runc/checkpoint.go

137 lines
4.5 KiB
Go
Raw Normal View History

// +build linux
package main
import (
"fmt"
"strconv"
"strings"
"github.com/opencontainers/runc/libcontainer"
"github.com/opencontainers/runtime-spec/specs-go"
"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"},
checkpoint: support lazy migration With the help of userfaultfd CRIU supports lazy migration. Lazy migration means that memory pages are only transferred from the migration source to the migration destination on page fault. This enables to reduce the downtime during process or container migration to a minimum as the memory does not need to be transferred during migration. Lazy migration currently depends on userfaultfd being available on the current Linux kernel and if the used CRIU version supports lazy migration. Both dependencies can be checked by querying CRIU via RPC if the lazy migration feature is available. Using feature checking instead of version comparison enables runC to use CRIU features from the criu-dev branch. This way the user can decide if lazy migration should be available by choosing the right kernel and CRIU branch. To use lazy migration the CRIU process during dump needs to dump everything besides the memory pages and then it opens a network port waiting for remote page fault requests: # runc checkpoint httpd --lazy-pages --page-server 0.0.0.0:27 \ --status-fd /tmp/postcopy-pipe In this example CRIU will hang/wait once it has opened the network port and wait for network connection. As runC waits for CRIU to finish it will also hang until the lazy migration has finished. To know when the restore on the destination side can start the '--status-fd' parameter is used: #️ runc checkpoint --help | grep status --status-fd value criu writes \0 to this FD once lazy-pages is ready The parameter '--status-fd' is directly from CRIU and this way the process outside of runC which controls the migration knows exactly when to transfer the checkpoint (without memory pages) to the destination and that the restore can be started. On the destination side it is necessary to start CRIU in 'lazy-pages' mode like this: # criu lazy-pages --page-server --address 192.168.122.3 --port 27 \ -D checkpoint and tell runC to do a lazy restore: # runc restore -d --image-path checkpoint --work-path checkpoint \ --lazy-pages httpd If both processes on the restore side have the same working directory 'criu lazy-pages' creates a unix domain socket where it waits for requests from the actual restore. runC starts CRIU restore in lazy restore mode and talks to 'criu lazy-pages' that it wants to restore memory pages on demand. CRIU continues to restore the process and once the process is running and accesses the first non-existing memory page the 'criu lazy-pages' server will request the page from the source system. Thus all pages from the source system will be transferred to the destination system. Once all pages have been transferred runC on the source system will end and the container will have finished migration. This can also be combined with CRIU's pre-copy support. The combination of pre-copy and post-copy (lazy migration) provides the possibility to migrate containers with minimal downtimes. Some additional background about post-copy migration can be found in these articles: https://lisas.de/~adrian/?p=1253 https://lisas.de/~adrian/?p=1183 Signed-off-by: Adrian Reber <areber@redhat.com>
2017-07-24 23:43:14 +08:00
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 isRootless() {
return fmt.Errorf("runc checkpoint requires root")
}
container, err := getContainer(context)
if err != nil {
return err
}
status, err := container.Status()
if err != nil {
return err
}
if status == libcontainer.Created {
fatalf("Container cannot be checkpointed in created state")
}
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 {
var nsmask int
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
}