1
0
Fork 0
mirror of https://github.com/moby/moby.git synced 2022-11-09 12:21:53 -05:00
moby--moby/container/container_windows.go
Kir Kolyshkin 7120976d74 Implement none, private, and shareable ipc modes
Since the commit d88fe447df ("Add support for sharing /dev/shm/ and
/dev/mqueue between containers") container's /dev/shm is mounted on the
host first, then bind-mounted inside the container. This is done that
way in order to be able to share this container's IPC namespace
(and the /dev/shm mount point) with another container.

Unfortunately, this functionality breaks container checkpoint/restore
(even if IPC is not shared). Since /dev/shm is an external mount, its
contents is not saved by `criu checkpoint`, and so upon restore any
application that tries to access data under /dev/shm is severily
disappointed (which usually results in a fatal crash).

This commit solves the issue by introducing new IPC modes for containers
(in addition to 'host' and 'container:ID'). The new modes are:

 - 'shareable':	enables sharing this container's IPC with others
		(this used to be the implicit default);

 - 'private':	disables sharing this container's IPC.

In 'private' mode, container's /dev/shm is truly mounted inside the
container, without any bind-mounting from the host, which solves the
issue.

While at it, let's also implement 'none' mode. The motivation, as
eloquently put by Justin Cormack, is:

> I wondered a while back about having a none shm mode, as currently it is
> not possible to have a totally unwriteable container as there is always
> a /dev/shm writeable mount. It is a bit of a niche case (and clearly
> should never be allowed to be daemon default) but it would be trivial to
> add now so maybe we should...

...so here's yet yet another mode:

 - 'none':	no /dev/shm mount inside the container (though it still
		has its own private IPC namespace).

Now, to ultimately solve the abovementioned checkpoint/restore issue, we'd
need to make 'private' the default mode, but unfortunately it breaks the
backward compatibility. So, let's make the default container IPC mode
per-daemon configurable (with the built-in default set to 'shareable'
for now). The default can be changed either via a daemon CLI option
(--default-shm-mode) or a daemon.json configuration file parameter
of the same name.

Note one can only set either 'shareable' or 'private' IPC modes as a
daemon default (i.e. in this context 'host', 'container', or 'none'
do not make much sense).

Some other changes this patch introduces are:

1. A mount for /dev/shm is added to default OCI Linux spec.

2. IpcMode.Valid() is simplified to remove duplicated code that parsed
   'container:ID' form. Note the old version used to check that ID does
   not contain a semicolon -- this is no longer the case (tests are
   modified accordingly). The motivation is we should either do a
   proper check for container ID validity, or don't check it at all
   (since it is checked in other places anyway). I chose the latter.

3. IpcMode.Container() is modified to not return container ID if the
   mode value does not start with "container:", unifying the check to
   be the same as in IpcMode.IsContainer().

3. IPC mode unit tests (runconfig/hostconfig_test.go) are modified
   to add checks for newly added values.

[v2: addressed review at https://github.com/moby/moby/pull/34087#pullrequestreview-51345997]
[v3: addressed review at https://github.com/moby/moby/pull/34087#pullrequestreview-53902833]
[v4: addressed the case of upgrading from older daemon, in this case
     container.HostConfig.IpcMode is unset and this is valid]
[v5: document old and new IpcMode values in api/swagger.yaml]
[v6: add the 'none' mode, changelog entry to docs/api/version-history.md]

Signed-off-by: Kir Kolyshkin <kolyshkin@gmail.com>
2017-08-14 10:50:39 +03:00

211 lines
6.4 KiB
Go

// +build windows
package container
import (
"fmt"
"os"
"path/filepath"
"github.com/docker/docker/api/types"
containertypes "github.com/docker/docker/api/types/container"
"github.com/docker/docker/pkg/system"
)
const (
containerSecretMountPath = `C:\ProgramData\Docker\secrets`
containerInternalSecretMountPath = `C:\ProgramData\Docker\internal\secrets`
containerInternalConfigsDirPath = `C:\ProgramData\Docker\internal\configs`
)
// ExitStatus provides exit reasons for a container.
type ExitStatus struct {
// The exit code with which the container exited.
ExitCode int
}
// UnmountIpcMount unmounts Ipc related mounts.
// This is a NOOP on windows.
func (container *Container) UnmountIpcMount(unmount func(pth string) error) error {
return nil
}
// IpcMounts returns the list of Ipc related mounts.
func (container *Container) IpcMounts() []Mount {
return nil
}
// CreateSecretSymlinks creates symlinks to files in the secret mount.
func (container *Container) CreateSecretSymlinks() error {
for _, r := range container.SecretReferences {
if r.File == nil {
continue
}
resolvedPath, _, err := container.ResolvePath(getSecretTargetPath(r))
if err != nil {
return err
}
if err := system.MkdirAll(filepath.Dir(resolvedPath), 0, ""); err != nil {
return err
}
if err := os.Symlink(filepath.Join(containerInternalSecretMountPath, r.SecretID), resolvedPath); err != nil {
return err
}
}
return nil
}
// SecretMounts returns the mount for the secret path.
// All secrets are stored in a single mount on Windows. Target symlinks are
// created for each secret, pointing to the files in this mount.
func (container *Container) SecretMounts() []Mount {
var mounts []Mount
if len(container.SecretReferences) > 0 {
mounts = append(mounts, Mount{
Source: container.SecretMountPath(),
Destination: containerInternalSecretMountPath,
Writable: false,
})
}
return mounts
}
// UnmountSecrets unmounts the fs for secrets
func (container *Container) UnmountSecrets() error {
return os.RemoveAll(container.SecretMountPath())
}
// CreateConfigSymlinks creates symlinks to files in the config mount.
func (container *Container) CreateConfigSymlinks() error {
for _, configRef := range container.ConfigReferences {
if configRef.File == nil {
continue
}
resolvedPath, _, err := container.ResolvePath(configRef.File.Name)
if err != nil {
return err
}
if err := system.MkdirAll(filepath.Dir(resolvedPath), 0, ""); err != nil {
return err
}
if err := os.Symlink(filepath.Join(containerInternalConfigsDirPath, configRef.ConfigID), resolvedPath); err != nil {
return err
}
}
return nil
}
// ConfigMounts returns the mount for configs.
// All configs are stored in a single mount on Windows. Target symlinks are
// created for each config, pointing to the files in this mount.
func (container *Container) ConfigMounts() []Mount {
var mounts []Mount
if len(container.ConfigReferences) > 0 {
mounts = append(mounts, Mount{
Source: container.ConfigsDirPath(),
Destination: containerInternalConfigsDirPath,
Writable: false,
})
}
return mounts
}
// DetachAndUnmount unmounts all volumes.
// On Windows it only delegates to `UnmountVolumes` since there is nothing to
// force unmount.
func (container *Container) DetachAndUnmount(volumeEventLog func(name, action string, attributes map[string]string)) error {
return container.UnmountVolumes(volumeEventLog)
}
// TmpfsMounts returns the list of tmpfs mounts
func (container *Container) TmpfsMounts() ([]Mount, error) {
var mounts []Mount
return mounts, nil
}
// UpdateContainer updates configuration of a container. Callers must hold a Lock on the Container.
func (container *Container) UpdateContainer(hostConfig *containertypes.HostConfig) error {
resources := hostConfig.Resources
if resources.CPUShares != 0 ||
resources.Memory != 0 ||
resources.NanoCPUs != 0 ||
resources.CgroupParent != "" ||
resources.BlkioWeight != 0 ||
len(resources.BlkioWeightDevice) != 0 ||
len(resources.BlkioDeviceReadBps) != 0 ||
len(resources.BlkioDeviceWriteBps) != 0 ||
len(resources.BlkioDeviceReadIOps) != 0 ||
len(resources.BlkioDeviceWriteIOps) != 0 ||
resources.CPUPeriod != 0 ||
resources.CPUQuota != 0 ||
resources.CPURealtimePeriod != 0 ||
resources.CPURealtimeRuntime != 0 ||
resources.CpusetCpus != "" ||
resources.CpusetMems != "" ||
len(resources.Devices) != 0 ||
len(resources.DeviceCgroupRules) != 0 ||
resources.DiskQuota != 0 ||
resources.KernelMemory != 0 ||
resources.MemoryReservation != 0 ||
resources.MemorySwap != 0 ||
resources.MemorySwappiness != nil ||
resources.OomKillDisable != nil ||
resources.PidsLimit != 0 ||
len(resources.Ulimits) != 0 ||
resources.CPUCount != 0 ||
resources.CPUPercent != 0 ||
resources.IOMaximumIOps != 0 ||
resources.IOMaximumBandwidth != 0 {
return fmt.Errorf("resource updating isn't supported on Windows")
}
// update HostConfig of container
if hostConfig.RestartPolicy.Name != "" {
if container.HostConfig.AutoRemove && !hostConfig.RestartPolicy.IsNone() {
return fmt.Errorf("Restart policy cannot be updated because AutoRemove is enabled for the container")
}
container.HostConfig.RestartPolicy = hostConfig.RestartPolicy
}
return nil
}
// cleanResourcePath cleans a resource path by removing C:\ syntax, and prepares
// to combine with a volume path
func cleanResourcePath(path string) string {
if len(path) >= 2 {
c := path[0]
if path[1] == ':' && ('a' <= c && c <= 'z' || 'A' <= c && c <= 'Z') {
path = path[2:]
}
}
return filepath.Join(string(os.PathSeparator), path)
}
// BuildHostnameFile writes the container's hostname file.
func (container *Container) BuildHostnameFile() error {
return nil
}
// EnableServiceDiscoveryOnDefaultNetwork Enable service discovery on default network
func (container *Container) EnableServiceDiscoveryOnDefaultNetwork() bool {
return true
}
// GetMountPoints gives a platform specific transformation to types.MountPoint. Callers must hold a Container lock.
func (container *Container) GetMountPoints() []types.MountPoint {
mountPoints := make([]types.MountPoint, 0, len(container.MountPoints))
for _, m := range container.MountPoints {
mountPoints = append(mountPoints, types.MountPoint{
Type: m.Type,
Name: m.Name,
Source: m.Path(),
Destination: m.Destination,
Driver: m.Driver,
RW: m.RW,
})
}
return mountPoints
}