ceph-csi/pkg/cephfs/volumemounter.go

263 lines
6.6 KiB
Go
Raw Normal View History

/*
Copyright 2018 The Ceph-CSI Authors.
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
*/
package cephfs
import (
"context"
"errors"
"fmt"
"os"
"os/exec"
2019-02-27 19:29:20 +00:00
"regexp"
"strconv"
"strings"
"sync"
"github.com/ceph/ceph-csi/pkg/util"
"k8s.io/klog"
)
const (
volumeMounterFuse = "fuse"
volumeMounterKernel = "kernel"
)
var (
availableMounters []string
// maps a mountpoint to PID of its FUSE daemon
fusePidMap = make(map[string]int)
fusePidMapMtx sync.Mutex
2019-02-27 19:29:20 +00:00
fusePidRx = regexp.MustCompile(`(?m)^ceph-fuse\[(.+)\]: starting fuse$`)
)
// Load available ceph mounters installed on system into availableMounters
// Called from driver.go's Run()
func loadAvailableMounters(conf *util.Config) error {
// #nosec
fuseMounterProbe := exec.Command("ceph-fuse", "--version")
// #nosec
kernelMounterProbe := exec.Command("mount.ceph")
err := kernelMounterProbe.Run()
if err == nil {
kernelVersion, _, err := execCommand(context.TODO(), "uname", "-r")
if err != nil {
return err
}
vers := strings.Split(string(kernelVersion), ".")
majorVers, err := strconv.Atoi(vers[0])
if err != nil {
return err
}
minorVers, err := strconv.Atoi(vers[1])
if err != nil {
return err
}
if conf.ForceKernelCephFS || majorVers > 4 || (majorVers >= 4 && minorVers >= 17) {
klog.Infof("loaded mounter: %s", volumeMounterKernel)
availableMounters = append(availableMounters, volumeMounterKernel)
} else {
klog.Infof("kernel version < 4.17 might not support quota feature, hence not loading kernel client")
}
}
if fuseMounterProbe.Run() == nil {
klog.Infof("loaded mounter: %s", volumeMounterFuse)
availableMounters = append(availableMounters, volumeMounterFuse)
}
if len(availableMounters) == 0 {
return errors.New("no ceph mounters found on system")
}
return nil
}
type volumeMounter interface {
mount(ctx context.Context, mountPoint string, cr *util.Credentials, volOptions *volumeOptions) error
name() string
}
func newMounter(volOptions *volumeOptions) (volumeMounter, error) {
// Get the mounter from the configuration
wantMounter := volOptions.Mounter
// Verify that it's available
var chosenMounter string
for _, availMounter := range availableMounters {
if availMounter == wantMounter {
chosenMounter = wantMounter
break
}
}
if chosenMounter == "" {
// Otherwise pick whatever is left
chosenMounter = availableMounters[0]
klog.Infof("requested mounter: %s, chosen mounter: %s", wantMounter, chosenMounter)
}
// Create the mounter
switch chosenMounter {
case volumeMounterFuse:
return &fuseMounter{}, nil
case volumeMounterKernel:
return &kernelMounter{}, nil
}
return nil, fmt.Errorf("unknown mounter '%s'", chosenMounter)
}
type fuseMounter struct{}
func mountFuse(ctx context.Context, mountPoint string, cr *util.Credentials, volOptions *volumeOptions) error {
Make CephFS plugin stateless reusing RADOS based journal scheme This is a part of the stateless set of commits for CephCSI. This commit removes the dependency on config maps to store cephFS provisioned volumes, and instead relies on RADOS based objects and keys, and required CSI VolumeID encoding to detect the provisioned volumes. Changes: - Provide backward compatibility to provisioned volumes by older plugin versions (1.0.0 or older) - Remove Create/Delete support for statically provisioned volumes (fixes #382) - Added namespace support to RADOS OMaps and used the same to store RADOS CSI objects and keys in the CephFS metadata pool - Added support to mention fsname for CephFS provisioning (fixes #359) - Changed field name in CSI Identifier to 'location', to denote a pool or fscid - Updated mounter cache to use new scheme - Required Helm manifests are updated - Required documentation and other manifests are updated - Made driver option 'metadatastorage' as optional, as fresh installs do not need to specify the same Testing done: - Create/Mount/Delete PVC - Create/Delete 5 PVCs - Mount version 1.0.0 PVC - Delete version 1.0.0 PV - Mount Statically defined PV/PVC/Pod - Mount Statically defined version 1.0.0 PV/PVC/Pod - Delete Statically defined version 1.0.0 PV/PVC/Pod - Node restart when mounted to test mountcache - Use InstanceID other than 'default' - RBD basic round of tests, as namespace is added to OMaps - csitest against ceph-fs plugin - NOTE: CephFS plugin still does not detect and address already created volumes but of a different size - Test not providing any value to the metadata storage parameter Signed-off-by: ShyamsundarR <srangana@redhat.com>
2019-05-28 19:03:18 +00:00
args := []string{
mountPoint,
2019-02-13 12:57:16 +00:00
"-m", volOptions.Monitors,
"-c", util.CephConfigPath,
"-n", cephEntityClientPrefix + cr.ID, "--keyfile=" + cr.KeyFile,
"-r", volOptions.RootPath,
"-o", "nonempty",
}
if volOptions.FuseMountOptions != "" {
args = append(args, ","+volOptions.FuseMountOptions)
}
Make CephFS plugin stateless reusing RADOS based journal scheme This is a part of the stateless set of commits for CephCSI. This commit removes the dependency on config maps to store cephFS provisioned volumes, and instead relies on RADOS based objects and keys, and required CSI VolumeID encoding to detect the provisioned volumes. Changes: - Provide backward compatibility to provisioned volumes by older plugin versions (1.0.0 or older) - Remove Create/Delete support for statically provisioned volumes (fixes #382) - Added namespace support to RADOS OMaps and used the same to store RADOS CSI objects and keys in the CephFS metadata pool - Added support to mention fsname for CephFS provisioning (fixes #359) - Changed field name in CSI Identifier to 'location', to denote a pool or fscid - Updated mounter cache to use new scheme - Required Helm manifests are updated - Required documentation and other manifests are updated - Made driver option 'metadatastorage' as optional, as fresh installs do not need to specify the same Testing done: - Create/Mount/Delete PVC - Create/Delete 5 PVCs - Mount version 1.0.0 PVC - Delete version 1.0.0 PV - Mount Statically defined PV/PVC/Pod - Mount Statically defined version 1.0.0 PV/PVC/Pod - Delete Statically defined version 1.0.0 PV/PVC/Pod - Node restart when mounted to test mountcache - Use InstanceID other than 'default' - RBD basic round of tests, as namespace is added to OMaps - csitest against ceph-fs plugin - NOTE: CephFS plugin still does not detect and address already created volumes but of a different size - Test not providing any value to the metadata storage parameter Signed-off-by: ShyamsundarR <srangana@redhat.com>
2019-05-28 19:03:18 +00:00
if volOptions.FsName != "" {
args = append(args, "--client_mds_namespace="+volOptions.FsName)
}
_, stderr, err := execCommand(ctx, "ceph-fuse", args[:]...)
if err != nil {
return err
}
// Parse the output:
// We need "starting fuse" meaning the mount is ok
// and PID of the ceph-fuse daemon for unmount
2019-02-27 19:29:20 +00:00
match := fusePidRx.FindSubmatch(stderr)
if len(match) != 2 {
return fmt.Errorf("ceph-fuse failed: %s", stderr)
}
2019-02-27 19:29:20 +00:00
pid, err := strconv.Atoi(string(match[1]))
if err != nil {
return fmt.Errorf("failed to parse FUSE daemon PID: %v", err)
}
fusePidMapMtx.Lock()
fusePidMap[mountPoint] = pid
fusePidMapMtx.Unlock()
return nil
}
func (m *fuseMounter) mount(ctx context.Context, mountPoint string, cr *util.Credentials, volOptions *volumeOptions) error {
if err := util.CreateMountPoint(mountPoint); err != nil {
return err
}
return mountFuse(ctx, mountPoint, cr, volOptions)
}
func (m *fuseMounter) name() string { return "Ceph FUSE driver" }
type kernelMounter struct{}
func mountKernel(ctx context.Context, mountPoint string, cr *util.Credentials, volOptions *volumeOptions) error {
if err := execCommandErr(ctx, "modprobe", "ceph"); err != nil {
return err
}
Make CephFS plugin stateless reusing RADOS based journal scheme This is a part of the stateless set of commits for CephCSI. This commit removes the dependency on config maps to store cephFS provisioned volumes, and instead relies on RADOS based objects and keys, and required CSI VolumeID encoding to detect the provisioned volumes. Changes: - Provide backward compatibility to provisioned volumes by older plugin versions (1.0.0 or older) - Remove Create/Delete support for statically provisioned volumes (fixes #382) - Added namespace support to RADOS OMaps and used the same to store RADOS CSI objects and keys in the CephFS metadata pool - Added support to mention fsname for CephFS provisioning (fixes #359) - Changed field name in CSI Identifier to 'location', to denote a pool or fscid - Updated mounter cache to use new scheme - Required Helm manifests are updated - Required documentation and other manifests are updated - Made driver option 'metadatastorage' as optional, as fresh installs do not need to specify the same Testing done: - Create/Mount/Delete PVC - Create/Delete 5 PVCs - Mount version 1.0.0 PVC - Delete version 1.0.0 PV - Mount Statically defined PV/PVC/Pod - Mount Statically defined version 1.0.0 PV/PVC/Pod - Delete Statically defined version 1.0.0 PV/PVC/Pod - Node restart when mounted to test mountcache - Use InstanceID other than 'default' - RBD basic round of tests, as namespace is added to OMaps - csitest against ceph-fs plugin - NOTE: CephFS plugin still does not detect and address already created volumes but of a different size - Test not providing any value to the metadata storage parameter Signed-off-by: ShyamsundarR <srangana@redhat.com>
2019-05-28 19:03:18 +00:00
args := []string{
"-t", "ceph",
fmt.Sprintf("%s:%s", volOptions.Monitors, volOptions.RootPath),
mountPoint,
Make CephFS plugin stateless reusing RADOS based journal scheme This is a part of the stateless set of commits for CephCSI. This commit removes the dependency on config maps to store cephFS provisioned volumes, and instead relies on RADOS based objects and keys, and required CSI VolumeID encoding to detect the provisioned volumes. Changes: - Provide backward compatibility to provisioned volumes by older plugin versions (1.0.0 or older) - Remove Create/Delete support for statically provisioned volumes (fixes #382) - Added namespace support to RADOS OMaps and used the same to store RADOS CSI objects and keys in the CephFS metadata pool - Added support to mention fsname for CephFS provisioning (fixes #359) - Changed field name in CSI Identifier to 'location', to denote a pool or fscid - Updated mounter cache to use new scheme - Required Helm manifests are updated - Required documentation and other manifests are updated - Made driver option 'metadatastorage' as optional, as fresh installs do not need to specify the same Testing done: - Create/Mount/Delete PVC - Create/Delete 5 PVCs - Mount version 1.0.0 PVC - Delete version 1.0.0 PV - Mount Statically defined PV/PVC/Pod - Mount Statically defined version 1.0.0 PV/PVC/Pod - Delete Statically defined version 1.0.0 PV/PVC/Pod - Node restart when mounted to test mountcache - Use InstanceID other than 'default' - RBD basic round of tests, as namespace is added to OMaps - csitest against ceph-fs plugin - NOTE: CephFS plugin still does not detect and address already created volumes but of a different size - Test not providing any value to the metadata storage parameter Signed-off-by: ShyamsundarR <srangana@redhat.com>
2019-05-28 19:03:18 +00:00
}
optionsStr := fmt.Sprintf("name=%s,secretfile=%s", cr.ID, cr.KeyFile)
Make CephFS plugin stateless reusing RADOS based journal scheme This is a part of the stateless set of commits for CephCSI. This commit removes the dependency on config maps to store cephFS provisioned volumes, and instead relies on RADOS based objects and keys, and required CSI VolumeID encoding to detect the provisioned volumes. Changes: - Provide backward compatibility to provisioned volumes by older plugin versions (1.0.0 or older) - Remove Create/Delete support for statically provisioned volumes (fixes #382) - Added namespace support to RADOS OMaps and used the same to store RADOS CSI objects and keys in the CephFS metadata pool - Added support to mention fsname for CephFS provisioning (fixes #359) - Changed field name in CSI Identifier to 'location', to denote a pool or fscid - Updated mounter cache to use new scheme - Required Helm manifests are updated - Required documentation and other manifests are updated - Made driver option 'metadatastorage' as optional, as fresh installs do not need to specify the same Testing done: - Create/Mount/Delete PVC - Create/Delete 5 PVCs - Mount version 1.0.0 PVC - Delete version 1.0.0 PV - Mount Statically defined PV/PVC/Pod - Mount Statically defined version 1.0.0 PV/PVC/Pod - Delete Statically defined version 1.0.0 PV/PVC/Pod - Node restart when mounted to test mountcache - Use InstanceID other than 'default' - RBD basic round of tests, as namespace is added to OMaps - csitest against ceph-fs plugin - NOTE: CephFS plugin still does not detect and address already created volumes but of a different size - Test not providing any value to the metadata storage parameter Signed-off-by: ShyamsundarR <srangana@redhat.com>
2019-05-28 19:03:18 +00:00
if volOptions.FsName != "" {
optionsStr += fmt.Sprintf(",mds_namespace=%s", volOptions.FsName)
Make CephFS plugin stateless reusing RADOS based journal scheme This is a part of the stateless set of commits for CephCSI. This commit removes the dependency on config maps to store cephFS provisioned volumes, and instead relies on RADOS based objects and keys, and required CSI VolumeID encoding to detect the provisioned volumes. Changes: - Provide backward compatibility to provisioned volumes by older plugin versions (1.0.0 or older) - Remove Create/Delete support for statically provisioned volumes (fixes #382) - Added namespace support to RADOS OMaps and used the same to store RADOS CSI objects and keys in the CephFS metadata pool - Added support to mention fsname for CephFS provisioning (fixes #359) - Changed field name in CSI Identifier to 'location', to denote a pool or fscid - Updated mounter cache to use new scheme - Required Helm manifests are updated - Required documentation and other manifests are updated - Made driver option 'metadatastorage' as optional, as fresh installs do not need to specify the same Testing done: - Create/Mount/Delete PVC - Create/Delete 5 PVCs - Mount version 1.0.0 PVC - Delete version 1.0.0 PV - Mount Statically defined PV/PVC/Pod - Mount Statically defined version 1.0.0 PV/PVC/Pod - Delete Statically defined version 1.0.0 PV/PVC/Pod - Node restart when mounted to test mountcache - Use InstanceID other than 'default' - RBD basic round of tests, as namespace is added to OMaps - csitest against ceph-fs plugin - NOTE: CephFS plugin still does not detect and address already created volumes but of a different size - Test not providing any value to the metadata storage parameter Signed-off-by: ShyamsundarR <srangana@redhat.com>
2019-05-28 19:03:18 +00:00
}
if volOptions.KernelMountOptions != "" {
optionsStr += fmt.Sprintf(",%s", volOptions.KernelMountOptions)
}
Make CephFS plugin stateless reusing RADOS based journal scheme This is a part of the stateless set of commits for CephCSI. This commit removes the dependency on config maps to store cephFS provisioned volumes, and instead relies on RADOS based objects and keys, and required CSI VolumeID encoding to detect the provisioned volumes. Changes: - Provide backward compatibility to provisioned volumes by older plugin versions (1.0.0 or older) - Remove Create/Delete support for statically provisioned volumes (fixes #382) - Added namespace support to RADOS OMaps and used the same to store RADOS CSI objects and keys in the CephFS metadata pool - Added support to mention fsname for CephFS provisioning (fixes #359) - Changed field name in CSI Identifier to 'location', to denote a pool or fscid - Updated mounter cache to use new scheme - Required Helm manifests are updated - Required documentation and other manifests are updated - Made driver option 'metadatastorage' as optional, as fresh installs do not need to specify the same Testing done: - Create/Mount/Delete PVC - Create/Delete 5 PVCs - Mount version 1.0.0 PVC - Delete version 1.0.0 PV - Mount Statically defined PV/PVC/Pod - Mount Statically defined version 1.0.0 PV/PVC/Pod - Delete Statically defined version 1.0.0 PV/PVC/Pod - Node restart when mounted to test mountcache - Use InstanceID other than 'default' - RBD basic round of tests, as namespace is added to OMaps - csitest against ceph-fs plugin - NOTE: CephFS plugin still does not detect and address already created volumes but of a different size - Test not providing any value to the metadata storage parameter Signed-off-by: ShyamsundarR <srangana@redhat.com>
2019-05-28 19:03:18 +00:00
args = append(args, "-o", optionsStr)
return execCommandErr(ctx, "mount", args[:]...)
}
func (m *kernelMounter) mount(ctx context.Context, mountPoint string, cr *util.Credentials, volOptions *volumeOptions) error {
if err := util.CreateMountPoint(mountPoint); err != nil {
return err
}
return mountKernel(ctx, mountPoint, cr, volOptions)
}
func (m *kernelMounter) name() string { return "Ceph kernel client" }
func bindMount(ctx context.Context, from, to string, readOnly bool, mntOptions []string) error {
mntOptionSli := strings.Join(mntOptions, ",")
if err := execCommandErr(ctx, "mount", "-o", mntOptionSli, from, to); err != nil {
return fmt.Errorf("failed to bind-mount %s to %s: %v", from, to, err)
}
if readOnly {
mntOptionSli += ",remount"
if err := execCommandErr(ctx, "mount", "-o", mntOptionSli, to); err != nil {
return fmt.Errorf("failed read-only remount of %s: %v", to, err)
}
}
return nil
}
func unmountVolume(ctx context.Context, mountPoint string) error {
if err := execCommandErr(ctx, "umount", mountPoint); err != nil {
return err
}
fusePidMapMtx.Lock()
pid, ok := fusePidMap[mountPoint]
if ok {
delete(fusePidMap, mountPoint)
}
fusePidMapMtx.Unlock()
if ok {
p, err := os.FindProcess(pid)
if err != nil {
klog.Warningf(util.Log(ctx, "failed to find process %d: %v"), pid, err)
} else {
if _, err = p.Wait(); err != nil {
klog.Warningf(util.Log(ctx, "%d is not a child process: %v"), pid, err)
}
}
}
return nil
}