft 会维护本文。 本系列文章为以下参与者的原创作品：
- Nelly Kiboi | 服务工程师
- Saverio Proto | 首席客户体验工程师
- Don High | 首席客户工程师
- LaBrina Loving | 首席服务工程师
- Ken Kilty | 首席 TPM
- Russell de Pina | 首席 TPM
- Colin Mixon | 产品经理
- Ketan Chawda | 高级客户工程师
- Naveed Kharadi | 客户体验工程师
- Erin Schaffer | 内容开发人员 2