K8S部署rook+ceph云原生存储

1. 概念

1.1. Ceph

   Ceph是一个开源的分布式存储系统,具有高扩展性、高性能、高可靠性等特点,提供良好的性能、可靠性和可扩展性。支持对象存储、快存储和文件系统,是目前为云平台提供存储的理想方案。

1.2. Rook

   Rook是一个自我管理的分布式存储编排系统,它本身并不是存储系统,在存储和k8s之间搭建了一个桥梁,使存储系统的搭建或者维护变得特别简单,Rook将分布式存储系统转变为自我管理、自我扩展、自我修复的存储服务。它让一些存储的操作,比如部署、配置、扩容、升级、迁移、灾难恢复、监视和资源管理变得自动化,无需人工处理。并且Rook支持cSsl,可以利用CSI做一些Pvc的快照、扩容等操作。

1.3. 架构

Rook Components on Kubernetes

  • 各组件说明
    • Operator:Rook控制端,监控存储守护进程,确保存储集群的健康
    • Agent:在每个存储节点创建,配置了FlexVolume插件和Kubernetes 的存储卷控制框架(CSI)进行集成
    • OSD:提供存储,每块硬盘可以看做一个osd
    • Mon:监控ceph集群的存储情况,记录集群的拓扑,数据存储的位置信息
    • MDS:负责跟踪文件存储的层次结构
    • RGW:Rest API结构,提供对象存储接口
    • MGR:为外界提供统一入口

2. 部署rook+ceph

2.1. 准备事项

2.1.1. 建议配置
环境测试生产环境
服务器配置(≥3)4核8G8核16G,或更高
硬盘(数据盘)每台1×20G每台2×500G,或更高
网卡——千兆网卡

注1:学习用建议阿里云–抢占式实例–张家口区域,4核8G仅需5分钱/小时

注2:osd仅支持裸盘挂载,不支持目录格式

2.1.2. 本文环境
环境说明
k8s版本1.20.4
rook版本1.6.5
ceph节点k8s-node2,k8s-node3,k8s-node4
硬盘vdb(20G),vdc(20G)
2.1.3. 注意事项
  • 防火墙
  • 文件描述符大小
  • 时间同步
  • 检查污点:kubectl describe node |grep Taint
  • 挂载磁盘:lsblk -f,必须为裸盘
  • ceph节点最好不要部署其它服务,如harbor,istio等,可能存在冲突
2.1.4. 克隆rook项目到本地
git clone --single-branch --branch v1.6.5 https://github.com/rook/rook.git
2.1.5. 同步镜像

注:所涉及的镜像都比较大,且gcr镜像下载速度极慢,非常有必要存到自建仓库里,加速pod的创建。

  • 查看所需镜像
- 进入工作目录
cd /root/rook/cluster/examples/kubernetes/ceph

vim operator.yaml
-------------------------------
  # ROOK_CSI_CEPH_IMAGE: "quay.io/cephcsi/cephcsi:v3.3.1"
  # ROOK_CSI_REGISTRAR_IMAGE: "k8s.gcr.io/sig-storage/csi-node-driver-registrar:v2.0.1"
  # ROOK_CSI_RESIZER_IMAGE: "k8s.gcr.io/sig-storage/csi-resizer:v1.0.1"
  # ROOK_CSI_PROVISIONER_IMAGE: "k8s.gcr.io/sig-storage/csi-provisioner:v2.0.4"
  # ROOK_CSI_SNAPSHOTTER_IMAGE: "k8s.gcr.io/sig-storage/csi-snapshotter:v4.0.0"
  # ROOK_CSI_ATTACHER_IMAGE: "k8s.gcr.io/sig-storage/csi-attacher:v3.0.2"

vim cluster.yaml
--------------------------------
    image: ceph/ceph:v15.2.13

   通过香港节点拉取镜像,推到阿里云个人仓库,再转到自建的Harbor仓库,也比较麻烦。

自建Harbor仓库(上一篇文章)

- 重新标记tag
docker tag quay.io/cephcsi/cephcsi:v3.3.1 hb.leayun.cn/rook-ceph/cephcsi:v3.3.1

- 推送到私有仓库
docker push hb.leayun.cn/rook-ceph/cephcsi:v3.3.1

2.2. 安装rook+ceph集群

2.2.1. 创建crd&common
  • 直接创建crd&common
kubectl create -f crds.yaml -f common.yaml 
2.2.2. 创建operator(rook)
  • 修改配置
vim operator.yaml
-------------------------------------

- 修改镜像(已同步到私有仓库)
  ROOK_CSI_CEPH_IMAGE: "hb.leayun.cn/rook-ceph/cephcsi:v3.3.1"
  ROOK_CSI_REGISTRAR_IMAGE: "hb.leayun.cn/rook-ceph/csi-node-driver-registrar:v2.0.1"
  ROOK_CSI_RESIZER_IMAGE: "hb.leayun.cn/rook-ceph/csi-resizer:v1.0.1"
  ROOK_CSI_PROVISIONER_IMAGE: "hb.leayun.cn/rook-ceph/csi-provisioner:v2.0.4"
  ROOK_CSI_SNAPSHOTTER_IMAGE: "hb.leayun.cn/rook-ceph/csi-snapshotter:v4.0.0"
  ROOK_CSI_ATTACHER_IMAGE: "hb.leayun.cn/rook-ceph/csi-attacher:v3.0.2"
  
- 开启自动发现磁盘(用于后期扩展)
  ROOK_ENABLE_DISCOVERY_DAEMON: "true"
  • 部署operator
kubectl create -f operator.yaml
2.2.3. 创建cluster(ceph)
  • 修改配置:等待operator容器和discover容器启动,配置osd节点
vim cluster.yaml
-------------------------------------

- 修改镜像
    image: hb.leayun.cn/rook-ceph/ceph:v15.2.13

- 改为false,并非使用所有节点所有磁盘作为osd
- 启用deviceFilter
- 按需配置config
- 会自动跳过非裸盘
  storage: # cluster level storage configuration and selection
    useAllNodes: false
    useAllDevices: false
    deviceFilter:
    config:
    nodes:
    - name: "k8s-node2"
      deviceFilter: "vdb"
    - name: "k8s-node3"
      deviceFilter: "vdb"
    - name: "k8s-node4"
      deviceFilter: "^vd."  #自动匹配vd开头的裸盘
  • 部署cluster
kubectl create -f cluster.yaml
  • 查看集群部署进度
- 实时查看pod创建进度
kubectl get pod -n rook-ceph -w

- 实时查看集群创建进度
kubectl get cephcluster -n rook-ceph rook-ceph -w

- 详细描述
kubectl describe cephcluster -n rook-ceph rook-ceph
  • 待osd-x的容器启动,表示安装成功

image-20210621183546501

image-20210621182916337

2.2.4. 添加裸盘测试
  • 给k8s-node4节点挂载一块裸盘

image-20210621185209968

  • 系统会自动检测到裸盘并创建osd:kubectl get cephcluster -n rook-ceph rook-ceph -w

image-20210621185259799

2.2.5. 安装ceph客户端工具
- 进入工作目录
cd /root/rook/cluster/examples/kubernetes/ceph

- 创建toolbox
kubectl  create -f toolbox.yaml -n rook-ceph

- 查看pod
kubectl  get pod -n rook-ceph -l app=rook-ceph-tools

- 进入pod
kubectl -n rook-ceph exec -it deploy/rook-ceph-tools -- bash

- 查看集群状态
ceph status

- 查看osd状态
ceph osd status

- 集群空间用量
ceph df
2.2.6. 暴露ceph-dashboard
  • 为ceph-dashboard创建一个NodePort(原service无法修改,因此需另建service)
vim rook-dashboard.yaml
--------------------------------------

apiVersion: v1
kind: Service
metadata:
  labels:
    app: rook-ceph-mgr
    ceph_daemon_id: a
    rook_cluster: rook-ceph
  name: rook-ceph-mgr-dashboard-np
  namespace: rook-ceph
spec:
  ports:
  - name: http-dashboard
    port: 7000
    protocol: TCP
    targetPort: 7000
    nodePort: 30700
  selector:
    app: rook-ceph-mgr
    ceph_daemon_id: a
  sessionAffinity: None
  type: NodePort
  

---------------------------
- 创建
kubectl apply -f rook-dashboard.yaml
  • 查看ceph-dashboard登录密码
kubectl -n rook-ceph get secret rook-ceph-dashboard-password -o jsonpath="{['data']['password']}" | base64 --decode && echo
  • 登录

image-20210621191239853

2.2.7. 消除HEALTH_WARN警告
  • 查看警告详情
    • AUTH_INSECURE_GLOBAL_ID_RECLAIM_ALLOWED: mons are allowing insecure global_id reclaim
    • MON_DISK_LOW: mons a,b,c are low on available space

image-20210621191403196

官方解决方案:https://docs.ceph.com/en/latest/rados/operations/health-checks/

  • AUTH_INSECURE_GLOBAL_ID_RECLAIM_ALLOWED
- 进入toolbox
kubectl -n rook-ceph exec -it deploy/rook-ceph-tools -- bash

ceph config set mon auth_allow_insecure_global_id_reclaim false
  • MON_DISK_LOW:根分区使用率过高,清理即可。

image-20210621195325516

3. Ceph存储使用

3.1. 三种存储类型

存储类型特征应用场景典型设备
块存储(RBD)存储速度较快
不支持共享存储 [ReadWriteOnce]
虚拟机硬盘硬盘
Raid
文件存储(CephFS)存储速度慢(需经操作系统处理再转为块存储)
支持共享存储 [ReadWriteMany]
文件共享FTP
NFS
对象存储(Object)具备块存储的读写性能和文件存储的共享特性
操作系统不能直接访问,只能通过应用程序级别的API访问
图片存储
视频存储
OSS

3.2. 块存储使用

3.2.1. 创建CephBlockPool和StorageClass
  • 文件路径:/root/rook/cluster/examples/kubernetes/ceph/csi/rbd/storageclass.yaml

  • CephBlockPool和StorageClass都位于storageclass.yaml 文件

  • 配置文件简要解读:

vim storageclass.yaml
-------------------------------------------

apiVersion: ceph.rook.io/v1
kind: CephBlockPool
metadata:
  name: replicapool
  namespace: rook-ceph
spec:
  failureDomain: host # host级容灾
  replicated:
    size: 3           # 默认三个副本
    requireSafeReplicaSize: true  # 强制高可用,如果size为1则需改为false

---
apiVersion: storage.k8s.io/v1
kind: StorageClass                       # sc无需指定命名空间
metadata:
  name: rook-ceph-block
provisioner: rook-ceph.rbd.csi.ceph.com   # 存储驱动
parameters:
  clusterID: rook-ceph # namespace:cluster
  pool: replicapool                       # 关联到CephBlockPool
  imageFormat: "2"
  imageFeatures: layering

  csi.storage.k8s.io/provisioner-secret-name: rook-csi-rbd-provisioner
  csi.storage.k8s.io/provisioner-secret-namespace: rook-ceph # namespace:cluster
  csi.storage.k8s.io/controller-expand-secret-name: rook-csi-rbd-provisioner
  csi.storage.k8s.io/controller-expand-secret-namespace: rook-ceph # namespace:cluster
  csi.storage.k8s.io/node-stage-secret-name: rook-csi-rbd-node
  csi.storage.k8s.io/node-stage-secret-namespace: rook-ceph # namespace:cluster
  
  csi.storage.k8s.io/fstype: ext4   

allowVolumeExpansion: true               # 是否允许扩容
reclaimPolicy: Delete                    # PV回收策略
  • 创建CephBlockPool和StorageClass
kubectl create -f storageclass.yaml
  • 查看
- 查看sc
kubectl get sc

- 查看CephBlockPool(也可在dashboard中查看)
kubectl get cephblockpools -n rook-ceph
3.2.2. 块存储使用示例
  • Deployment单副本+PersistentVolumeClaim
vim nginx-deploy-rbd.yaml
-------------------------------

apiVersion: apps/v1
kind: Deployment
metadata:
  labels:
    app: nginx-deploy-rbd
  name: nginx-deploy-rbd
spec:
  replicas: 1
  selector:
    matchLabels:
      app: nginx-deploy-rbd
  template:
    metadata:
      labels:
        app: nginx-deploy-rbd
    spec:
      containers:
      - image: nginx
        name: nginx
        volumeMounts:
        - name: data
          mountPath: /usr/share/nginx/html
      volumes:
      - name: data
        persistentVolumeClaim:
          claimName: nginx-rbd-pvc
---
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: nginx-rbd-pvc
spec:
  storageClassName: "rook-ceph-block"
  accessModes:
    - ReadWriteOnce
  resources:
    requests:
      storage: 5Gi
  • StatefulSet多副本+volumeClaimTemplates
vim nginx-ss-rbd.yaml
-------------------------------

apiVersion: apps/v1
kind: StatefulSet
metadata:
  name: nginx-ss-rbd
spec:
  selector:
    matchLabels:
      app: nginx-ss-rbd 
  serviceName: "nginx"
  replicas: 3 
  template:
    metadata:
      labels:
        app: nginx-ss-rbd 
    spec:
      containers:
      - name: nginx
        image: nginx
        ports:
        - containerPort: 80
          name: web
        volumeMounts:
        - name: www
          mountPath: /usr/share/nginx/html
  volumeClaimTemplates:
  - metadata:
      name: www
    spec:
      accessModes: [ "ReadWriteOnce" ]
      storageClassName: "rook-ceph-block"
      resources:
        requests:
          storage: 2Gi

image-20210622161110029

3.3. 共享文件存储

3.3.1. 部署MDS

   创建Cephfs文件系统需要先部署MDS服务,该服务负责处理文件系统中的元数据。

  • 文件路径:/root/rook/cluster/examples/kubernetes/ceph/filesystem.yaml
apiVersion: ceph.rook.io/v1
kind: CephFilesystem
metadata:
  name: myfs
  namespace: rook-ceph 
spec:
  metadataPool:
    replicated:
      size: 3                         # 元数据副本数
      requireSafeReplicaSize: true
    parameters:
      compression_mode:
        none
  dataPools:
    - failureDomain: host
      replicated:
        size: 3                     # 存储数据的副本数
        requireSafeReplicaSize: true
      parameters:
        compression_mode:
          none
  preserveFilesystemOnDelete: true
  metadataServer:
    activeCount: 1                # MDS实例的副本数,默认1,生产环境建议设置为3
3.3.2. 创建StorageClass
  • 文件路径:/root/rook/cluster/examples/kubernetes/ceph/csi/cephfs/storageclass.yaml
kubectl apply -f storageclass.yaml
3.3.3. 共享文件存储使用示例
  • Deployment多副本+ PersistentVolumeClaim
vim nginx-deploy-cephfs.yaml
-------------------------------

apiVersion: apps/v1
kind: Deployment
metadata:
  labels:
    app: nginx-deploy-cephfs
  name: nginx-deploy-cephfs
spec:
  replicas: 3
  selector:
    matchLabels:
      app: nginx-deploy-cephfs
  template:
    metadata:
      labels:
        app: nginx-deploy-cephfs
    spec:
      containers:
      - image: nginx
        name: nginx
        volumeMounts:
        - name: data
          mountPath: /usr/share/nginx/html
      volumes:
      - name: data
        persistentVolumeClaim:
          claimName: nginx-cephfs-pvc
---
apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: nginx-cephfs-pvc
spec:
  storageClassName: "rook-cephfs"
  accessModes:
    - ReadWriteMany
  resources:
    requests:
      storage: 1Gi

image-20210622164001844

4. 扩容、快照、克隆

后续补充

5. rook+ceph集群清理

参考官网:rook+ceph集群清理官方指南

5.1. 删除pvc、sc及对应的存储资源

- 按需删除pvc、pv
kubectl get pvc -n [namespace] | awk '{print $1};' | xargs kubectl delete pvc -n [namespace]
kubectl get pv | grep Released | awk '{print $1};' | xargs kubectl delete pv

- 删除块存储及SC
kubectl delete -n rook-ceph cephblockpool replicapool
kubectl delete storageclass rook-ceph-block

- 删除文件系统存储及SC
kubectl delete -n rook-ceph cephfilesystem myfs
kubectl delete storageclass rook-cephfs

- 删除对象存储

5.2. 删除ceph集群

kubectl -n rook-ceph patch cephcluster rook-ceph --type merge -p '{"spec":{"cleanupPolicy":{"confirmation":"yes-really-destroy-data"}}}'

kubectl -n rook-ceph delete cephcluster rook-ceph

- 若无法删除,手动终结
for CRD in $(kubectl get crd -n rook-ceph | awk '/ceph.rook.io/ {print $1}'); do
    kubectl get -n rook-ceph "$CRD" -o name | \
    xargs -I {} kubectl patch {} --type merge -p '{"metadata":{"finalizers": [null]}}' -n rook-ceph
done

- 查看ceph资源
kubectl -n rook-ceph get cephcluster

5.3. 删除rook集群crd

- 工作目录
cd /root/rook/cluster/examples/kubernetes/ceph

kubectl delete -f operator.yaml
kubectl delete -f common.yaml
kubectl delete -f crds.yaml

5.4. 检查资源是否清除完毕

kubectl api-resources --verbs=list --namespaced -o name \
  | xargs -n 1 kubectl get --show-kind --ignore-not-found -n rook-ceph

5.5. 清楚主机上的数据

5.5.1. 准备清除脚本
  • 参考官方示例,按需修改
vim rook-clean.sh
----------------------------

#!/usr/bin/env bash
DISK="/dev/vdb"
# Zap the disk to a fresh, usable state (zap-all is important, b/c MBR has to be clean)
# You will have to run this step for all disks.
sgdisk --zap-all $DISK
# Clean hdds with dd
dd if=/dev/zero of="$DISK" bs=1M count=100 oflag=direct,dsync
# Clean disks such as ssd with blkdiscard instead of dd
blkdiscard $DISK

# These steps only have to be run once on each node
# If rook sets up osds using ceph-volume, teardown leaves some devices mapped that lock the disks.
ls /dev/mapper/ceph-* | xargs -I% -- dmsetup remove %
# ceph-volume setup can leave ceph-<UUID> directories in /dev (unnecessary clutter)
rm -rf /dev/ceph-*

--------------------------
chmod +x rook-clean.sh
5.5.2. 编写Ansible剧本
  • hosts
[master]
172.16.16.11 node_name=k8s-master1
172.16.16.12 node_name=k8s-master2
172.16.16.13 node_name=k8s-master3

[node]
172.16.16.101 node_name=k8s-node1
172.16.16.102 node_name=k8s-node2
172.16.16.103 node_name=k8s-node3
172.16.16.104 node_name=k8s-node4

[osd]
172.16.16.102 node_name=k8s-node2
172.16.16.103 node_name=k8s-node3
172.16.16.104 node_name=k8s-node4
  • delete-old-rookfile.yml
- hosts:
  - master
  - node
  tasks:
  - name: 清除rook目录
    shell: rm -rf /var/lib/rook/*
  • reset-rook-osd.yml
- hosts:
  - osd
  tasks:
  - name: 分发清理脚本
    copy: src=/root/ansible-install-k8s-master-v1.20/rook-clean.sh dest=~/ mode=u+x
  - name: 执行清理脚本
    shell: /bin/bash ~/rook-clean.sh
  • 执行
ansible-playbook -i hosts delete-old-rookfile.yml -uroot -k
ansible-playbook -i hosts reset-rook-osd.yml -uroot -k

若本篇内容对您有所帮助,请三连点赞,关注,收藏支持下~

评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值