remount old mount point when csi plugin unexpect exit

cmd/cephfs/main.go

@@ -31,6 +31,7 @@ var (
 	nodeID          = flag.String("nodeid", "", "node id")
 	volumeMounter   = flag.String("volumemounter", "", "default volume mounter (possible options are 'kernel', 'fuse')")
 	metadataStorage = flag.String("metadatastorage", "", "metadata persistence method [node|k8s_configmap]")
+	mountCacheDir   = flag.String("mountcachedir", "", "mount info cache save dir")
 )
 
 func init() {
@@ -56,7 +57,7 @@ func main() {
 	}
 
 	driver := cephfs.NewDriver()
-	driver.Run(*driverName, *nodeID, *endpoint, *volumeMounter, cp)
+	driver.Run(*driverName, *nodeID, *endpoint, *volumeMounter, *mountCacheDir, cp)
 
 	os.Exit(0)
 }

deploy/cephfs/helm/templates/nodeplugin-clusterrole.yaml

@@ -10,6 +10,9 @@ metadata:
     release: {{ .Release.Name }}
     heritage: {{ .Release.Service }}
 rules:
+  - apiGroups: [""]
+    resources: ["configmaps"]
+    verbs: ["get", "list"]
   - apiGroups: [""]
     resources: ["nodes"]
     verbs: ["get", "list", "update"]

deploy/cephfs/helm/templates/nodeplugin-daemonset.yaml

@@ -70,6 +70,7 @@ spec:
             - "--v=5"
             - "--drivername=$(DRIVER_NAME)"
             - "--metadatastorage=k8s_configmap"
+            - "--mountcachedir=/mount-cache-dir"
           env:
             - name: HOST_ROOTFS
               value: "/rootfs"
@@ -83,6 +84,8 @@ spec:
               value: "unix:/{{ .Values.socketDir }}/{{ .Values.socketFile }}"
           imagePullPolicy: {{ .Values.nodeplugin.plugin.image.imagePullPolicy }}
           volumeMounts:
+            - name: mount-cache-dir
+              mountPath: /mount-cache-dir
             - name: plugin-dir
               mountPath: {{ .Values.socketDir }}
             - name: pods-mount-dir
@@ -103,6 +106,8 @@ spec:
           resources:
 {{ toYaml .Values.nodeplugin.plugin.resources | indent 12 }}
       volumes:
+        - name: mount-cache-dir
+          emptyDir: {}
         - name: plugin-dir
           hostPath:
             path: {{ .Values.socketDir }}

deploy/cephfs/kubernetes/csi-cephfsplugin.yaml

@@ -55,6 +55,7 @@ spec:
             - "--v=5"
             - "--drivername=cephfs.csi.ceph.com"
             - "--metadatastorage=k8s_configmap"
+            - "--mountcachedir=/mount-cache-dir"
           env:
             - name: NODE_ID
               valueFrom:
@@ -68,6 +69,8 @@ spec:
               value: unix:///csi/csi.sock
           imagePullPolicy: "IfNotPresent"
           volumeMounts:
+            - name: mount-cache-dir
+              mountPath: /mount-cache-dir
             - name: plugin-dir
               mountPath: /csi
             - name: csi-plugins-dir
@@ -84,6 +87,8 @@ spec:
             - name: host-dev
               mountPath: /dev
       volumes:
+        - name: mount-cache-dir
+          emptyDir: {}
         - name: plugin-dir
           hostPath:
             path: /var/lib/kubelet/plugins/cephfs.csi.ceph.com/

deploy/cephfs/kubernetes/csi-nodeplugin-rbac.yaml

@@ -10,6 +10,9 @@ apiVersion: rbac.authorization.k8s.io/v1
 metadata:
   name: cephfs-csi-nodeplugin
 rules:
+  - apiGroups: [""]
+    resources: ["configmaps"]
+    verbs: ["get", "list"]
   - apiGroups: [""]
     resources: ["nodes"]
     verbs: ["get", "list", "update"]

docs/deploy-cephfs.md

@@ -34,6 +34,7 @@ Option              | Default value         | Description
 `--nodeid`          | _empty_               | This node's ID
 `--volumemounter`   | _empty_               | default volume mounter. Available options are `kernel` and `fuse`. This is the mount method used if volume parameters don't specify otherwise. If left unspecified, the driver will first probe for `ceph-fuse` in system's path and will choose Ceph kernel client if probing failed.
 `--metadatastorage` | _empty_               | Whether metadata should be kept on node as file or in a k8s configmap (`node` or `k8s_configmap`)
+`--mountcachedir` | _empty_               | volume mount cache info save dir. If left unspecified, the dirver will not record mount info, or it will save mount info and when driver restart it will remount volume it cached.
 
 **Available environmental variables:**
 

pkg/cephfs/driver.go

@@ -77,7 +77,7 @@ func NewNodeServer(d *csicommon.CSIDriver) *NodeServer {
 
 // Run start a non-blocking grpc controller,node and identityserver for
 // ceph CSI driver which can serve multiple parallel requests
-func (fs *Driver) Run(driverName, nodeID, endpoint, volumeMounter string, cachePersister util.CachePersister) {
+func (fs *Driver) Run(driverName, nodeID, endpoint, volumeMounter, mountCacheDir string, cachePersister util.CachePersister) {
 	klog.Infof("Driver: %v version: %v", driverName, version)
 
 	// Configuration
@@ -105,6 +105,13 @@ func (fs *Driver) Run(driverName, nodeID, endpoint, volumeMounter string, cacheP
 		klog.Fatalf("failed to write ceph configuration file: %v", err)
 	}
 
+	initVolumeMountCache(driverName, mountCacheDir, cachePersister)
+	if mountCacheDir != "" {
+		if err := remountCachedVolumes(); err != nil {
+			klog.Warningf("failed to remount cached volumes: %v", err)
+			//ignore remount fail
+		}
+	}
 	// Initialize default library driver
 
 	fs.cd = csicommon.NewCSIDriver(driverName, version, nodeID)