mirror of
https://github.com/ceph/ceph-csi.git
synced 2024-11-15 02:40:23 +00:00
d98516e9d8
Add VolumeGroupLocks in the CSI Controller Server so that operations are protected against concurrent requests for the same VolumeGroupSnapshot. Signed-off-by: Niels de Vos <ndevos@ibm.com>
248 lines
7.4 KiB
Go
248 lines
7.4 KiB
Go
/*
|
|
Copyright 2024 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"
|
|
|
|
"github.com/container-storage-interface/spec/lib/go/csi"
|
|
"google.golang.org/grpc/codes"
|
|
"google.golang.org/grpc/status"
|
|
|
|
"github.com/ceph/ceph-csi/internal/rbd/types"
|
|
"github.com/ceph/ceph-csi/internal/util"
|
|
"github.com/ceph/ceph-csi/internal/util/log"
|
|
)
|
|
|
|
// CreateVolumeGroupSnapshot receives a list of volume handles and is requested
|
|
// to create a list of snapshots that are created at the same time. This is
|
|
// similar (although not exactly the same) to consistency groups.
|
|
//
|
|
// RBD has a limitation that an image can only belong to a single group. It is
|
|
// therefore required to create a temporary group, add all images, create the
|
|
// group snapshot and remove all images from the group again. This leaves the
|
|
// group and its snapshot around, the group snapshot can be inspected to list
|
|
// the snapshots of the images.
|
|
func (cs *ControllerServer) CreateVolumeGroupSnapshot(
|
|
ctx context.Context,
|
|
req *csi.CreateVolumeGroupSnapshotRequest,
|
|
) (*csi.CreateVolumeGroupSnapshotResponse, error) {
|
|
var (
|
|
err error
|
|
vg types.VolumeGroup
|
|
groupSnapshot types.VolumeGroupSnapshot
|
|
|
|
// the VG and VGS should not have the same name
|
|
vgName = req.GetName() + "-vg" // stable temporary name
|
|
vgsName = req.GetName()
|
|
)
|
|
|
|
// Existence and conflict checks
|
|
if acquired := cs.VolumeGroupLocks.TryAcquire(vgsName); !acquired {
|
|
log.ErrorLog(ctx, util.SnapshotOperationAlreadyExistsFmt, vgsName)
|
|
|
|
return nil, status.Errorf(codes.Aborted, util.SnapshotOperationAlreadyExistsFmt, vgsName)
|
|
}
|
|
defer cs.VolumeGroupLocks.Release(vgsName)
|
|
|
|
mgr := NewManager(cs.Driver.GetInstanceID(), req.GetParameters(), req.GetSecrets())
|
|
defer mgr.Destroy(ctx)
|
|
|
|
// resolve all volumes, free them later on
|
|
volumes := make([]types.Volume, len(req.GetSourceVolumeIds()))
|
|
defer func() {
|
|
for _, volume := range volumes {
|
|
if vg != nil {
|
|
// 'normal' cleanup, remove all images from the group
|
|
vgErr := vg.RemoveVolume(ctx, volume)
|
|
if vgErr != nil {
|
|
log.ErrorLog(
|
|
ctx,
|
|
"failed to remove volume %q from volume group %q: %v",
|
|
volume, vg, vgErr)
|
|
}
|
|
}
|
|
|
|
// free all allocated volumes
|
|
if volume != nil {
|
|
volume.Destroy(ctx)
|
|
}
|
|
}
|
|
|
|
if vg != nil {
|
|
// the VG should always be deleted, volumes can only belong to a single VG
|
|
log.DebugLog(ctx, "removing temporary volume group %q", vg)
|
|
|
|
vgErr := vg.Delete(ctx)
|
|
if vgErr != nil {
|
|
log.ErrorLog(ctx, "failed to remove temporary volume group %q: %v", vg, vgErr)
|
|
}
|
|
|
|
// free the resources of the VolumeGroup
|
|
vg.Destroy(ctx)
|
|
}
|
|
}()
|
|
for i, id := range req.GetSourceVolumeIds() {
|
|
var vol types.Volume
|
|
vol, err = mgr.GetVolumeByID(ctx, id)
|
|
if err != nil {
|
|
return nil, status.Errorf(
|
|
codes.InvalidArgument,
|
|
"failed to find required volume %q for volume group snapshot %q: %s",
|
|
id,
|
|
vgsName,
|
|
err.Error())
|
|
}
|
|
volumes[i] = vol
|
|
}
|
|
|
|
log.DebugLog(ctx, "all %d Volumes for VolumeGroup %q have been found", len(volumes), vgsName)
|
|
|
|
groupSnapshot, err = mgr.GetVolumeGroupSnapshotByName(ctx, vgsName)
|
|
if groupSnapshot != nil {
|
|
defer groupSnapshot.Destroy(ctx)
|
|
|
|
csiVGS, csiErr := groupSnapshot.ToCSI(ctx)
|
|
if csiErr != nil {
|
|
return nil, status.Error(codes.Aborted, csiErr.Error())
|
|
}
|
|
|
|
return &csi.CreateVolumeGroupSnapshotResponse{
|
|
GroupSnapshot: csiVGS,
|
|
}, nil
|
|
}
|
|
if err != nil {
|
|
log.DebugLog(ctx, "need to create new volume group snapshot, "+
|
|
"failed to get existing one with name %q: %v", vgsName, err)
|
|
}
|
|
|
|
// create a temporary VolumeGroup with a different name
|
|
vg, err = mgr.CreateVolumeGroup(ctx, vgName)
|
|
if err != nil {
|
|
return nil, status.Errorf(
|
|
codes.Internal,
|
|
"failed to create volume group %q: %s",
|
|
vgName,
|
|
err.Error())
|
|
}
|
|
// vg.Destroy(ctx) is called in a defer further up ^
|
|
|
|
log.DebugLog(ctx, "VolumeGroup %q has been created: %+v", vgName, vg)
|
|
|
|
// add images to the group
|
|
for _, volume := range volumes {
|
|
err = vg.AddVolume(ctx, volume)
|
|
if err != nil {
|
|
return nil, status.Error(codes.Aborted, err.Error())
|
|
}
|
|
}
|
|
|
|
groupSnapshot, err = mgr.CreateVolumeGroupSnapshot(ctx, vg, vgsName)
|
|
if err != nil {
|
|
return nil, status.Error(codes.Aborted, err.Error())
|
|
}
|
|
defer groupSnapshot.Destroy(ctx)
|
|
|
|
csiVGS, err := groupSnapshot.ToCSI(ctx)
|
|
if err != nil {
|
|
return nil, status.Error(codes.Aborted, err.Error())
|
|
}
|
|
|
|
return &csi.CreateVolumeGroupSnapshotResponse{
|
|
GroupSnapshot: csiVGS,
|
|
}, nil
|
|
}
|
|
|
|
func (cs *ControllerServer) DeleteVolumeGroupSnapshot(
|
|
ctx context.Context,
|
|
req *csi.DeleteVolumeGroupSnapshotRequest,
|
|
) (*csi.DeleteVolumeGroupSnapshotResponse, error) {
|
|
// FIXME: more checking of the request in needed
|
|
// 1. verify that all snapshots in the request are all snapshots in the group
|
|
// 2. delete the group
|
|
|
|
groupSnapshotID := req.GetGroupSnapshotId()
|
|
|
|
// Existence and conflict checks
|
|
if acquired := cs.VolumeGroupLocks.TryAcquire(groupSnapshotID); !acquired {
|
|
log.ErrorLog(ctx, util.SnapshotOperationAlreadyExistsFmt, groupSnapshotID)
|
|
|
|
return nil, status.Errorf(codes.Aborted, util.SnapshotOperationAlreadyExistsFmt, groupSnapshotID)
|
|
}
|
|
defer cs.VolumeGroupLocks.Release(groupSnapshotID)
|
|
|
|
mgr := NewManager(cs.Driver.GetInstanceID(), nil, req.GetSecrets())
|
|
defer mgr.Destroy(ctx)
|
|
|
|
groupSnapshot, err := mgr.GetVolumeGroupSnapshotByID(ctx, groupSnapshotID)
|
|
if err != nil {
|
|
return nil, status.Errorf(
|
|
codes.Internal,
|
|
"failed to get volume group snapshot with id %q: %v",
|
|
groupSnapshotID, err)
|
|
}
|
|
defer groupSnapshot.Destroy(ctx)
|
|
|
|
err = groupSnapshot.Delete(ctx)
|
|
if err != nil {
|
|
return nil, status.Errorf(
|
|
codes.Internal,
|
|
"failed to delete volume group snapshot %q: %v",
|
|
groupSnapshot, err)
|
|
}
|
|
|
|
return &csi.DeleteVolumeGroupSnapshotResponse{}, nil
|
|
}
|
|
|
|
// GetVolumeGroupSnapshot is sortof optional, only used for
|
|
// static/pre-provisioned VolumeGroupSnapshots.
|
|
func (cs *ControllerServer) GetVolumeGroupSnapshot(
|
|
ctx context.Context,
|
|
req *csi.GetVolumeGroupSnapshotRequest,
|
|
) (*csi.GetVolumeGroupSnapshotResponse, error) {
|
|
groupSnapshotID := req.GetGroupSnapshotId()
|
|
|
|
// Existence and conflict checks
|
|
if acquired := cs.VolumeGroupLocks.TryAcquire(groupSnapshotID); !acquired {
|
|
log.ErrorLog(ctx, util.SnapshotOperationAlreadyExistsFmt, groupSnapshotID)
|
|
|
|
return nil, status.Errorf(codes.Aborted, util.SnapshotOperationAlreadyExistsFmt, groupSnapshotID)
|
|
}
|
|
defer cs.VolumeGroupLocks.Release(groupSnapshotID)
|
|
|
|
mgr := NewManager(cs.Driver.GetInstanceID(), nil, req.GetSecrets())
|
|
defer mgr.Destroy(ctx)
|
|
|
|
groupSnapshot, err := mgr.GetVolumeGroupSnapshotByID(ctx, groupSnapshotID)
|
|
if err != nil {
|
|
return nil, status.Errorf(
|
|
codes.Internal,
|
|
"failed to get volume group snapshot with id %q: %v",
|
|
groupSnapshotID, err)
|
|
}
|
|
defer groupSnapshot.Destroy(ctx)
|
|
|
|
csiVGS, err := groupSnapshot.ToCSI(ctx)
|
|
if err != nil {
|
|
return nil, status.Error(codes.Aborted, err.Error())
|
|
}
|
|
|
|
return &csi.GetVolumeGroupSnapshotResponse{
|
|
GroupSnapshot: csiVGS,
|
|
}, nil
|
|
}
|