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。
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。
2.1.1. 建议配置
环境 | 测试 | 生产环境 |
---|---|---|
服务器配置(≥3) | 4核8G | 8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
kubectl describe node |grep Taint
lsblk -f
,必须为裸盘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。
- 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
2.2.1. 创建crd&common
kubectl create -f crds.yaml -f common.yaml
2.2.2. 创建operator(rook)
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"
kubectl create -f operator.yaml
2.2.3. 创建cluster(ceph)
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
kubectl create -f cluster.yaml
- 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
2.2.4. 添加裸盘测试
kubectl get cephcluster -n rook-ceph rook-ceph -w
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
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
kubectl -n rook-ceph get secret rook-ceph-dashboard-password -o jsonpath="{['data']['password']}" | base64 --decode && echo
2.2.7. 消除HEALTH_WARN警告
官方解决方案:https://docs.ceph.com/en/latest/rados/operations/health-checks/
- 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
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 |
3.2.1. 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
kubectl create -f storageclass.yaml
- 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
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
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
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。
/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
/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
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
后续补充
参考官网:rook+ceph集群清理官方指南
- 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
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
- 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
kubectl api-resources --verbs=list --namespaced -o name \
| xargs -n 1 kubectl get --show-kind --ignore-not-found -n rook-ceph
5.5.1. 准备清除脚本
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- directories in /dev (unnecessary clutter)
rm -rf /dev/ceph-*
--------------------------
chmod +x rook-clean.sh
5.5.2. 编写Ansible剧本
[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
- hosts:
- master
- node
tasks:
- name: 清除rook目录
shell: rm -rf /var/lib/rook/*
- 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
ansible-playbook -i hosts delete-old-rookfile.yml -uroot -k
ansible-playbook -i hosts reset-rook-osd.yml -uroot -k
转载至https://blog.csdn.net/qq_14999375/article/details/118110564?ops_request_misc=%257B%2522request%255Fid%2522%253A%2522162839308616780264088666%2522%252C%2522scm%2522%253A%252220140713.130102334.pc%255Fall.%2522%257D&request_id=162839308616780264088666&biz_id=0&utm_medium=distribute.pc_search_result.none-task-blog-2~all~first_rank_v2~rank_v29-2-118110564.first_rank_v2_pc_rank_v29&utm_term=rook+ceph+ROOK_CSI_REGISTRAR_IMAGE&spm=1018.2226.3001.4187