作者:申红磊,QingCloud 容器解决方案架构师,开源项目爱好者,KubeSphere Member
在正式阅读本文之前,先友情提醒一下:不建议您在生产环境中使用 NFS 存储(特别是 Kubernetes 1.20 或以上版本),原因如下:
- selfLink was empty 在 K8s 集群 v1.20 之前都存在,在 v1.20 之后被删除问题。
- 还有可能引起 failed to obtain lock 和 input/output error 等问题,从而导致 Pod CrashLoopBackOff。此外,部分应用不兼容 NFS,例如 Prometheus 等。
安装 NFS Server
#安装 NFS 服务器端
$ sudo apt-get update #执行以下命令确保使用最新软件包
$ sudo apt-get install nfs-kernel-server
#安装 NFS 客户端
$ sudo apt-get install nfs-common
# yum
$ yum install -y nfs-utils
创建共享目录
先查看配置文件 /etc/exports:
$ cat /etc/exports
# /etc/exports: the access control list for filesystems which may be exported
# to NFS clients. See exports(5).
#
# Example for NFSv2 and NFSv3:
# /srv/homes hostname1(rw,sync,no_subtree_check) hostname2(ro,sync,no_subtree_check)
#
# Example for NFSv4:
# /srv/nfs4 gss/krb5i(rw,sync,fsid=0,crossmnt,no_subtree_check)
# /srv/nfs4/homes gss/krb5i(rw,sync,no_subtree_check)
创建共享目标并赋权:
#目录ksha3.2
$ mkdir /ksha3.2
$ chmod 777 /ksha3.2
#目录demo
$ mkdir /demo
$ chmod 777 /demo
#目录/home/ubuntu/nfs/ks3.1
$ mkdir /home/ubuntu/nfs/ks3.1
$ chmod 777 /home/ubuntu/nfs/ks3.1
添加到配置文件中 /etc/exports:
$ vi /etc/exports
....
/home/ubuntu/nfs/ks3.1 *(rw,sync,no_subtree_check)
/mount/ksha3.2 *(rw,sync,no_subtree_check)
/mount/demo *(rw,sync,no_subtree_check)
#/mnt/ks3.2 139.198.186.39(insecure,rw,sync,anonuid=500,anongid=500)
#/mnt/demo 139.198.167.103(rw,sync,no_subtree_check)
#此文件的配置格式为:<输出目录> [客户端1 选项(访问权限,用户映射,其他)] [客户端2 选项(访问权限,用户映射,其他)]
注意:如果共享目录创建无效或者遗漏创建,使用时会报异常如下:
$ mount.nfs: access denied by server while mounting 139.198.168.114:/mnt/demo1
#客户端挂载 nfs 共享目录的时候提示 mount.nfs: access denied by server while mounting。
#问题原因:服务器端的共享目录没有设置允许其他人访问的权限,或者客户端挂载的目录没有权限。
#解决办法:在服务器端修改共享目录的权限,成功连接。
验证
#更新配置文件,重新加载 /etc/exports 的配置:
$ exportfs -rv
#在 nfs server 上测试 查看本机共享的目录:
$ showmount -e 127.0.0.1
$ showmount -e localhost
$ showmount -e 127.0.0.1
Export list for 127.0.0.1:
/mount/demo *
/mount/ksha3.2 *
/home/ubuntu/nfs/ks3.1 *
使用
在其它网络通的机器上,使用 NFS 共享目录。
安装客户端
如果不安装客户端会报:bad option; for several filesystems (e.g. nfs, cifs) you might need a /sbin/mount. helper program.
#apt-get
$ apt-get install nfs-common
#yum
$ yum install nfs-utils
在用户机器上进行测试:
#本机创建挂载目录
#目录ksha3.2
$ mkdir /root/aa
$ chmod 777 /root/aa
#使用 nfs client 上测试,将 /root/aa 目录挂载到远端 139.198.168.114 的共享存储目录/mount/ksha3.2上
$ mount -t nfs 139.198.168.114:/mount/ksha3.2 /root/aa
查看:
#使用 df -h 查看
$ df -h | grep aa
139.198.168.114:/mount/ksha3.2 146G 5.3G 140G 4% /root/aa
解绑:
$ umount -t nfs 139.198.168.114:/mnt/ksha3.2 /root/aa
#如果卸载的时候提示:umount:/mnt:device is busy;解决方法:需要退出挂载目录再进行卸载,或者是否NFS server宕机了
#需要强制卸载:mount –lf /mnt
#此命令也可以:fuser –km /mnt 不建议用
KubeSphere 对接 NFS 动态分配器
可安装 NFS 客户端程序,为了方便用户对接 NFS 服务端,KubeSphere 可安装 NFS 动态分配器,支持动态分配存储卷,分配和回收存储卷过程简便,可对接一个或者多个 NFS 服务端。
当然也可以使用 Kubernetes 官方方法对接 NFS 服务端,这是一种静态分配存储卷方法,分配和回收存储卷过程复杂,可对接多个 NFS 服务端。
前提条件
用户对接 NFS 服务端时应确保 KubeSphere 各节点有权限挂载 NFS 服务端文件夹。
操作步骤
以下步骤示例中,NFS 服务端 IP 为 139.198.168.114,NFS 共享文件夹为 /mnt/ksha3.2。
安装 NFS 动态分配器
首先请在这里下载 rbac.yaml,或者直接执行命令:
$ kubectl apply -f https://raw.githubusercontent.com/kubernetes-incubator/external-storage/master/nfs-client/deploy/rbac.yaml
serviceaccount/nfs-client-provisioner created
clusterrole.rbac.authorization.k8s.io/nfs-client-provisioner-runner created
clusterrolebinding.rbac.authorization.k8s.io/run-nfs-client-provisioner created
role.rbac.authorization.k8s.io/leader-locking-nfs-client-provisioner created
rolebinding.rbac.authorization.k8s.io/leader-locking-nfs-client-provisioner created
然后下载官方的 nfs provisoner 用途的 deployment。deployment 文件中有几处,请根据自己的情况做修改:
apiVersion: apps/v1
kind: Deployment
metadata:
name: nfs-client-provisioner
labels:
app: nfs-client-provisioner
# replace with namespace where provisioner is deployed
namespace: default
spec:
replicas: 1
strategy:
type: Recreate
selector:
matchLabels:
app: nfs-client-provisioner
template:
metadata:
labels:
app: nfs-client-provisioner
spec:
serviceAccountName: nfs-client-provisioner
containers:
- name: nfs-client-provisioner
image: quay.io/external_storage/nfs-client-provisioner:latest
volumeMounts:
- name: nfs-client-root
mountPath: /persistentvolumes
env:
- name: PROVISIONER_NAME
value: nfs/provisioner-229
- name: NFS_SERVER
value: 139.198.168.114
- name: NFS_PATH
value: /mount/ksha3.2
volumes:
- name: nfs-client-root
nfs:
server: 139.198.168.114
path: /mount/ksha3.2
执行创建:
$ kubectl apply -f deployment.yaml
deployment.apps/nfs-client-provisioner created
接下来下载官方的 class.yaml,然后创建 StorageClass,需要根据实际情况修改参数:
#请根据上方deployment部署时候的provisioner_name做对应的修改,或者没有修改,就不用动
apiVersion: storage.k8s.io/v1
kind: StorageClass
metadata:
name: managed-nfs-storage
annotations:
"storageclass.kubernetes.io/is-default-class": "false"
provisioner: nfs/provisioner-229 # or choose another name, must match deployment's env PROVISIONER_NAME'
parameters:
archiveOnDelete: "false"
执行创建:kubectl apply -f storageclass.yaml
。
如果想让这个 NFS 作为默认的 Provisioner, 那么就添加如下的 annotation:
annotations:
"storageclass.kubernetes.io/is-default-class": "true"
或者标记一个 StorageClass 为默认的 StorageClass, 你需要添加 / 设置注解 storageclass.kubernetes.io/is-default-class=true
。
$ kubectl patch storageclass <your-class-name> -p '{"metadata": {"annotations":{"storageclass.kubernetes.io/is-default-class":"true"}}}'
# 检查集群中是否已存在 default Storage Class
$ kubectl get sc
NAME PROVISIONER AGE
glusterfs (default) kubernetes.io/glusterfs 3d4h
验证安装结果
执行以下命令,查看 NFS 动态分配器容器组是否正常运行。
$ kubectl get po -A | grep nfs-client
default nfs-client-provisioner-7d69b9f45f-ks94m 1/1 Running 0 9m3s
查看 NFS 存储类型
$ kubectl get sc managed-nfs-storage
NAME PROVISIONER RECLAIMPOLICY VOLUMEBINDINGMODE ALLOWVOLUMEEXPANSION AGE
managed-nfs-storage nfs/provisioner-229 Delete Immediate false 6m28s
创建和挂载 NFS 存储卷
现在可以通过动态创建 NFS 存储卷和工作负载挂载 NFS 存储卷了。
kind: PersistentVolumeClaim
apiVersion: v1
metadata:
name: demo4nfs
namespace: ddddd
annotations:
kubesphere.io/creator: admin
volume.beta.kubernetes.io/storage-provisioner: nfs/provisioner-229
finalizers:
- kubernetes.io/pvc-protection
spec:
accessModes:
- ReadWriteOnce
resources:
requests:
storage: 10Gi
storageClassName: managed-nfs-storage
volumeMode: Filesystem
unexpected error getting claim reference: selfLink was empty, can’t make reference
问题现象
使用 NFS 创建 PV 时,PVC 一直是处于 Pending 状态。
查看 PVC:
$ kubectl get pvc -n ddddd
NAME STATUS VOLUME CAPACITY ACCESS MODES STORAGECLASS AGE
demo4nfs Bound pvc-a561ce85-fc0d-42af-948e-6894ac000264 10Gi RWO managed-nfs-storage 32m
查看详细信息:
#查看当前pvc的状态信息,发现是在等待volume的创建
$ kubectl get pvc -n ddddd
查看 nfs-client-provisioner 的日志,是 seltlink was empty 的问题,selfLink was empty 在 K8s 集群 v1.20 之前都存在,在 v1.20 之后被删除,需要在 /etc/kubernetes/manifests/kube-apiserver.yaml
中添加参数。
$ kubectl get pod -n default
NAME READY STATUS RESTARTS AGE
nfs-client-provisioner-7d69b9f45f-ks94m 1/1 Running 0 27m
$ kubectl logs -f nfs-client-provisioner-7d69b9f45f-ks94m
I0622 09:41:33.606000 1 leaderelection.go:185] attempting to acquire leader lease default/nfs-provisioner-229...
E0622 09:41:33.612745 1 event.go:259] Could not construct reference to: '&v1.Endpoints{TypeMeta:v1.TypeMeta{Kind:"", APIVersion:""}, ObjectMeta:v1.ObjectMeta{Name:"nfs-provisioner-229", GenerateName:"", Namespace:"default", SelfLink:"", UID:"e8f19e28-f17f-4b22-9bb8-d4cbe20c796b", ResourceVersion:"23803580", Generation:0, CreationTimestamp:v1.Time{Time:time.Time{wall:0x0, ext:63791487693, loc:(*time.Location)(0x1956800)}}, DeletionTimestamp:(*v1.Time)(nil), DeletionGracePeriodSeconds:(*int64)(nil), Labels:map[string]string(nil), Annotations:map[string]string{"control-plane.alpha.kubernetes.io/leader":"{\"holderIdentity\":\"nfs-client-provisioner-7d69b9f45f-ks94m_8081b417-f20f-11ec-bff3-f64a9f402fda\",\"leaseDurationSeconds\":15,\"acquireTime\":\"2022-06-22T09:41:33Z\",\"renewTime\":\"2022-06-22T09:41:33Z\",\"leaderTransitions\":0}"}, OwnerReferences:[]v1.OwnerReference(nil), Initializers:(*v1.Initializers)(nil), Finalizers:[]string(nil), ClusterName:""}, Subsets:[]v1.EndpointSubset(nil)}' due to: 'selfLink was empty, can't make reference'. Will not report event: 'Normal' 'LeaderElection' 'nfs-client-provisioner-7d69b9f45f-ks94m_8081b417-f20f-11ec-bff3-f64a9f402fda became leader'
I0622 09:41:33.612829 1 leaderelection.go:194] successfully acquired lease default/nfs-provisioner-229
I0622 09:41:33.612973 1 controller.go:631] Starting provisioner controller nfs/provisioner-229_nfs-client-provisioner-7d69b9f45f-ks94m_8081b417-f20f-11ec-bff3-f64a9f402fda!
I0622 09:41:33.713170 1 controller.go:680] Started provisioner controller nfs/provisioner-229_nfs-client-provisioner-7d69b9f45f-ks94m_8081b417-f20f-11ec-bff3-f64a9f402fda!
I0622 09:53:33.461902 1 controller.go:987] provision "ddddd/demo4nfs" class "managed-nfs-storage": started
E0622 09:53:33.464213 1 controller.go:1004] provision "ddddd/demo4nfs" class "managed-nfs-storage": unexpected error getting claim reference: selfLink was empty, can't make reference
I0622 09:56:33.623717 1 controller.go:987] provision "ddddd/demo4nfs" class "managed-nfs-storage": started
E0622 09:56:33.625852 1 controller.go:1004] provision "ddddd/demo4nfs" class "managed-nfs-storage": unexpected error getting claim reference: selfLink was empty, can't make reference
解决办法
在 kube-apiserver.yaml 文件中添加参数 - --feature-gates=RemoveSelfLink=false
。
#使用命令查找一下 kube-apiserver.yaml的位置
$ find / -name kube-apiserver.yaml
/data/kubernetes/manifests/kube-apiserver.yaml
#在文件中添加 - --feature-gates=RemoveSelfLink=false ,如下图
$ cat /data/kubernetes/manifests/kube-apiserver.yaml
apiVersion: v1
kind: Pod
metadata:
annotations:
kubeadm.kubernetes.io/kube-apiserver.advertise-address.endpoint: 192.168.100.25:6443
creationTimestamp: null
labels:
component: kube-apiserver
tier: control-plane
name: kube-apiserver
namespace: kube-system
spec:
containers:
- command:
- kube-apiserver
- --feature-gates=RemoveSelfLink=false
- --advertise-address=0.0.0.0
- --allow-privileged=true
- --authorization-mode=Node,RBAC
- --client-ca-file=/etc/kubernetes/pki/ca.crt
- --enable-admission-plugins=NodeRestriction
#添加之后使用kubeadm部署的集群会自动加载部署pod
#kubeadm安装的apiserver是Static Pod,它的配置文件被修改后,立即生效。
#Kubelet 会监听该文件的变化,当您修改了 /etc/kubenetes/manifest/kube-apiserver.yaml 文件之后,kubelet 将自动终止原有的 #kube-apiserver-{nodename} 的 Pod,并自动创建一个使用了新配置参数的 Pod 作为替代。
#如果您有多个 Kubernetes Master 节点,您需要在每一个 Master 节点上都修改该文件,并使各节点上的参数保持一致。
#这里需注意如果api-server启动失败 需重新在执行一遍
$ kubectl apply -f /etc/kubernetes/manifests/kube-apiserver.yaml