Files
runc/checkpoint.go
T
Akihiro Suda 06f789cf26 Disable rootless mode except RootlessCgMgr when executed as the root in userns
This PR decomposes `libcontainer/configs.Config.Rootless bool` into `RootlessEUID bool` and
`RootlessCgroups bool`, so as to make "runc-in-userns" to be more compatible with "rootful" runc.

`RootlessEUID` denotes that runc is being executed as a non-root user (euid != 0) in
the current user namespace. `RootlessEUID` is almost identical to the former `Rootless`
except cgroups stuff.

`RootlessCgroups` denotes that runc is unlikely to have the full access to cgroups.
`RootlessCgroups` is set to false if runc is executed as the root (euid == 0) in the initial namespace.
Otherwise `RootlessCgroups` is set to true.
(Hint: if `RootlessEUID` is true, `RootlessCgroups` becomes true as well)

When runc is executed as the root (euid == 0) in an user namespace (e.g. by Docker-in-LXD, Podman, Usernetes),
`RootlessEUID` is set to false but `RootlessCgroups` is set to true.
So, "runc-in-userns" behaves almost same as "rootful" runc except that cgroups errors are ignored.

This PR does not have any impact on CLI flags and `state.json`.

Note about CLI:
* Now `runc --rootless=(auto|true|false)` CLI flag is only used for setting `RootlessCgroups`.
* Now `runc spec --rootless` is only required when `RootlessEUID` is set to true.
  For runc-in-userns, `runc spec`  without `--rootless` should work, when sufficient numbers of
  UID/GID are mapped.

Note about `$XDG_RUNTIME_DIR` (e.g. `/run/user/1000`):
* `$XDG_RUNTIME_DIR` is ignored if runc is being executed as the root (euid == 0) in the initial namespace, for backward compatibility.
  (`/run/runc` is used)
* If runc is executed as the root (euid == 0) in an user namespace, `$XDG_RUNTIME_DIR` is honored if `$USER != "" && $USER != "root"`.
  This allows unprivileged users to allow execute runc as the root in userns, without mounting writable `/run/runc`.

Note about `state.json`:
* `rootless` is set to true when `RootlessEUID == true && RootlessCgroups == true`.

Signed-off-by: Akihiro Suda <suda.akihiro@lab.ntt.co.jp>
2018-09-07 15:05:03 +09:00

141 lines
4.7 KiB
Go

// +build linux
package main
import (
"fmt"
"os"
"strconv"
"strings"
"github.com/opencontainers/runc/libcontainer"
"github.com/opencontainers/runc/libcontainer/system"
"github.com/opencontainers/runtime-spec/specs-go"
"github.com/sirupsen/logrus"
"github.com/urfave/cli"
"golang.org/x/sys/unix"
)
var checkpointCommand = cli.Command{
Name: "checkpoint",
Usage: "checkpoint a running container",
ArgsUsage: `<container-id>
Where "<container-id>" is the name for the instance of the container to be
checkpointed.`,
Description: `The checkpoint command saves the state of the container instance.`,
Flags: []cli.Flag{
cli.StringFlag{Name: "image-path", Value: "", Usage: "path for saving criu image files"},
cli.StringFlag{Name: "work-path", Value: "", Usage: "path for saving work files and logs"},
cli.StringFlag{Name: "parent-path", Value: "", Usage: "path for previous criu image files in pre-dump"},
cli.BoolFlag{Name: "leave-running", Usage: "leave the process running after checkpointing"},
cli.BoolFlag{Name: "tcp-established", Usage: "allow open tcp connections"},
cli.BoolFlag{Name: "ext-unix-sk", Usage: "allow external unix sockets"},
cli.BoolFlag{Name: "shell-job", Usage: "allow shell jobs"},
cli.BoolFlag{Name: "lazy-pages", Usage: "use userfaultfd to lazily restore memory pages"},
cli.StringFlag{Name: "status-fd", Value: "", Usage: "criu writes \\0 to this FD once lazy-pages is ready"},
cli.StringFlag{Name: "page-server", Value: "", Usage: "ADDRESS:PORT of the page server"},
cli.BoolFlag{Name: "file-locks", Usage: "handle file locks, for safety"},
cli.BoolFlag{Name: "pre-dump", Usage: "dump container's memory information only, leave the container running after this"},
cli.StringFlag{Name: "manage-cgroups-mode", Value: "", Usage: "cgroups mode: 'soft' (default), 'full' and 'strict'"},
cli.StringSliceFlag{Name: "empty-ns", Usage: "create a namespace, but don't restore its properties"},
cli.BoolFlag{Name: "auto-dedup", Usage: "enable auto deduplication of memory images"},
},
Action: func(context *cli.Context) error {
if err := checkArgs(context, 1, exactArgs); err != nil {
return err
}
// XXX: Currently this is untested with rootless containers.
if os.Geteuid() != 0 || system.RunningInUserNS() {
logrus.Warn("runc checkpoint is untested with rootless containers")
}
container, err := getContainer(context)
if err != nil {
return err
}
status, err := container.Status()
if err != nil {
return err
}
if status == libcontainer.Created || status == libcontainer.Stopped {
fatalf("Container cannot be checkpointed in %s state", status.String())
}
defer destroy(container)
options := criuOptions(context)
// these are the mandatory criu options for a container
setPageServer(context, options)
setManageCgroupsMode(context, options)
if err := setEmptyNsMask(context, options); err != nil {
return err
}
if err := container.Checkpoint(options); err != nil {
return err
}
return nil
},
}
func getCheckpointImagePath(context *cli.Context) string {
imagePath := context.String("image-path")
if imagePath == "" {
imagePath = getDefaultImagePath(context)
}
return imagePath
}
func setPageServer(context *cli.Context, options *libcontainer.CriuOpts) {
// xxx following criu opts are optional
// The dump image can be sent to a criu page server
if psOpt := context.String("page-server"); psOpt != "" {
addressPort := strings.Split(psOpt, ":")
if len(addressPort) != 2 {
fatal(fmt.Errorf("Use --page-server ADDRESS:PORT to specify page server"))
}
portInt, err := strconv.Atoi(addressPort[1])
if err != nil {
fatal(fmt.Errorf("Invalid port number"))
}
options.PageServer = libcontainer.CriuPageServerInfo{
Address: addressPort[0],
Port: int32(portInt),
}
}
}
func setManageCgroupsMode(context *cli.Context, options *libcontainer.CriuOpts) {
if cgOpt := context.String("manage-cgroups-mode"); cgOpt != "" {
switch cgOpt {
case "soft":
options.ManageCgroupsMode = libcontainer.CRIU_CG_MODE_SOFT
case "full":
options.ManageCgroupsMode = libcontainer.CRIU_CG_MODE_FULL
case "strict":
options.ManageCgroupsMode = libcontainer.CRIU_CG_MODE_STRICT
default:
fatal(fmt.Errorf("Invalid manage cgroups mode"))
}
}
}
var namespaceMapping = map[specs.LinuxNamespaceType]int{
specs.NetworkNamespace: unix.CLONE_NEWNET,
}
func setEmptyNsMask(context *cli.Context, options *libcontainer.CriuOpts) error {
/* Runc doesn't manage network devices and their configuration */
nsmask := unix.CLONE_NEWNET
for _, ns := range context.StringSlice("empty-ns") {
f, exists := namespaceMapping[specs.LinuxNamespaceType(ns)]
if !exists {
return fmt.Errorf("namespace %q is not supported", ns)
}
nsmask |= f
}
options.EmptyNs = uint32(nsmask)
return nil
}