The following pull request was submitted through Github.
It can be accessed and reviewed at: https://github.com/lxc/lxd/pull/2890
This e-mail was sent by the LXC bot, direct replies will not reach the author
unless they happen to be subscribed to this list.
=== Description (from pull-request) ===
Signed-off-by: Christian Brauner <christian.brau...@ubuntu.com>
From 53c312d1f985d94c6ccd0807744f42482730259d Mon Sep 17 00:00:00 2001
From: Christian Brauner <christian.brau...@ubuntu.com>
Date: Wed, 15 Feb 2017 17:38:53 +0100
Subject: [PATCH] storage: lock Container{U,M}ount for zfs, lvm
Signed-off-by: Christian Brauner <christian.brau...@ubuntu.com>
---
lxd/storage.go | 4 +--
lxd/storage_btrfs.go | 14 +++++-----
lxd/storage_lvm.go | 72 +++++++++++++++++++++++++++++++++++++++------------
lxd/storage_zfs.go | 73 +++++++++++++++++++++++++++++++++++++++-------------
4 files changed, 119 insertions(+), 44 deletions(-)
diff --git a/lxd/storage.go b/lxd/storage.go
index f90cc74..9023031 100644
--- a/lxd/storage.go
+++ b/lxd/storage.go
@@ -22,8 +22,8 @@ import (
log "gopkg.in/inconshreveable/log15.v2"
)
-var imageCreationInPool = map[string]chan bool{}
-var imageCreationInPoolLock sync.Mutex
+var lxdStorageLockMap = map[string]chan bool{}
+var lxdStorageLock sync.Mutex
/* Some interesting filesystems */
const (
diff --git a/lxd/storage_btrfs.go b/lxd/storage_btrfs.go
index d847f34..ebcecdb 100644
--- a/lxd/storage_btrfs.go
+++ b/lxd/storage_btrfs.go
@@ -571,23 +571,23 @@ func (s *storageBtrfs) ContainerCreateFromImage(container
container, fingerprint
// ${LXD_DIR}/images/<fingerprint>
imageMntPoint := getImageMountPoint(s.pool.Name, fingerprint)
imageStoragePoolLockID := fmt.Sprintf("%s/%s", s.pool.Name, fingerprint)
- imageCreationInPoolLock.Lock()
- if waitChannel, ok := imageCreationInPool[imageStoragePoolLockID]; ok {
- imageCreationInPoolLock.Unlock()
+ lxdStorageLock.Lock()
+ if waitChannel, ok := lxdStorageLockMap[imageStoragePoolLockID]; ok {
+ lxdStorageLock.Unlock()
if _, ok := <-waitChannel; ok {
shared.LogWarnf("Value transmitted over image lock
semaphore?")
}
} else {
- imageCreationInPool[imageStoragePoolLockID] = make(chan bool)
+ lxdStorageLockMap[imageStoragePoolLockID] = make(chan bool)
var imgerr error
if !shared.PathExists(imageMntPoint) ||
!s.isBtrfsPoolVolume(imageMntPoint) {
imgerr = s.ImageCreate(fingerprint)
}
- if waitChannel, ok :=
imageCreationInPool[imageStoragePoolLockID]; ok {
+ if waitChannel, ok :=
lxdStorageLockMap[imageStoragePoolLockID]; ok {
close(waitChannel)
- delete(imageCreationInPool, imageStoragePoolLockID)
+ delete(lxdStorageLockMap, imageStoragePoolLockID)
}
- imageCreationInPoolLock.Unlock()
+ lxdStorageLock.Unlock()
if imgerr != nil {
return imgerr
}
diff --git a/lxd/storage_lvm.go b/lxd/storage_lvm.go
index 4afda74..7ae48ce 100644
--- a/lxd/storage_lvm.go
+++ b/lxd/storage_lvm.go
@@ -547,23 +547,23 @@ func (s *storageLvm) ContainerCreateFromImage(container
container, fingerprint s
imageLvmDevPath := getLvmDevPath(s.pool.Name,
storagePoolVolumeApiEndpointImages, fingerprint)
imageStoragePoolLockID := fmt.Sprintf("%s/%s", s.pool.Name, fingerprint)
- imageCreationInPoolLock.Lock()
- if waitChannel, ok := imageCreationInPool[imageStoragePoolLockID]; ok {
- imageCreationInPoolLock.Unlock()
+ lxdStorageLock.Lock()
+ if waitChannel, ok := lxdStorageLockMap[imageStoragePoolLockID]; ok {
+ lxdStorageLock.Unlock()
if _, ok := <-waitChannel; ok {
shared.LogWarnf("Value transmitted over image lock
semaphore?")
}
} else {
- imageCreationInPool[imageStoragePoolLockID] = make(chan bool)
+ lxdStorageLockMap[imageStoragePoolLockID] = make(chan bool)
var imgerr error
if !shared.PathExists(imageMntPoint) ||
!shared.PathExists(imageLvmDevPath) {
imgerr = s.ImageCreate(fingerprint)
}
- if waitChannel, ok :=
imageCreationInPool[imageStoragePoolLockID]; ok {
+ if waitChannel, ok :=
lxdStorageLockMap[imageStoragePoolLockID]; ok {
close(waitChannel)
- delete(imageCreationInPool, imageStoragePoolLockID)
+ delete(lxdStorageLockMap, imageStoragePoolLockID)
}
- imageCreationInPoolLock.Unlock()
+ lxdStorageLock.Unlock()
if imgerr != nil {
return imgerr
}
@@ -759,31 +759,69 @@ func (s *storageLvm) ContainerMount(name string, path
string) (bool, error) {
mountOptions := s.volume.Config["block.mount_options"]
containerMntPoint := getContainerMountPoint(s.pool.Name, name)
- if shared.IsMountPoint(containerMntPoint) {
+ containerMountLockID := fmt.Sprintf("mount/%s/%s", s.pool.Name, name)
+ lxdStorageLock.Lock()
+ if waitChannel, ok := lxdStorageLockMap[containerMountLockID]; ok {
+ lxdStorageLock.Unlock()
+ if _, ok := <-waitChannel; ok {
+ shared.LogWarnf("Value transmitted over image lock
semaphore?")
+ }
+ // Give the benefit of the doubt and assume that the other
+ // thread actually succeeded in mounting the storage volume.
return false, nil
}
- err := tryMount(containerLvmPath, containerMntPoint, lvFsType, 0,
mountOptions)
- if err != nil {
- return false, fmt.Errorf("Error mounting snapshot LV path='%s':
%s", path, err)
+ lxdStorageLockMap[containerMountLockID] = make(chan bool)
+ var imgerr error
+ ourMount := false
+ if !shared.IsMountPoint(containerMntPoint) {
+ imgerr = tryMount(containerLvmPath, containerMntPoint,
lvFsType, 0, mountOptions)
+ ourMount = true
+ }
+ if waitChannel, ok := lxdStorageLockMap[containerMountLockID]; ok {
+ close(waitChannel)
+ delete(lxdStorageLockMap, containerMountLockID)
+ }
+ lxdStorageLock.Unlock()
+ if imgerr != nil {
+ return false, imgerr
}
- return true, nil
+ return ourMount, nil
}
func (s *storageLvm) ContainerUmount(name string, path string) (bool, error) {
containerMntPoint := getContainerMountPoint(s.pool.Name, name)
- if !shared.IsMountPoint(containerMntPoint) {
+ containerUmountLockID := fmt.Sprintf("umount/%s/%s", s.pool.Name, name)
+ lxdStorageLock.Lock()
+ if waitChannel, ok := lxdStorageLockMap[containerUmountLockID]; ok {
+ lxdStorageLock.Unlock()
+ if _, ok := <-waitChannel; ok {
+ shared.LogWarnf("Value transmitted over image lock
semaphore?")
+ }
+ // Give the benefit of the doubt and assume that the other
+ // thread actually succeeded in unmounting the storage volume.
return false, nil
}
- err := tryUnmount(containerMntPoint, 0)
- if err != nil {
- return false, fmt.Errorf("failed to unmount container path
'%s': %s", path, err)
+ lxdStorageLockMap[containerUmountLockID] = make(chan bool)
+ var imgerr error
+ ourUmount := false
+ if shared.IsMountPoint(containerMntPoint) {
+ imgerr = tryUnmount(containerMntPoint, 0)
+ ourUmount = true
+ }
+ if waitChannel, ok := lxdStorageLockMap[containerUmountLockID]; ok {
+ close(waitChannel)
+ delete(lxdStorageLockMap, containerUmountLockID)
+ }
+ lxdStorageLock.Unlock()
+ if imgerr != nil {
+ return false, imgerr
}
- return true, nil
+ return ourUmount, nil
}
func (s *storageLvm) ContainerRename(container container, newContainerName
string) error {
diff --git a/lxd/storage_zfs.go b/lxd/storage_zfs.go
index 0e83884..0ad6eda 100644
--- a/lxd/storage_zfs.go
+++ b/lxd/storage_zfs.go
@@ -398,33 +398,70 @@ func (s *storageZfs) ContainerMount(name string, path
string) (bool, error) {
fs := fmt.Sprintf("containers/%s", name)
containerPoolVolumeMntPoint := getContainerMountPoint(s.pool.Name, name)
- // Just in case the container filesystem got unmounted
- if shared.IsMountPoint(containerPoolVolumeMntPoint) {
+ containerMountLockID := fmt.Sprintf("mount/%s/%s", s.pool.Name, name)
+ lxdStorageLock.Lock()
+ if waitChannel, ok := lxdStorageLockMap[containerMountLockID]; ok {
+ lxdStorageLock.Unlock()
+ if _, ok := <-waitChannel; ok {
+ shared.LogWarnf("Value transmitted over image lock
semaphore?")
+ }
+ // Give the benefit of the doubt and assume that the other
+ // thread actually succeeded in mounting the storage volume.
return false, nil
}
- err := s.zfsPoolVolumeMount(fs)
- if err != nil {
- return false, err
+ lxdStorageLockMap[containerMountLockID] = make(chan bool)
+ var imgerr error
+ ourMount := false
+ if !shared.IsMountPoint(containerPoolVolumeMntPoint) {
+ imgerr = s.zfsPoolVolumeMount(fs)
+ ourMount = true
+ }
+ if waitChannel, ok := lxdStorageLockMap[containerMountLockID]; ok {
+ close(waitChannel)
+ delete(lxdStorageLockMap, containerMountLockID)
+ }
+ lxdStorageLock.Unlock()
+ if imgerr != nil {
+ return false, imgerr
}
- return true, nil
+ return ourMount, nil
}
func (s *storageZfs) ContainerUmount(name string, path string) (bool, error) {
fs := fmt.Sprintf("containers/%s", name)
containerPoolVolumeMntPoint := getContainerMountPoint(s.pool.Name, name)
- if !shared.IsMountPoint(containerPoolVolumeMntPoint) {
+ containerUmountLockID := fmt.Sprintf("umount/%s/%s", s.pool.Name, name)
+ lxdStorageLock.Lock()
+ if waitChannel, ok := lxdStorageLockMap[containerUmountLockID]; ok {
+ lxdStorageLock.Unlock()
+ if _, ok := <-waitChannel; ok {
+ shared.LogWarnf("Value transmitted over image lock
semaphore?")
+ }
+ // Give the benefit of the doubt and assume that the other
+ // thread actually succeeded in unmounting the storage volume.
return false, nil
}
- err := s.zfsPoolVolumeUmount(fs)
- if err != nil {
- return false, err
+ lxdStorageLockMap[containerUmountLockID] = make(chan bool)
+ var imgerr error
+ ourUmount := false
+ if shared.IsMountPoint(containerPoolVolumeMntPoint) {
+ imgerr = s.zfsPoolVolumeUmount(fs)
+ ourUmount = true
+ }
+ if waitChannel, ok := lxdStorageLockMap[containerUmountLockID]; ok {
+ close(waitChannel)
+ delete(lxdStorageLockMap, containerUmountLockID)
+ }
+ lxdStorageLock.Unlock()
+ if imgerr != nil {
+ return false, imgerr
}
- return true, nil
+ return ourUmount, nil
}
// Things we do have to care about
@@ -477,23 +514,23 @@ func (s *storageZfs) ContainerCreateFromImage(container
container, fingerprint s
fsImage := fmt.Sprintf("images/%s", fingerprint)
imageStoragePoolLockID := fmt.Sprintf("%s/%s", s.pool.Name, fingerprint)
- imageCreationInPoolLock.Lock()
- if waitChannel, ok := imageCreationInPool[imageStoragePoolLockID]; ok {
- imageCreationInPoolLock.Unlock()
+ lxdStorageLock.Lock()
+ if waitChannel, ok := lxdStorageLockMap[imageStoragePoolLockID]; ok {
+ lxdStorageLock.Unlock()
if _, ok := <-waitChannel; ok {
shared.LogWarnf("Value transmitted over image lock
semaphore?")
}
} else {
- imageCreationInPool[imageStoragePoolLockID] = make(chan bool)
+ lxdStorageLockMap[imageStoragePoolLockID] = make(chan bool)
var imgerr error
if !s.zfsPoolVolumeExists(fsImage) {
imgerr = s.ImageCreate(fingerprint)
}
- if waitChannel, ok :=
imageCreationInPool[imageStoragePoolLockID]; ok {
+ if waitChannel, ok :=
lxdStorageLockMap[imageStoragePoolLockID]; ok {
close(waitChannel)
- delete(imageCreationInPool, imageStoragePoolLockID)
+ delete(lxdStorageLockMap, imageStoragePoolLockID)
}
- imageCreationInPoolLock.Unlock()
+ lxdStorageLock.Unlock()
if imgerr != nil {
return imgerr
}
_______________________________________________
lxc-devel mailing list
lxc-devel@lists.linuxcontainers.org
http://lists.linuxcontainers.org/listinfo/lxc-devel
