1. 存储方案 三种存储方案： 基于目录隔离公用一个 JuiceFS Elasticsearch 的节点共用一个 JuiceFS，通过子目录挂载不同的 Elasticsearch 节点。 /0/ 对应节点 Node-0 /1/ 对应节点 Node-1 /2/ 对应节点 Node-2 这种方式的好处主要是，易于扩展、配置方便。 基于 JuiceFS 隔离节点数据 Elasticsearch 每个节点都对接一个独立的 JuiceF

1. 创建 Dataset 1 2 3 4 5 6 7 8 9 10 kubectl apply -f - <<EOF apiVersion: v1 kind: Secret metadata: name: my-s3 type: Opaque stringData: aws.accessKeyId: xxx aws.secretKey: xxx EOF 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 kubectl apply -f - <<EOF apiVersion: data.fluid.io/v1alpha1 kind: Dataset metadata: name: my-s3 spec: mounts: - mountPoint: s3://BUCKET/ name: s3 options: alluxio.underfs.s3.endpoint: ks3-cn-beijing-internal.ksyun.com alluxio.underfs.s3.disable.dns.buckets: "false" encryptOptions: - name: aws.accessKeyId valueFrom: secretKeyRef: name: my-s3 key: aws.accessKeyId - name: aws.secretKey valueFrom: secretKeyRef: name: my-s3 key: aws.secretKey accessModes: - ReadWriteMany EOF 2. 创建 Runtime 1 2 3 4 5 6 7 8 9

1. 分析 Fluid 挂载 NFS 存储 查看 Fuse Pod 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 kubectl get pod nfs-demo-fuse-f9wg8 -oyaml apiVersion: v1 kind: Pod metadata: generateName: nfs-demo-fuse- spec: containers: - command: - /usr/local/bin/entrypoint.sh env: - name: FLUID_RUNTIME_TYPE value: thin - name: FLUID_RUNTIME_NS value: default - name: FLUID_RUNTIME_NAME value: nfs-demo - name: MOUNT_POINT value: /runtime-mnt/thin/default/nfs-demo/thin-fuse - name: MOUNT_OPTIONS value: ro image: fluidcloudnative/nfs:v0.1 imagePullPolicy: IfNotPresent lifecycle: preStop: exec: command: - sh - -c - umount /runtime-mnt/thin/default/nfs-demo/thin-fuse name: thin-fuse securityContext: privileged: true volumeMounts:

1. 创建 Dataset 1 2 3 4 5 6 7 8 9 10 kubectl apply -f - <<EOF apiVersion: data.fluid.io/v1alpha1 kind: Dataset metadata: name: nfs-demo spec: mounts: - mountPoint: x.x.x.x:/x-x/ name: nfs-demo EOF 2. 创建 Runtime 1 2 3 4 5 6 7 8 9 10 11 12 13 14 kubectl apply -f - <<EOF apiVersion: data.fluid.io/v1alpha1 kind: ThinRuntimeProfile metadata: name: nfs spec: fileSystemType: nfs fuse: image: fluidcloudnative/nfs imageTag: v0.1 imagePullPolicy: IfNotPresent command: - "/usr/local/bin/entrypoint.sh" EOF 1 2 3 4 5 6 7 8 kubectl apply -f - <<EOF apiVersion: data.fluid.io/v1alpha1 kind: ThinRuntime metadata: name: nfs-demo spec: profileName: nfs EOF 3. 创建测试 Pod 1 2 3 4 5 6 7 8 9 10 11 12

本篇主要记录创建 JuiceFS PVC 的脚本，方便快速配置。组件部署可以参考 使用 Fluid 和 JuiceFS 在 Kubernetes 管理数据 。 1. 设置环境变量 桶的配置 1 2 3 4 5 6 export ACCESS_KEY= export SECRET_KEY= export BUCKET= export ENDPOINT=ks3-cn-beijing-internal.ksyun.com export BUCKET_ENPOINT=$BUCKET.$ENDPOINT export PROVIDER=ks3 Workload 的配置 1 2 3 4 5 export NAMESPACE= export PVC_NAME= export NODE_SELECTOR_KEY= export NODE_SELECTOR_VALUE= 镜像的配置 export JUICEFS_IMAGE=juicedata/juicefs-fuse export DEMO_IMAGE=shaowenchen/demo-ubuntu 元数据的配置 如果是 Redis 配置 1 2 3 4 export REDIS_PASSWORD= #ip:port/database export REDIS_ENDPOINT= export META_SERVER=redis://:$REDIS_PASSWORD@$REDIS_ENDPOINT 如