longhorn默认存储目录/var/lib/longhorn，使用yaml文件安装也在节点目录下

longhorn它将创建自己的命名空间和存储类别

每个节点都需要安装iscsi-initiator-utils，ubuntu需要安装open-iscsi yum -y install iscsi-initiator-utils 

您可以下载指定版本。也可以直接安装这里下载修改配置 #kubectl apply -f https://raw.githubusercontent.com/longhorn/longhorn/v1.1.0/deploy/longhorn.yaml wget https://raw.githubusercontent.com/longhorn/longhorn/v1.1.0/deploy/longhorn.yaml 

1. 修改pv保留策略

修改pv为了保留策略，否则删除pvc的时候pv就会被删除 #在文件中找到StorageClass.data.reclaimPolicy修改为字段Retain reclaimPolicy: Retain 

2. 修改longhorn-frontend端口，方便访问

修改longhorn-frontend端口 #类型也需要修改NodePort type: NodePort #找到longhorn-frontend的nodePort字段 nodePort: 30179 

3. 修改存储路径

#找到DaemonSet资源下的spec.template.containers.volumeMounts字段 将/var/lib/longhorn/修改为需要更改的存储路径，记将文件中的所有/var/lib/longhorn//用变更的存储路径替换此路径 

1. 部署

[root@master01 volumes]# kubectl apply -f longhorn.yaml  namespace/longhorn-system created serviceaccount/longhorn-service-account created clusterrole.rbac.authorization.k8s.io/longhorn-role created clusterrolebinding.rbac.authorization.k8s.io/longhorn-bind created customresourcedefinition.apiextensions.k8s.io/engines.longhorn.io created customresourcedefinition.apiextensions.k8s.io/replicas.longhorn.io created customresourcedefinition.apiextensions.k8s.io/settings.longhorn.io created customresourcedefinition.apiextensions.k8s.io/volumes.longhorn.io created customresourcedefinition.apiextensions.k8s.io/engineimages.longhorn.io created customresourcedefinition.apiextensions.k8s.io/nodes.longhorn.io created customresourcedefinition.apiextensions.k8s.io/instancemanagers.longhorn.io created customresourcedefinition.apiextensions.k8s.io/sharemanagers.longhorn.io created configmap/longhorn-default-setting created podsecuritypolicy.policy/longhorn-psp created role.rbac.authorization.k8s.io/longhorn-psp-role created rolebinding.rbac.authorization.k8s.io/longhorn-psp-binding created configmap/longhorn-storageclass created daemonset.apps/longhorn-manager created service/longhorn-backend created deployment.apps/longhorn-ui created service/longhorn-frontend created deployment.apps/longhorn-driver-deployer created 

2. 查看部署结果

[root@master01 volumes]# kubectl get ns NAME              STATUS   AGE default           Active   105d dev               Active   10d kube-node-lease   Active   105d kube-public       Active   105d kube-system       Active   105d longhorn-system   Active   3m8s  [root@master01 volumes]# kubectl get pod -n longhorn-system NAME                                       READY   STATUS    RESTARTS   AGE si-attacher-5dcdcd5984-cvk6k              1/1     Running   0          2m38s
csi-attacher-5dcdcd5984-fp8q4              1/1     Running   0          2m38s
csi-attacher-5dcdcd5984-kjjg9              1/1     Running   0          2m38s
csi-provisioner-5c9dfb6446-df4qw           1/1     Running   0          2m38s
csi-provisioner-5c9dfb6446-dhjhn           1/1     Running   0          2m38s
csi-provisioner-5c9dfb6446-szcml           1/1     Running   0          2m38s
csi-resizer-54d484bf8-ggjxg                1/1     Running   0          2m38s
csi-resizer-54d484bf8-jg9p9                1/1     Running   0          2m38s
csi-resizer-54d484bf8-k74gp                1/1     Running   0          2m38s
csi-snapshotter-96bfff7c9-555kp            1/1     Running   0          2m37s
csi-snapshotter-96bfff7c9-dcc96            1/1     Running   0          2m37s
csi-snapshotter-96bfff7c9-pd8hb            1/1     Running   0          2m37s
engine-image-ei-cf743f9c-fkgx8             1/1     Running   0          3m2s
engine-image-ei-cf743f9c-rtgml             1/1     Running   0          3m2s
engine-image-ei-cf743f9c-zhdtv             1/1     Running   0          3m2s
instance-manager-e-206640f3                1/1     Running   0          2m58s
instance-manager-e-d25b6c2c                1/1     Running   0          2m54s
instance-manager-e-e04840b1                1/1     Running   0          3m1s
instance-manager-r-2e491629                1/1     Running   0          3m1s
instance-manager-r-384e572d                1/1     Running   0          2m57s
instance-manager-r-70cc0617                1/1     Running   0          2m53s
longhorn-csi-plugin-bb2bw                  2/2     Running   0          2m37s
longhorn-csi-plugin-mmchb                  2/2     Running   0          2m37s
longhorn-csi-plugin-vrpq6                  2/2     Running   0          2m37s
longhorn-driver-deployer-cdb7464c6-cwt98   1/1     Running   0          3m17s
longhorn-manager-52gn6                     1/1     Running   2          3m18s
longhorn-manager-597xc                     1/1     Running   2          3m18s
longhorn-manager-zltr8                     1/1     Running   1          3m18s
longhorn-ui-849c455d79-x9265               1/1     Running   0          3m17s

查看存储类
[root@master01 volumes]# kubectl get sc
NAME       PROVISIONER          RECLAIMPOLICY   VOLUMEBINDINGMODE   ALLOWVOLUMEEXPANSION   AGE
longhorn   driver.longhorn.io   Retain          Immediate           true                   3m44s

直接访问集群中任何一个节点加端口就可以了 http://192.168.8.10:30179/#/dashboard 可以手动创建pv，部署好这个服务之后，之后在longhorn存储类中创建的pvc如果没有匹配的pv，都会自动创建pv，与pvc想匹配的。所以不需要咱们手动创建