mirror of
https://github.com/ceph/ceph-csi.git
synced 2024-11-26 08:10:20 +00:00
649aeb7aaf
if the PVC access mode is ReadOnlyMany or single node readonly, mounting the rbd device path to the staging path as readonly to avoid the write operation. If the PVC acccess mode is readonly, mapping rbd images as readonly. Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
1019 lines
30 KiB
Go
1019 lines
30 KiB
Go
/*
|
|
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 rbd
|
|
|
|
import (
|
|
"context"
|
|
"encoding/json"
|
|
"fmt"
|
|
"io/ioutil"
|
|
"os"
|
|
"os/exec"
|
|
"path/filepath"
|
|
"strconv"
|
|
"strings"
|
|
"time"
|
|
|
|
"github.com/ceph/ceph-csi/internal/util"
|
|
|
|
"github.com/ceph/go-ceph/rados"
|
|
librbd "github.com/ceph/go-ceph/rbd"
|
|
"github.com/container-storage-interface/spec/lib/go/csi"
|
|
"github.com/golang/protobuf/ptypes"
|
|
"github.com/golang/protobuf/ptypes/timestamp"
|
|
"github.com/pborman/uuid"
|
|
"github.com/pkg/errors"
|
|
"k8s.io/apimachinery/pkg/util/sets"
|
|
"k8s.io/cloud-provider/volume/helpers"
|
|
"k8s.io/klog"
|
|
)
|
|
|
|
const (
|
|
imageWatcherStr = "watcher="
|
|
// The following three values are used for 30 seconds timeout
|
|
// while waiting for RBD Watcher to expire.
|
|
rbdImageWatcherInitDelay = 1 * time.Second
|
|
rbdImageWatcherFactor = 1.4
|
|
rbdImageWatcherSteps = 10
|
|
rbdDefaultMounter = "rbd"
|
|
|
|
// Output strings returned during invocation of "ceph rbd task add remove <imagespec>" when
|
|
// command is not supported by ceph manager. Used to check errors and recover when the command
|
|
// is unsupported.
|
|
rbdTaskRemoveCmdInvalidString1 = "no valid command found"
|
|
rbdTaskRemoveCmdInvalidString2 = "Error EINVAL: invalid command"
|
|
rbdTaskRemoveCmdAccessDeniedMessage = "Error EACCES:"
|
|
|
|
// Encryption statuses for RbdImage
|
|
rbdImageEncrypted = "encrypted"
|
|
rbdImageRequiresEncryption = "requiresEncryption"
|
|
// image metadata key for encryption
|
|
encryptionMetaKey = ".rbd.csi.ceph.com/encrypted"
|
|
)
|
|
|
|
// rbdVolume represents a CSI volume and its RBD image specifics
|
|
type rbdVolume struct {
|
|
// RbdImageName is the name of the RBD image backing this rbdVolume. This does not have a
|
|
// JSON tag as it is not stashed in JSON encoded config maps in v1.0.0
|
|
// VolID is the volume ID that is exchanged with CSI drivers, identifying this rbdVol
|
|
// RequestName is the CSI generated volume name for the rbdVolume. This does not have a
|
|
// JSON tag as it is not stashed in JSON encoded config maps in v1.0.0
|
|
// VolName and MonValueFromSecret are retained from older plugin versions (<= 1.0.0)
|
|
// for backward compatibility reasons
|
|
// JournalPool is the ceph pool in which the CSI Journal is stored
|
|
// Pool is where the image journal and image is stored, and could be the same as `JournalPool`
|
|
// (retained as Pool instead of renaming to ImagePool or such, as this is referenced in the code extensively)
|
|
// DataPool is where the data for images in `Pool` are stored, this is used as the `--data-pool`
|
|
// argument when the pool is created, and is not used anywhere else
|
|
TopologyPools *[]util.TopologyConstrainedPool
|
|
TopologyRequirement *csi.TopologyRequirement
|
|
Topology map[string]string
|
|
RbdImageName string
|
|
NamePrefix string
|
|
VolID string `json:"volID"`
|
|
Monitors string `json:"monitors"`
|
|
JournalPool string
|
|
Pool string `json:"pool"`
|
|
DataPool string
|
|
imageFeatureSet librbd.FeatureSet
|
|
AdminID string `json:"adminId"`
|
|
UserID string `json:"userId"`
|
|
Mounter string `json:"mounter"`
|
|
ClusterID string `json:"clusterId"`
|
|
RequestName string
|
|
VolName string `json:"volName"`
|
|
MonValueFromSecret string `json:"monValueFromSecret"`
|
|
VolSize int64 `json:"volSize"`
|
|
DisableInUseChecks bool `json:"disableInUseChecks"`
|
|
Encrypted bool
|
|
readOnly bool
|
|
KMS util.EncryptionKMS
|
|
|
|
// conn is a connection to the Ceph cluster obtained from a ConnPool
|
|
conn *util.ClusterConnection
|
|
// an opened IOContext, call .openIoctx() before using
|
|
ioctx *rados.IOContext
|
|
}
|
|
|
|
// rbdSnapshot represents a CSI snapshot and its RBD snapshot specifics
|
|
type rbdSnapshot struct {
|
|
// SourceVolumeID is the volume ID of RbdImageName, that is exchanged with CSI drivers
|
|
// RbdImageName is the name of the RBD image, that is this rbdSnapshot's source image
|
|
// RbdSnapName is the name of the RBD snapshot backing this rbdSnapshot
|
|
// SnapID is the snapshot ID that is exchanged with CSI drivers, identifying this rbdSnapshot
|
|
// RequestName is the CSI generated snapshot name for the rbdSnapshot
|
|
// JournalPool is the ceph pool in which the CSI snapshot Journal is stored
|
|
// Pool is where the image snapshot journal and snapshot is stored, and could be the same as `JournalPool`
|
|
SourceVolumeID string
|
|
RbdImageName string
|
|
NamePrefix string
|
|
RbdSnapName string
|
|
SnapID string
|
|
Monitors string
|
|
JournalPool string
|
|
Pool string
|
|
CreatedAt *timestamp.Timestamp
|
|
SizeBytes int64
|
|
ClusterID string
|
|
RequestName string
|
|
}
|
|
|
|
var (
|
|
supportedFeatures = sets.NewString(librbd.FeatureNameLayering)
|
|
)
|
|
|
|
// Connect an rbdVolume to the Ceph cluster
|
|
func (rv *rbdVolume) Connect(cr *util.Credentials) error {
|
|
if rv.conn != nil {
|
|
return nil
|
|
}
|
|
|
|
conn := &util.ClusterConnection{}
|
|
if err := conn.Connect(rv.Monitors, cr); err != nil {
|
|
return err
|
|
}
|
|
|
|
rv.conn = conn
|
|
return nil
|
|
}
|
|
|
|
// Destroy cleans up the rbdVolume and closes the connection to the Ceph
|
|
// cluster in case one was setup.
|
|
func (rv *rbdVolume) Destroy() {
|
|
if rv.ioctx != nil {
|
|
rv.ioctx.Destroy()
|
|
}
|
|
if rv.conn != nil {
|
|
rv.conn.Destroy()
|
|
}
|
|
}
|
|
|
|
// String returns the image-spec (pool/image) format of the image
|
|
func (rv *rbdVolume) String() string {
|
|
return fmt.Sprintf("%s/%s", rv.Pool, rv.RbdImageName)
|
|
}
|
|
|
|
// String returns the snap-spec (pool/image@snap) format of the snapshot
|
|
func (rs *rbdSnapshot) String() string {
|
|
return fmt.Sprintf("%s/%s@%s", rs.Pool, rs.RbdImageName, rs.RbdSnapName)
|
|
}
|
|
|
|
// imageSpec returns the image-spec (pool/image) format of the snapshot
|
|
func (rs *rbdSnapshot) imageSpec() string {
|
|
return fmt.Sprintf("%s/%s", rs.Pool, rs.RbdImageName)
|
|
}
|
|
|
|
// createImage creates a new ceph image with provision and volume options.
|
|
func createImage(ctx context.Context, pOpts *rbdVolume, cr *util.Credentials) error {
|
|
volSzMiB := fmt.Sprintf("%dM", util.RoundOffVolSize(pOpts.VolSize))
|
|
options := librbd.NewRbdImageOptions()
|
|
|
|
logMsg := "rbd: create %s size %s (features: %s) using mon %s"
|
|
if pOpts.DataPool != "" {
|
|
logMsg += fmt.Sprintf(", data pool %s", pOpts.DataPool)
|
|
err := options.SetString(librbd.RbdImageOptionDataPool, pOpts.DataPool)
|
|
if err != nil {
|
|
return errors.Wrapf(err, "failed to set data pool")
|
|
}
|
|
}
|
|
klog.V(4).Infof(util.Log(ctx, logMsg),
|
|
pOpts, volSzMiB, pOpts.imageFeatureSet.Names(), pOpts.Monitors)
|
|
|
|
if pOpts.imageFeatureSet != 0 {
|
|
err := options.SetUint64(librbd.RbdImageOptionFeatures, uint64(pOpts.imageFeatureSet))
|
|
if err != nil {
|
|
return errors.Wrapf(err, "failed to set image features")
|
|
}
|
|
}
|
|
|
|
err := pOpts.Connect(cr)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
err = pOpts.openIoctx()
|
|
if err != nil {
|
|
return errors.Wrapf(err, "failed to get IOContext")
|
|
}
|
|
|
|
err = librbd.CreateImage(pOpts.ioctx, pOpts.RbdImageName,
|
|
uint64(util.RoundOffVolSize(pOpts.VolSize)*helpers.MiB), options)
|
|
if err != nil {
|
|
return errors.Wrapf(err, "failed to create rbd image")
|
|
}
|
|
|
|
return nil
|
|
}
|
|
|
|
func (rv *rbdVolume) openIoctx() error {
|
|
if rv.ioctx != nil {
|
|
return nil
|
|
}
|
|
|
|
ioctx, err := rv.conn.GetIoctx(rv.Pool)
|
|
if err != nil {
|
|
// GetIoctx() can return util.ErrPoolNotFound
|
|
return err
|
|
}
|
|
|
|
rv.ioctx = ioctx
|
|
|
|
return nil
|
|
}
|
|
|
|
// open the rbdVolume after it has been connected.
|
|
// ErrPoolNotFound or ErrImageNotFound are returned in case the pool or image
|
|
// can not be found, other errors will contain more details about other issues
|
|
// (permission denied, ...) and are expected to relate to configuration issues.
|
|
func (rv *rbdVolume) open() (*librbd.Image, error) {
|
|
err := rv.openIoctx()
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
|
|
image, err := librbd.OpenImage(rv.ioctx, rv.RbdImageName, librbd.NoSnapshot)
|
|
if err != nil {
|
|
if err == librbd.ErrNotFound {
|
|
err = ErrImageNotFound{rv.RbdImageName, err}
|
|
}
|
|
return nil, err
|
|
}
|
|
return image, nil
|
|
}
|
|
|
|
// rbdStatus checks if there is watcher on the image.
|
|
// It returns true if there is a watcher on the image, otherwise returns false.
|
|
func rbdStatus(ctx context.Context, pOpts *rbdVolume, cr *util.Credentials) (bool, string, error) {
|
|
var output string
|
|
var cmd []byte
|
|
|
|
klog.V(4).Infof(util.Log(ctx, "rbd: status %s using mon %s"), pOpts, pOpts.Monitors)
|
|
args := []string{"status", pOpts.String(), "-m", pOpts.Monitors, "--id", cr.ID, "--keyfile=" + cr.KeyFile}
|
|
cmd, err := execCommand("rbd", args)
|
|
output = string(cmd)
|
|
|
|
if err, ok := err.(*exec.Error); ok {
|
|
if errors.Is(err.Err, exec.ErrNotFound) {
|
|
klog.Errorf(util.Log(ctx, "rbd cmd not found"))
|
|
// fail fast if command not found
|
|
return false, output, err
|
|
}
|
|
}
|
|
|
|
// If command never succeed, returns its last error.
|
|
if err != nil {
|
|
return false, output, err
|
|
}
|
|
|
|
if strings.Contains(output, imageWatcherStr) {
|
|
klog.V(4).Infof(util.Log(ctx, "rbd: watchers on %s: %s"), pOpts, output)
|
|
return true, output, nil
|
|
}
|
|
klog.Warningf(util.Log(ctx, "rbd: no watchers on %s"), pOpts)
|
|
return false, output, nil
|
|
}
|
|
|
|
// rbdManagerTaskDelete adds a ceph manager task to delete an rbd image, thus deleting
|
|
// it asynchronously. If command is not found returns a bool set to false
|
|
func rbdManagerTaskDeleteImage(ctx context.Context, pOpts *rbdVolume, cr *util.Credentials) (bool, error) {
|
|
var output []byte
|
|
|
|
args := []string{"rbd", "task", "add", "remove",
|
|
pOpts.String(),
|
|
"--id", cr.ID,
|
|
"--keyfile=" + cr.KeyFile,
|
|
"-m", pOpts.Monitors,
|
|
}
|
|
|
|
output, err := execCommand("ceph", args)
|
|
if err != nil {
|
|
switch {
|
|
case strings.Contains(string(output), rbdTaskRemoveCmdInvalidString1) &&
|
|
strings.Contains(string(output), rbdTaskRemoveCmdInvalidString2):
|
|
klog.Warningf(util.Log(ctx, "cluster with cluster ID (%s) does not support Ceph manager based rbd image"+
|
|
" deletion (minimum ceph version required is v14.2.3)"), pOpts.ClusterID)
|
|
case strings.HasPrefix(string(output), rbdTaskRemoveCmdAccessDeniedMessage):
|
|
klog.Warningf(util.Log(ctx, "access denied to Ceph MGR-based RBD image deletion "+
|
|
"on cluster ID (%s)"), pOpts.ClusterID)
|
|
default:
|
|
klog.Warningf(util.Log(ctx, "uncaught error while scheduling an image deletion task: %s"), err)
|
|
}
|
|
return false, err
|
|
}
|
|
|
|
return true, err
|
|
}
|
|
|
|
// deleteImage deletes a ceph image with provision and volume options.
|
|
func deleteImage(ctx context.Context, pOpts *rbdVolume, cr *util.Credentials) error {
|
|
image := pOpts.RbdImageName
|
|
found, _, err := rbdStatus(ctx, pOpts, cr)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
if found {
|
|
klog.Errorf(util.Log(ctx, "rbd %s is still being used"), pOpts)
|
|
return fmt.Errorf("rbd %s is still being used", pOpts)
|
|
}
|
|
|
|
klog.V(4).Infof(util.Log(ctx, "rbd: rm %s using mon %s"), pOpts, pOpts.Monitors)
|
|
|
|
// attempt to use Ceph manager based deletion support if available
|
|
rbdCephMgrSupported, err := rbdManagerTaskDeleteImage(ctx, pOpts, cr)
|
|
if rbdCephMgrSupported && err != nil {
|
|
klog.Errorf(util.Log(ctx, "failed to add task to delete rbd image: %s, %v"), pOpts, err)
|
|
return err
|
|
}
|
|
|
|
if !rbdCephMgrSupported {
|
|
var ioctx *rados.IOContext
|
|
ioctx, err = pOpts.conn.GetIoctx(pOpts.Pool)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
defer ioctx.Destroy()
|
|
|
|
rbdImage := librbd.GetImage(ioctx, image)
|
|
err = rbdImage.Remove()
|
|
if err != nil {
|
|
klog.Errorf(util.Log(ctx, "failed to delete rbd image: %s, error: %v"), pOpts, err)
|
|
}
|
|
}
|
|
|
|
return err
|
|
}
|
|
|
|
// updateSnapWithImageInfo updates provided rbdSnapshot with information from on-disk data
|
|
// regarding the same
|
|
func updateSnapWithImageInfo(ctx context.Context, rbdSnap *rbdSnapshot, cr *util.Credentials) error {
|
|
snapInfo, err := rbdSnap.getSnapInfo(ctx, rbdSnap.Monitors, cr)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
rbdSnap.SizeBytes = snapInfo.Size
|
|
|
|
tm, err := time.Parse(time.ANSIC, snapInfo.Timestamp)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
rbdSnap.CreatedAt, err = ptypes.TimestampProto(tm)
|
|
|
|
return err
|
|
}
|
|
|
|
// genSnapFromSnapID generates a rbdSnapshot structure from the provided identifier, updating
|
|
// the structure with elements from on-disk snapshot metadata as well
|
|
func genSnapFromSnapID(ctx context.Context, rbdSnap *rbdSnapshot, snapshotID string, cr *util.Credentials) error {
|
|
var (
|
|
options map[string]string
|
|
vi util.CSIIdentifier
|
|
)
|
|
options = make(map[string]string)
|
|
|
|
rbdSnap.SnapID = snapshotID
|
|
|
|
err := vi.DecomposeCSIID(rbdSnap.SnapID)
|
|
if err != nil {
|
|
klog.Errorf(util.Log(ctx, "error decoding snapshot ID (%s) (%s)"), err, rbdSnap.SnapID)
|
|
return err
|
|
}
|
|
|
|
rbdSnap.ClusterID = vi.ClusterID
|
|
options["clusterID"] = rbdSnap.ClusterID
|
|
|
|
rbdSnap.Monitors, _, err = getMonsAndClusterID(ctx, options)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
rbdSnap.Pool, err = util.GetPoolName(rbdSnap.Monitors, cr, vi.LocationID)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
rbdSnap.JournalPool = rbdSnap.Pool
|
|
|
|
j, err := snapJournal.Connect(rbdSnap.Monitors, cr)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
defer j.Destroy()
|
|
|
|
imageAttributes, err := j.GetImageAttributes(
|
|
ctx, rbdSnap.Pool, vi.ObjectUUID, true)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
rbdSnap.RequestName = imageAttributes.RequestName
|
|
rbdSnap.RbdImageName = imageAttributes.SourceName
|
|
rbdSnap.RbdSnapName = imageAttributes.ImageName
|
|
|
|
// convert the journal pool ID to name, for use in DeleteSnapshot cases
|
|
if imageAttributes.JournalPoolID != util.InvalidPoolID {
|
|
rbdSnap.JournalPool, err = util.GetPoolName(rbdSnap.Monitors, cr, imageAttributes.JournalPoolID)
|
|
if err != nil {
|
|
// TODO: If pool is not found we may leak the image (as DeleteSnapshot will return success)
|
|
return err
|
|
}
|
|
}
|
|
|
|
err = updateSnapWithImageInfo(ctx, rbdSnap, cr)
|
|
|
|
return err
|
|
}
|
|
|
|
// genVolFromVolID generates a rbdVolume structure from the provided identifier, updating
|
|
// the structure with elements from on-disk image metadata as well
|
|
func genVolFromVolID(ctx context.Context, volumeID string, cr *util.Credentials, secrets map[string]string) (*rbdVolume, error) {
|
|
var (
|
|
options map[string]string
|
|
vi util.CSIIdentifier
|
|
rbdVol *rbdVolume
|
|
)
|
|
options = make(map[string]string)
|
|
|
|
// rbdVolume fields that are not filled up in this function are:
|
|
// Mounter, MultiNodeWritable
|
|
rbdVol = &rbdVolume{VolID: volumeID}
|
|
|
|
err := vi.DecomposeCSIID(rbdVol.VolID)
|
|
if err != nil {
|
|
err = fmt.Errorf("error decoding volume ID (%s) (%s)", err, rbdVol.VolID)
|
|
return nil, ErrInvalidVolID{err}
|
|
}
|
|
|
|
rbdVol.ClusterID = vi.ClusterID
|
|
options["clusterID"] = rbdVol.ClusterID
|
|
|
|
rbdVol.Monitors, _, err = getMonsAndClusterID(ctx, options)
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
|
|
rbdVol.Pool, err = util.GetPoolName(rbdVol.Monitors, cr, vi.LocationID)
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
|
|
err = rbdVol.Connect(cr)
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
rbdVol.JournalPool = rbdVol.Pool
|
|
|
|
j, err := volJournal.Connect(rbdVol.Monitors, cr)
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
defer j.Destroy()
|
|
|
|
imageAttributes, err := j.GetImageAttributes(
|
|
ctx, rbdVol.Pool, vi.ObjectUUID, false)
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
|
|
if imageAttributes.KmsID != "" {
|
|
rbdVol.Encrypted = true
|
|
rbdVol.KMS, err = util.GetKMS(imageAttributes.KmsID, secrets)
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
}
|
|
rbdVol.RequestName = imageAttributes.RequestName
|
|
rbdVol.RbdImageName = imageAttributes.ImageName
|
|
|
|
// convert the journal pool ID to name, for use in DeleteVolume cases
|
|
if imageAttributes.JournalPoolID >= 0 {
|
|
rbdVol.JournalPool, err = util.GetPoolName(rbdVol.Monitors, cr, imageAttributes.JournalPoolID)
|
|
if err != nil {
|
|
// TODO: If pool is not found we may leak the image (as DeleteVolume will return success)
|
|
return nil, err
|
|
}
|
|
}
|
|
|
|
err = rbdVol.getImageInfo()
|
|
|
|
return rbdVol, err
|
|
}
|
|
|
|
func execCommand(command string, args []string) ([]byte, error) {
|
|
// #nosec
|
|
cmd := exec.Command(command, args...)
|
|
return cmd.CombinedOutput()
|
|
}
|
|
|
|
func getMonsAndClusterID(ctx context.Context, options map[string]string) (monitors, clusterID string, err error) {
|
|
var ok bool
|
|
|
|
if clusterID, ok = options["clusterID"]; !ok {
|
|
err = errors.New("clusterID must be set")
|
|
return
|
|
}
|
|
|
|
if monitors, err = util.Mons(csiConfigFile, clusterID); err != nil {
|
|
klog.Errorf(util.Log(ctx, "failed getting mons (%s)"), err)
|
|
err = errors.Wrapf(err, "failed to fetch monitor list using clusterID (%s)", clusterID)
|
|
return
|
|
}
|
|
|
|
return
|
|
}
|
|
|
|
// isLegacyVolumeID checks if passed in volume ID string conforms to volume ID naming scheme used
|
|
// by the version 1.0.0 (legacy) of the plugin, and returns true if found to be conforming
|
|
func isLegacyVolumeID(volumeID string) bool {
|
|
// Version 1.0.0 volumeID format: "csi-rbd-vol-" + UUID string
|
|
// length: 12 ("csi-rbd-vol-") + 36 (UUID string)
|
|
|
|
// length check
|
|
if len(volumeID) != 48 {
|
|
return false
|
|
}
|
|
|
|
// Header check
|
|
if !strings.HasPrefix(volumeID, "csi-rbd-vol-") {
|
|
return false
|
|
}
|
|
|
|
// Trailer UUID format check
|
|
if uuid.Parse(volumeID[12:]) == nil {
|
|
return false
|
|
}
|
|
|
|
return true
|
|
}
|
|
|
|
// upadateMons function is used to update the rbdVolume.Monitors for volumes that were provisioned
|
|
// using the 1.0.0 version (legacy) of the plugin.
|
|
func updateMons(rbdVol *rbdVolume, options, credentials map[string]string) error {
|
|
var ok bool
|
|
|
|
// read monitors and MonValueFromSecret from options, else check passed in rbdVolume for
|
|
// MonValueFromSecret key in credentials
|
|
monInSecret := ""
|
|
if options != nil {
|
|
if rbdVol.Monitors, ok = options["monitors"]; !ok {
|
|
rbdVol.Monitors = ""
|
|
}
|
|
if monInSecret, ok = options["monValueFromSecret"]; !ok {
|
|
monInSecret = ""
|
|
}
|
|
} else {
|
|
monInSecret = rbdVol.MonValueFromSecret
|
|
}
|
|
|
|
// if monitors are present in secrets and we have the credentials, use monitors from the
|
|
// credentials overriding monitors from other sources
|
|
if monInSecret != "" && credentials != nil {
|
|
monsFromSecret, ok := credentials[monInSecret]
|
|
if ok {
|
|
rbdVol.Monitors = monsFromSecret
|
|
}
|
|
}
|
|
|
|
if rbdVol.Monitors == "" {
|
|
return errors.New("either monitors or monValueFromSecret must be set")
|
|
}
|
|
|
|
return nil
|
|
}
|
|
|
|
func genVolFromVolumeOptions(ctx context.Context, volOptions, credentials map[string]string, disableInUseChecks, isLegacyVolume bool) (*rbdVolume, error) {
|
|
var (
|
|
ok bool
|
|
err error
|
|
namePrefix string
|
|
encrypted string
|
|
)
|
|
|
|
rbdVol := &rbdVolume{}
|
|
rbdVol.Pool, ok = volOptions["pool"]
|
|
if !ok {
|
|
return nil, errors.New("missing required parameter pool")
|
|
}
|
|
|
|
rbdVol.DataPool = volOptions["dataPool"]
|
|
if namePrefix, ok = volOptions["volumeNamePrefix"]; ok {
|
|
rbdVol.NamePrefix = namePrefix
|
|
}
|
|
|
|
if isLegacyVolume {
|
|
err = updateMons(rbdVol, volOptions, credentials)
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
} else {
|
|
rbdVol.Monitors, rbdVol.ClusterID, err = getMonsAndClusterID(ctx, volOptions)
|
|
if err != nil {
|
|
return nil, err
|
|
}
|
|
}
|
|
|
|
// if no image features is provided, it results in empty string
|
|
// which disable all RBD image features as we expected
|
|
|
|
imageFeatures, found := volOptions["imageFeatures"]
|
|
if found {
|
|
arr := strings.Split(imageFeatures, ",")
|
|
for _, f := range arr {
|
|
if !supportedFeatures.Has(f) {
|
|
return nil, fmt.Errorf("invalid feature %q for volume csi-rbdplugin, supported"+
|
|
" features are: %v", f, supportedFeatures)
|
|
}
|
|
}
|
|
rbdVol.imageFeatureSet = librbd.FeatureSetFromNames(arr)
|
|
}
|
|
|
|
klog.V(3).Infof(util.Log(ctx, "setting disableInUseChecks on rbd volume to: %v"), disableInUseChecks)
|
|
rbdVol.DisableInUseChecks = disableInUseChecks
|
|
|
|
rbdVol.Mounter, ok = volOptions["mounter"]
|
|
if !ok {
|
|
rbdVol.Mounter = rbdDefaultMounter
|
|
}
|
|
|
|
rbdVol.Encrypted = false
|
|
encrypted, ok = volOptions["encrypted"]
|
|
if ok {
|
|
rbdVol.Encrypted, err = strconv.ParseBool(encrypted)
|
|
if err != nil {
|
|
return nil, fmt.Errorf(
|
|
"invalid value set in 'encrypted': %s (should be \"true\" or \"false\")", encrypted)
|
|
}
|
|
|
|
if rbdVol.Encrypted {
|
|
// deliberately ignore if parsing failed as GetKMS will return default
|
|
// implementation of kmsID is empty
|
|
kmsID := volOptions["encryptionKMSID"]
|
|
rbdVol.KMS, err = util.GetKMS(kmsID, credentials)
|
|
if err != nil {
|
|
return nil, fmt.Errorf("invalid encryption kms configuration: %s", err)
|
|
}
|
|
}
|
|
}
|
|
|
|
return rbdVol, nil
|
|
}
|
|
|
|
func genSnapFromOptions(ctx context.Context, rbdVol *rbdVolume, snapOptions map[string]string) *rbdSnapshot {
|
|
var err error
|
|
|
|
rbdSnap := &rbdSnapshot{}
|
|
rbdSnap.Pool = rbdVol.Pool
|
|
rbdSnap.JournalPool = rbdVol.JournalPool
|
|
|
|
rbdSnap.Monitors, rbdSnap.ClusterID, err = getMonsAndClusterID(ctx, snapOptions)
|
|
if err != nil {
|
|
rbdSnap.Monitors = rbdVol.Monitors
|
|
rbdSnap.ClusterID = rbdVol.ClusterID
|
|
}
|
|
|
|
if namePrefix, ok := snapOptions["snapshotNamePrefix"]; ok {
|
|
rbdSnap.NamePrefix = namePrefix
|
|
}
|
|
|
|
return rbdSnap
|
|
}
|
|
|
|
// hasSnapshotFeature checks if Layering is enabled for this image
|
|
func (rv *rbdVolume) hasSnapshotFeature() bool {
|
|
return (uint64(rv.imageFeatureSet) & librbd.FeatureLayering) == librbd.FeatureLayering
|
|
}
|
|
|
|
func protectSnapshot(ctx context.Context, pOpts *rbdSnapshot, cr *util.Credentials) error {
|
|
var output []byte
|
|
|
|
klog.V(4).Infof(util.Log(ctx, "rbd: snap protect %s using mon %s"), pOpts, pOpts.Monitors)
|
|
args := []string{"snap", "protect", pOpts.String(), "--id", cr.ID, "-m", pOpts.Monitors, "--keyfile=" + cr.KeyFile}
|
|
|
|
output, err := execCommand("rbd", args)
|
|
|
|
if err != nil {
|
|
return errors.Wrapf(err, "failed to protect snapshot, command output: %s", string(output))
|
|
}
|
|
|
|
return nil
|
|
}
|
|
|
|
func createSnapshot(ctx context.Context, pOpts *rbdSnapshot, cr *util.Credentials) error {
|
|
var output []byte
|
|
|
|
klog.V(4).Infof(util.Log(ctx, "rbd: snap create %s using mon %s"), pOpts, pOpts.Monitors)
|
|
args := []string{"snap", "create", pOpts.String(), "--id", cr.ID, "-m", pOpts.Monitors, "--keyfile=" + cr.KeyFile}
|
|
|
|
output, err := execCommand("rbd", args)
|
|
|
|
if err != nil {
|
|
return errors.Wrapf(err, "failed to create snapshot, command output: %s", string(output))
|
|
}
|
|
|
|
return nil
|
|
}
|
|
|
|
func unprotectSnapshot(ctx context.Context, pOpts *rbdSnapshot, cr *util.Credentials) error {
|
|
var output []byte
|
|
|
|
klog.V(4).Infof(util.Log(ctx, "rbd: snap unprotect %s using mon %s"), pOpts, pOpts.Monitors)
|
|
args := []string{"snap", "unprotect", pOpts.String(), "--id", cr.ID, "-m", pOpts.Monitors, "--keyfile=" + cr.KeyFile}
|
|
|
|
output, err := execCommand("rbd", args)
|
|
|
|
if err != nil {
|
|
return errors.Wrapf(err, "failed to unprotect snapshot, command output: %s", string(output))
|
|
}
|
|
|
|
return nil
|
|
}
|
|
|
|
func deleteSnapshot(ctx context.Context, pOpts *rbdSnapshot, cr *util.Credentials) error {
|
|
var output []byte
|
|
|
|
klog.V(4).Infof(util.Log(ctx, "rbd: snap rm %s using mon %s"), pOpts, pOpts.Monitors)
|
|
args := []string{"snap", "rm", pOpts.String(), "--id", cr.ID, "-m", pOpts.Monitors, "--keyfile=" + cr.KeyFile}
|
|
|
|
output, err := execCommand("rbd", args)
|
|
|
|
if err != nil {
|
|
return errors.Wrapf(err, "failed to delete snapshot, command output: %s", string(output))
|
|
}
|
|
|
|
if err := undoSnapReservation(ctx, pOpts, cr); err != nil {
|
|
klog.Errorf(util.Log(ctx, "failed to remove reservation for snapname (%s) with backing snap (%s): %s"),
|
|
pOpts.RequestName, pOpts, err)
|
|
}
|
|
|
|
return nil
|
|
}
|
|
|
|
// nolint: interfacer
|
|
func restoreSnapshot(ctx context.Context, pVolOpts *rbdVolume, pSnapOpts *rbdSnapshot, cr *util.Credentials) error {
|
|
var output []byte
|
|
|
|
klog.V(4).Infof(util.Log(ctx, "rbd: clone %s using mon %s"), pVolOpts, pVolOpts.Monitors)
|
|
args := []string{"clone", pSnapOpts.String(), pVolOpts.String(),
|
|
"--id", cr.ID, "-m", pVolOpts.Monitors, "--keyfile=" + cr.KeyFile}
|
|
|
|
output, err := execCommand("rbd", args)
|
|
|
|
if err != nil {
|
|
return errors.Wrapf(err, "failed to restore snapshot, command output: %s", string(output))
|
|
}
|
|
|
|
return nil
|
|
}
|
|
|
|
// getSnapshotMetadata fetches on-disk metadata about the snapshot and populates the passed in
|
|
// rbdSnapshot structure
|
|
func getSnapshotMetadata(ctx context.Context, pSnapOpts *rbdSnapshot, cr *util.Credentials) error {
|
|
snapInfo, err := pSnapOpts.getSnapInfo(ctx, pSnapOpts.Monitors, cr)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
pSnapOpts.SizeBytes = snapInfo.Size
|
|
|
|
tm, err := time.Parse(time.ANSIC, snapInfo.Timestamp)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
pSnapOpts.CreatedAt, err = ptypes.TimestampProto(tm)
|
|
if err != nil {
|
|
return err
|
|
}
|
|
|
|
return nil
|
|
}
|
|
|
|
// getImageInfo queries rbd about the given image and returns its metadata, and returns
|
|
// ErrImageNotFound if provided image is not found
|
|
func (rv *rbdVolume) getImageInfo() error {
|
|
image, err := rv.open()
|
|
if err != nil {
|
|
return err
|
|
}
|
|
defer image.Close()
|
|
|
|
imageInfo, err := image.Stat()
|
|
if err != nil {
|
|
return err
|
|
}
|
|
// TODO: can rv.VolSize not be a uint64? Or initialize it to -1?
|
|
rv.VolSize = int64(imageInfo.Size)
|
|
|
|
features, err := image.GetFeatures()
|
|
if err != nil {
|
|
return err
|
|
}
|
|
rv.imageFeatureSet = librbd.FeatureSet(features)
|
|
|
|
return nil
|
|
}
|
|
|
|
// snapInfo strongly typed JSON spec for snap ls rbd output
|
|
type snapInfo struct {
|
|
ID int64 `json:"id"`
|
|
Name string `json:"name"`
|
|
Size int64 `json:"size"`
|
|
Timestamp string `json:"timestamp"`
|
|
}
|
|
|
|
/*
|
|
getSnapInfo queries rbd about the snapshots of the given image and returns its metadata, and
|
|
returns ErrImageNotFound if provided image is not found, and ErrSnapNotFound if provided snap
|
|
is not found in the images snapshot list
|
|
*/
|
|
func (rs *rbdSnapshot) getSnapInfo(ctx context.Context, monitors string, cr *util.Credentials) (snapInfo, error) {
|
|
// rbd --format=json snap ls [image-spec]
|
|
|
|
var (
|
|
snpInfo snapInfo
|
|
snaps []snapInfo
|
|
)
|
|
|
|
stdout, stderr, err := util.ExecCommand(
|
|
"rbd",
|
|
"-m", monitors,
|
|
"--id", cr.ID,
|
|
"--keyfile="+cr.KeyFile,
|
|
"-c", util.CephConfigPath,
|
|
"--format="+"json",
|
|
"snap", "ls", rs.imageSpec())
|
|
if err != nil {
|
|
klog.Errorf(util.Log(ctx, "failed getting snap (%s) information from image (%s): (%s)"),
|
|
rs.RbdSnapName, rs.imageSpec(), err)
|
|
if strings.Contains(string(stderr), "rbd: error opening image "+rs.RbdImageName+
|
|
": (2) No such file or directory") {
|
|
return snpInfo, ErrImageNotFound{rs.imageSpec(), err}
|
|
}
|
|
return snpInfo, err
|
|
}
|
|
|
|
err = json.Unmarshal(stdout, &snaps)
|
|
if err != nil {
|
|
klog.Errorf(util.Log(ctx, "failed to parse JSON output of image snap list (%s): (%s)"), rs.imageSpec(), err)
|
|
return snpInfo, fmt.Errorf("unmarshal failed: %+v. raw buffer response: %s", err, string(stdout))
|
|
}
|
|
|
|
for _, snap := range snaps {
|
|
if snap.Name == rs.RbdSnapName {
|
|
return snap, nil
|
|
}
|
|
}
|
|
|
|
return snpInfo, ErrSnapNotFound{rs.String(), fmt.Errorf("snap (%s) not found", rs)}
|
|
}
|
|
|
|
// rbdImageMetadataStash strongly typed JSON spec for stashed RBD image metadata
|
|
type rbdImageMetadataStash struct {
|
|
Version int `json:"Version"`
|
|
Pool string `json:"pool"`
|
|
ImageName string `json:"image"`
|
|
NbdAccess bool `json:"accessType"`
|
|
Encrypted bool `json:"encrypted"`
|
|
}
|
|
|
|
// file name in which image metadata is stashed
|
|
const stashFileName = "image-meta.json"
|
|
|
|
// spec returns the image-spec (pool/image) format of the image
|
|
func (ri *rbdImageMetadataStash) String() string {
|
|
return fmt.Sprintf("%s/%s", ri.Pool, ri.ImageName)
|
|
}
|
|
|
|
// stashRBDImageMetadata stashes required fields into the stashFileName at the passed in path, in
|
|
// JSON format
|
|
func stashRBDImageMetadata(volOptions *rbdVolume, path string) error {
|
|
var imgMeta = rbdImageMetadataStash{
|
|
Version: 2, // there are no checks for this at present
|
|
Pool: volOptions.Pool,
|
|
ImageName: volOptions.RbdImageName,
|
|
Encrypted: volOptions.Encrypted,
|
|
}
|
|
|
|
imgMeta.NbdAccess = false
|
|
if volOptions.Mounter == rbdTonbd && hasNBD {
|
|
imgMeta.NbdAccess = true
|
|
}
|
|
|
|
encodedBytes, err := json.Marshal(imgMeta)
|
|
if err != nil {
|
|
return fmt.Errorf("failed to marshall JSON image metadata for image (%s): (%v)", volOptions, err)
|
|
}
|
|
|
|
fPath := filepath.Join(path, stashFileName)
|
|
err = ioutil.WriteFile(fPath, encodedBytes, 0600)
|
|
if err != nil {
|
|
return fmt.Errorf("failed to stash JSON image metadata for image (%s) at path (%s): (%v)", volOptions, fPath, err)
|
|
}
|
|
|
|
return nil
|
|
}
|
|
|
|
// lookupRBDImageMetadataStash reads and returns stashed image metadata at passed in path
|
|
func lookupRBDImageMetadataStash(path string) (rbdImageMetadataStash, error) {
|
|
var imgMeta rbdImageMetadataStash
|
|
|
|
fPath := filepath.Join(path, stashFileName)
|
|
encodedBytes, err := ioutil.ReadFile(fPath) // #nosec - intended reading from fPath
|
|
if err != nil {
|
|
if !os.IsNotExist(err) {
|
|
return imgMeta, fmt.Errorf("failed to read stashed JSON image metadata from path (%s): (%v)", fPath, err)
|
|
}
|
|
|
|
return imgMeta, ErrMissingStash{err}
|
|
}
|
|
|
|
err = json.Unmarshal(encodedBytes, &imgMeta)
|
|
if err != nil {
|
|
return imgMeta, fmt.Errorf("failed to unmarshall stashed JSON image metadata from path (%s): (%v)", fPath, err)
|
|
}
|
|
|
|
return imgMeta, nil
|
|
}
|
|
|
|
// cleanupRBDImageMetadataStash cleans up any stashed metadata at passed in path
|
|
func cleanupRBDImageMetadataStash(path string) error {
|
|
fPath := filepath.Join(path, stashFileName)
|
|
if err := os.Remove(fPath); err != nil {
|
|
return fmt.Errorf("failed to cleanup stashed JSON data (%s): (%v)", fPath, err)
|
|
}
|
|
|
|
return nil
|
|
}
|
|
|
|
// resizeRBDImage resizes the given volume to new size
|
|
func resizeRBDImage(rbdVol *rbdVolume, cr *util.Credentials) error {
|
|
var output []byte
|
|
|
|
mon := rbdVol.Monitors
|
|
volSzMiB := fmt.Sprintf("%dM", util.RoundOffVolSize(rbdVol.VolSize))
|
|
|
|
args := []string{"resize", rbdVol.String(), "--size", volSzMiB, "--id", cr.ID, "-m", mon, "--keyfile=" + cr.KeyFile}
|
|
output, err := execCommand("rbd", args)
|
|
|
|
if err != nil {
|
|
return errors.Wrapf(err, "failed to resize rbd image, command output: %s", string(output))
|
|
}
|
|
|
|
return nil
|
|
}
|
|
|
|
func (rv *rbdVolume) GetMetadata(key string) (string, error) {
|
|
image, err := rv.open()
|
|
if err != nil {
|
|
return "", err
|
|
}
|
|
defer image.Close()
|
|
|
|
return image.GetMetadata(key)
|
|
}
|
|
|
|
func (rv *rbdVolume) SetMetadata(key, value string) error {
|
|
image, err := rv.open()
|
|
if err != nil {
|
|
return err
|
|
}
|
|
defer image.Close()
|
|
|
|
return image.SetMetadata(key, value)
|
|
}
|
|
|
|
// checkRbdImageEncrypted verifies if rbd image was encrypted when created
|
|
func (rv *rbdVolume) checkRbdImageEncrypted(ctx context.Context) (string, error) {
|
|
value, err := rv.GetMetadata(encryptionMetaKey)
|
|
if err != nil {
|
|
klog.Errorf(util.Log(ctx, "checking image %s encrypted state metadata failed: %s"), rv, err)
|
|
return "", err
|
|
}
|
|
|
|
encrypted := strings.TrimSpace(value)
|
|
klog.V(4).Infof(util.Log(ctx, "image %s encrypted state metadata reports %q"), rv, encrypted)
|
|
return encrypted, nil
|
|
}
|
|
|
|
func (rv *rbdVolume) ensureEncryptionMetadataSet(status string) error {
|
|
err := rv.SetMetadata(encryptionMetaKey, status)
|
|
if err != nil {
|
|
return fmt.Errorf("failed to save encryption status for %s: %v", rv, err)
|
|
}
|
|
|
|
return nil
|
|
}
|