// appToSystemd transforms the provided RuntimeApp+ImageManifest into systemd units
func appToSystemd(p *stage1commontypes.Pod, ra *schema.RuntimeApp, interactive bool, flavor string, privateUsers string) error {
app := ra.App
appName := ra.Name
imgName := p.AppNameToImageName(appName)
if len(app.Exec) == 0 {
return fmt.Errorf(`image %q has an empty "exec" (try --exec=BINARY)`, imgName)
}
workDir := "/"
if app.WorkingDirectory != "" {
workDir = app.WorkingDirectory
}
env := app.Environment
env.Set("AC_APP_NAME", appName.String())
if p.MetadataServiceURL != "" {
env.Set("AC_METADATA_URL", p.MetadataServiceURL)
}
if err := writeEnvFile(p, env, appName, privateUsers); err != nil {
return errwrap.Wrap(errors.New("unable to write environment file"), err)
}
var _uid, gid int
var err error
uidRange := uid.NewBlankUidRange()
if err := uidRange.Deserialize([]byte(privateUsers)); err != nil {
return errwrap.Wrap(errors.New("unable to deserialize uid range"), err)
}
if strings.HasPrefix(app.User, "/") {
var stat syscall.Stat_t
if err = syscall.Lstat(filepath.Join(common.AppRootfsPath(p.Root, appName),
app.User), &stat); err != nil {
return errwrap.Wrap(fmt.Errorf("unable to get uid from file %q",
app.User), err)
}
uidReal, _, err := uidRange.UnshiftRange(stat.Uid, 0)
if err != nil {
return errwrap.Wrap(errors.New("unable to determine real uid"), err)
}
_uid = int(uidReal)
} else {
_uid, err = strconv.Atoi(app.User)
if err != nil {
_uid, err = passwd.LookupUidFromFile(app.User,
filepath.Join(common.AppRootfsPath(p.Root, appName), "etc/passwd"))
if err != nil {
return errwrap.Wrap(fmt.Errorf("cannot lookup user %q", app.User), err)
}
}
}
if strings.HasPrefix(app.Group, "/") {
var stat syscall.Stat_t
if err = syscall.Lstat(filepath.Join(common.AppRootfsPath(p.Root, appName),
app.Group), &stat); err != nil {
return errwrap.Wrap(fmt.Errorf("unable to get gid from file %q",
app.Group), err)
}
_, gidReal, err := uidRange.UnshiftRange(0, stat.Gid)
if err != nil {
return errwrap.Wrap(errors.New("unable to determine real gid"), err)
}
gid = int(gidReal)
} else {
gid, err = strconv.Atoi(app.Group)
if err != nil {
gid, err = group.LookupGidFromFile(app.Group,
filepath.Join(common.AppRootfsPath(p.Root, appName), "etc/group"))
if err != nil {
return errwrap.Wrap(fmt.Errorf("cannot lookup group %q", app.Group), err)
}
}
}
execWrap := []string{"/appexec", common.RelAppRootfsPath(appName), workDir, RelEnvFilePath(appName),
strconv.Itoa(_uid), generateGidArg(gid, app.SupplementaryGIDs), "--"}
execStart := quoteExec(append(execWrap, app.Exec...))
opts := []*unit.UnitOption{
unit.NewUnitOption("Unit", "Description", fmt.Sprintf("Application=%v Image=%v", appName, imgName)),
unit.NewUnitOption("Unit", "DefaultDependencies", "false"),
unit.NewUnitOption("Unit", "Wants", fmt.Sprintf("reaper-%s.service", appName)),
unit.NewUnitOption("Service", "Restart", "no"),
unit.NewUnitOption("Service", "ExecStart", execStart),
unit.NewUnitOption("Service", "User", "0"),
unit.NewUnitOption("Service", "Group", "0"),
}
if interactive {
opts = append(opts, unit.NewUnitOption("Service", "StandardInput", "tty"))
opts = append(opts, unit.NewUnitOption("Service", "StandardOutput", "tty"))
opts = append(opts, unit.NewUnitOption("Service", "StandardError", "tty"))
} else {
opts = append(opts, unit.NewUnitOption("Service", "StandardOutput", "journal+console"))
opts = append(opts, unit.NewUnitOption("Service", "StandardError", "journal+console"))
opts = append(opts, unit.NewUnitOption("Service", "SyslogIdentifier", filepath.Base(app.Exec[0])))
}
// When an app fails, we shut down the pod
opts = append(opts, unit.NewUnitOption("Unit", "OnFailure", "halt.target"))
for _, eh := range app.EventHandlers {
var typ string
switch eh.Name {
case "pre-start":
typ = "ExecStartPre"
case "post-stop":
typ = "ExecStopPost"
default:
return fmt.Errorf("unrecognized eventHandler: %v", eh.Name)
}
exec := quoteExec(append(execWrap, eh.Exec...))
opts = append(opts, unit.NewUnitOption("Service", typ, exec))
}
// Some pre-start jobs take a long time, set the timeout to 0
opts = append(opts, unit.NewUnitOption("Service", "TimeoutStartSec", "0"))
var saPorts []types.Port
for _, p := range app.Ports {
if p.SocketActivated {
saPorts = append(saPorts, p)
}
}
for _, i := range app.Isolators {
switch v := i.Value().(type) {
case *types.ResourceMemory:
opts, err = cgroup.MaybeAddIsolator(opts, "memory", v.Limit())
if err != nil {
return err
}
case *types.ResourceCPU:
opts, err = cgroup.MaybeAddIsolator(opts, "cpu", v.Limit())
if err != nil {
return err
}
}
}
if len(saPorts) > 0 {
sockopts := []*unit.UnitOption{
unit.NewUnitOption("Unit", "Description", fmt.Sprintf("Application=%v Image=%v %s", appName, imgName, "socket-activated ports")),
unit.NewUnitOption("Unit", "DefaultDependencies", "false"),
unit.NewUnitOption("Socket", "BindIPv6Only", "both"),
unit.NewUnitOption("Socket", "Service", ServiceUnitName(appName)),
}
for _, sap := range saPorts {
var proto string
switch sap.Protocol {
case "tcp":
proto = "ListenStream"
case "udp":
proto = "ListenDatagram"
default:
return fmt.Errorf("unrecognized protocol: %v", sap.Protocol)
}
// We find the host port for the pod's port and use that in the
// socket unit file.
// This is so because systemd inside the pod will match based on
// the socket port number, and since the socket was created on the
// host, it will have the host port number.
port := findHostPort(*p.Manifest, sap.Name)
if port == 0 {
log.Printf("warning: no --port option for socket-activated port %q, assuming port %d as specified in the manifest", sap.Name, sap.Port)
port = sap.Port
}
sockopts = append(sockopts, unit.NewUnitOption("Socket", proto, fmt.Sprintf("%v", port)))
}
file, err := os.OpenFile(SocketUnitPath(p.Root, appName), os.O_WRONLY|os.O_CREATE, 0644)
if err != nil {
return errwrap.Wrap(errors.New("failed to create socket file"), err)
}
defer file.Close()
if _, err = io.Copy(file, unit.Serialize(sockopts)); err != nil {
return errwrap.Wrap(errors.New("failed to write socket unit file"), err)
}
if err = os.Symlink(path.Join("..", SocketUnitName(appName)), SocketWantPath(p.Root, appName)); err != nil {
return errwrap.Wrap(errors.New("failed to link socket want"), err)
}
opts = append(opts, unit.NewUnitOption("Unit", "Requires", SocketUnitName(appName)))
}
opts = append(opts, unit.NewUnitOption("Unit", "Requires", InstantiatedPrepareAppUnitName(appName)))
opts = append(opts, unit.NewUnitOption("Unit", "After", InstantiatedPrepareAppUnitName(appName)))
file, err := os.OpenFile(ServiceUnitPath(p.Root, appName), os.O_WRONLY|os.O_CREATE, 0644)
if err != nil {
return errwrap.Wrap(errors.New("failed to create service unit file"), err)
}
defer file.Close()
if _, err = io.Copy(file, unit.Serialize(opts)); err != nil {
return errwrap.Wrap(errors.New("failed to write service unit file"), err)
}
if err = os.Symlink(path.Join("..", ServiceUnitName(appName)), ServiceWantPath(p.Root, appName)); err != nil {
return errwrap.Wrap(errors.New("failed to link service want"), err)
}
if flavor == "kvm" {
// bind mount all shared volumes from /mnt/volumeName (we don't use mechanism for bind-mounting given by nspawn)
err := AppToSystemdMountUnits(common.Stage1RootfsPath(p.Root), appName, p.Manifest.Volumes, ra, UnitsDir)
if err != nil {
return errwrap.Wrap(errors.New("failed to prepare mount units"), err)
}
}
if err = writeAppReaper(p, appName.String()); err != nil {
return errwrap.Wrap(fmt.Errorf("failed to write app %q reaper service", appName), err)
}
return nil
}
// getArgsEnv returns the nspawn or lkvm args and env according to the flavor
// as the first two return values respectively.
func getArgsEnv(p *stage1commontypes.Pod, flavor string, canMachinedRegister bool, debug bool, n *networking.Networking) ([]string, []string, error) {
var args []string
env := os.Environ()
// We store the pod's flavor so we can later garbage collect it correctly
if err := os.Symlink(flavor, filepath.Join(p.Root, stage1initcommon.FlavorFile)); err != nil {
return nil, nil, errwrap.Wrap(errors.New("failed to create flavor symlink"), err)
}
// set hostname inside pod
// According to systemd manual (https://www.freedesktop.org/software/systemd/man/hostname.html) :
// "The /etc/hostname file configures the name of the local system that is set
// during boot using the sethostname system call"
if p.Hostname == "" {
p.Hostname = stage1initcommon.GetMachineID(p)
}
hostnamePath := filepath.Join(common.Stage1RootfsPath(p.Root), "etc/hostname")
if err := ioutil.WriteFile(hostnamePath, []byte(p.Hostname), 0644); err != nil {
return nil, nil, fmt.Errorf("error writing %s, %s", hostnamePath, err)
}
// systemd-nspawn needs /etc/machine-id to link the container's journal
// to the host. Since systemd-v230, /etc/machine-id is mandatory, see
// https://github.com/systemd/systemd/commit/e01ff70a77e781734e1e73a2238af2e9bf7967a8
mPath := filepath.Join(common.Stage1RootfsPath(p.Root), "etc", "machine-id")
machineID := strings.Replace(p.UUID.String(), "-", "", -1)
switch flavor {
case "kvm":
if p.PrivateUsers != "" {
return nil, nil, fmt.Errorf("flag --private-users cannot be used with an lkvm stage1")
}
// kernel and hypervisor binaries are located relative to the working directory
// of init (/var/lib/rkt/..../uuid)
// TODO: move to path.go
kernelPath := filepath.Join(common.Stage1RootfsPath(p.Root), "bzImage")
netDescriptions := kvm.GetNetworkDescriptions(n)
cpu, mem := kvm.GetAppsResources(p.Manifest.Apps)
// Parse hypervisor
hv, err := KvmCheckHypervisor(common.Stage1RootfsPath(p.Root))
if err != nil {
return nil, nil, err
}
// Set start command for hypervisor
StartCmd := hvlkvm.StartCmd
switch hv {
case "lkvm":
StartCmd = hvlkvm.StartCmd
case "qemu":
StartCmd = hvqemu.StartCmd
default:
return nil, nil, fmt.Errorf("unrecognized hypervisor")
}
hvStartCmd := StartCmd(
common.Stage1RootfsPath(p.Root),
p.UUID.String(),
kernelPath,
netDescriptions,
cpu,
mem,
debug,
)
if hvStartCmd == nil {
return nil, nil, fmt.Errorf("no hypervisor")
}
args = append(args, hvStartCmd...)
// lkvm requires $HOME to be defined,
// see https://github.com/coreos/rkt/issues/1393
if os.Getenv("HOME") == "" {
env = append(env, "HOME=/root")
}
if err := linkJournal(common.Stage1RootfsPath(p.Root), machineID); err != nil {
return nil, nil, errwrap.Wrap(errors.New("error linking pod's journal"), err)
}
// use only dynamic libraries provided in the image
// from systemd v231 there's a new internal libsystemd-shared-v231.so
// which is present in /usr/lib/systemd
env = append(env, "LD_LIBRARY_PATH="+filepath.Join(common.Stage1RootfsPath(p.Root), "usr/lib/systemd"))
return args, env, nil
case "coreos":
args = append(args, filepath.Join(common.Stage1RootfsPath(p.Root), interpBin))
args = append(args, filepath.Join(common.Stage1RootfsPath(p.Root), nspawnBin))
args = append(args, "--boot") // Launch systemd in the pod
args = append(args, "--notify-ready=yes") // From systemd v231
if context := os.Getenv(common.EnvSELinuxContext); context != "" {
args = append(args, fmt.Sprintf("-Z%s", context))
}
if context := os.Getenv(common.EnvSELinuxMountContext); context != "" {
args = append(args, fmt.Sprintf("-L%s", context))
}
if canMachinedRegister {
args = append(args, fmt.Sprintf("--register=true"))
} else {
args = append(args, fmt.Sprintf("--register=false"))
}
// use only dynamic libraries provided in the image
// from systemd v231 there's a new internal libsystemd-shared-v231.so
// which is present in /usr/lib/systemd
env = append(env, "LD_LIBRARY_PATH="+
filepath.Join(common.Stage1RootfsPath(p.Root), "usr/lib")+":"+
filepath.Join(common.Stage1RootfsPath(p.Root), "usr/lib/systemd"))
case "src":
args = append(args, filepath.Join(common.Stage1RootfsPath(p.Root), interpBin))
args = append(args, filepath.Join(common.Stage1RootfsPath(p.Root), nspawnBin))
args = append(args, "--boot") // Launch systemd in the pod
args = append(args, "--notify-ready=yes") // From systemd v231
if context := os.Getenv(common.EnvSELinuxContext); context != "" {
args = append(args, fmt.Sprintf("-Z%s", context))
}
if context := os.Getenv(common.EnvSELinuxMountContext); context != "" {
args = append(args, fmt.Sprintf("-L%s", context))
}
if canMachinedRegister {
args = append(args, fmt.Sprintf("--register=true"))
} else {
args = append(args, fmt.Sprintf("--register=false"))
}
// use only dynamic libraries provided in the image
// from systemd v231 there's a new internal libsystemd-shared-v231.so
// which is present in /usr/lib/systemd
env = append(env, "LD_LIBRARY_PATH="+
filepath.Join(common.Stage1RootfsPath(p.Root), "usr/lib")+":"+
filepath.Join(common.Stage1RootfsPath(p.Root), "usr/lib/systemd"))
case "host":
hostNspawnBin, err := common.LookupPath("systemd-nspawn", os.Getenv("PATH"))
if err != nil {
return nil, nil, err
}
// Check dynamically which version is installed on the host
// Support version >= 220
versionBytes, err := exec.Command(hostNspawnBin, "--version").CombinedOutput()
if err != nil {
return nil, nil, errwrap.Wrap(fmt.Errorf("unable to probe %s version", hostNspawnBin), err)
}
versionStr := strings.SplitN(string(versionBytes), "\n", 2)[0]
var version int
n, err := fmt.Sscanf(versionStr, "systemd %d", &version)
if err != nil {
return nil, nil, fmt.Errorf("cannot parse version: %q", versionStr)
}
if n != 1 || version < 220 {
return nil, nil, fmt.Errorf("rkt needs systemd-nspawn >= 220. %s version not supported: %v", hostNspawnBin, versionStr)
}
// Copy systemd, bash, etc. in stage1 at run-time
if err := installAssets(); err != nil {
return nil, nil, errwrap.Wrap(errors.New("cannot install assets from the host"), err)
}
args = append(args, hostNspawnBin)
args = append(args, "--boot") // Launch systemd in the pod
args = append(args, fmt.Sprintf("--register=true"))
if version >= 231 {
args = append(args, "--notify-ready=yes") // From systemd v231
}
if context := os.Getenv(common.EnvSELinuxContext); context != "" {
args = append(args, fmt.Sprintf("-Z%s", context))
}
if context := os.Getenv(common.EnvSELinuxMountContext); context != "" {
args = append(args, fmt.Sprintf("-L%s", context))
}
default:
return nil, nil, fmt.Errorf("unrecognized stage1 flavor: %q", flavor)
}
machineIDBytes := append([]byte(machineID), '\n')
if err := ioutil.WriteFile(mPath, machineIDBytes, 0644); err != nil {
log.FatalE("error writing /etc/machine-id", err)
}
// link journal only if the host is running systemd
if util.IsRunningSystemd() {
args = append(args, "--link-journal=try-guest")
keepUnit, err := util.RunningFromSystemService()
if err != nil {
if err == dlopen.ErrSoNotFound {
log.Print("warning: libsystemd not found even though systemd is running. Cgroup limits set by the environment (e.g. a systemd service) won't be enforced.")
} else {
return nil, nil, errwrap.Wrap(errors.New("error determining if we're running from a system service"), err)
}
}
if keepUnit {
args = append(args, "--keep-unit")
}
} else {
args = append(args, "--link-journal=no")
}
if !debug {
args = append(args, "--quiet") // silence most nspawn output (log_warning is currently not covered by this)
env = append(env, "SYSTEMD_LOG_LEVEL=err") // silence log_warning too
}
env = append(env, "SYSTEMD_NSPAWN_CONTAINER_SERVICE=rkt")
// TODO (alepuccetti) remove this line when rkt will use cgroup namespace
// If the kernel has the cgroup namespace enabled, systemd v232 will use it by default.
// This was introduced by https://github.com/systemd/systemd/pull/3809 and it will cause
// problems in rkt when cgns is enabled and cgroup-v1 is used. For more information see
// https://github.com/systemd/systemd/pull/3589#discussion_r70277625.
// The following line tells systemd-nspawn not to use cgroup namespace using the environment variable
// introduced by https://github.com/systemd/systemd/pull/3809.
env = append(env, "SYSTEMD_NSPAWN_USE_CGNS=no")
if p.InsecureOptions.DisablePaths {
env = append(env, "SYSTEMD_NSPAWN_API_VFS_WRITABLE=yes")
}
if len(p.PrivateUsers) > 0 {
args = append(args, "--private-users="+p.PrivateUsers)
}
nsargs, err := stage1initcommon.PodToNspawnArgs(p)
if err != nil {
return nil, nil, errwrap.Wrap(errors.New("failed to generate nspawn args"), err)
}
args = append(args, nsargs...)
// Arguments to systemd
args = append(args, "--")
args = append(args, "--default-standard-output=tty") // redirect all service logs straight to tty
if !debug {
args = append(args, "--log-target=null") // silence systemd output inside pod
args = append(args, "--show-status=0") // silence systemd initialization status output
}
return args, env, nil
}
// appToSystemd transforms the provided RuntimeApp+ImageManifest into systemd units
func appToSystemd(p *stage1commontypes.Pod, ra *schema.RuntimeApp, interactive bool, flavor string, privateUsers string) error {
app := ra.App
appName := ra.Name
imgName := p.AppNameToImageName(appName)
if len(app.Exec) == 0 {
return fmt.Errorf(`image %q has an empty "exec" (try --exec=BINARY)`, imgName)
}
workDir := "/"
if app.WorkingDirectory != "" {
workDir = app.WorkingDirectory
}
env := app.Environment
env.Set("AC_APP_NAME", appName.String())
if p.MetadataServiceURL != "" {
env.Set("AC_METADATA_URL", p.MetadataServiceURL)
}
envFilePath := EnvFilePath(p.Root, appName)
uidRange := user.NewBlankUidRange()
if err := uidRange.Deserialize([]byte(privateUsers)); err != nil {
return err
}
if err := writeEnvFile(p, env, appName, uidRange, '\n', envFilePath); err != nil {
return errwrap.Wrap(errors.New("unable to write environment file for systemd"), err)
}
u, g, err := parseUserGroup(p, ra, uidRange)
if err != nil {
return err
}
if err := generateSysusers(p, ra, u, g, uidRange); err != nil {
return errwrap.Wrap(errors.New("unable to generate sysusers"), err)
}
binPath, err := findBinPath(p, appName, *app, workDir, app.Exec[0])
if err != nil {
return err
}
var supplementaryGroups []string
for _, g := range app.SupplementaryGIDs {
supplementaryGroups = append(supplementaryGroups, strconv.Itoa(g))
}
capabilitiesStr, err := getAppCapabilities(app.Isolators)
if err != nil {
return err
}
noNewPrivileges := getAppNoNewPrivileges(app.Isolators)
execStart := append([]string{binPath}, app.Exec[1:]...)
execStartString := quoteExec(execStart)
opts := []*unit.UnitOption{
unit.NewUnitOption("Unit", "Description", fmt.Sprintf("Application=%v Image=%v", appName, imgName)),
unit.NewUnitOption("Unit", "DefaultDependencies", "false"),
unit.NewUnitOption("Unit", "Wants", fmt.Sprintf("reaper-%s.service", appName)),
unit.NewUnitOption("Service", "Restart", "no"),
unit.NewUnitOption("Service", "ExecStart", execStartString),
unit.NewUnitOption("Service", "RootDirectory", common.RelAppRootfsPath(appName)),
// MountFlags=shared creates a new mount namespace and (as unintuitive
// as it might seem) makes sure the mount is slave+shared.
unit.NewUnitOption("Service", "MountFlags", "shared"),
unit.NewUnitOption("Service", "WorkingDirectory", workDir),
unit.NewUnitOption("Service", "EnvironmentFile", RelEnvFilePath(appName)),
unit.NewUnitOption("Service", "User", strconv.Itoa(u)),
unit.NewUnitOption("Service", "Group", strconv.Itoa(g)),
unit.NewUnitOption("Service", "SupplementaryGroups", strings.Join(supplementaryGroups, " ")),
unit.NewUnitOption("Service", "CapabilityBoundingSet", strings.Join(capabilitiesStr, " ")),
unit.NewUnitOption("Service", "NoNewPrivileges", strconv.FormatBool(noNewPrivileges)),
// This helps working around a race
// (https://github.com/systemd/systemd/issues/2913) that causes the
// systemd unit name not getting written to the journal if the unit is
// short-lived and runs as non-root.
unit.NewUnitOption("Service", "SyslogIdentifier", appName.String()),
}
// Restrict access to sensitive paths (eg. procfs)
opts = protectSystemFiles(opts, appName)
if ra.ReadOnlyRootFS {
opts = append(opts, unit.NewUnitOption("Service", "ReadOnlyDirectories", common.RelAppRootfsPath(appName)))
}
// TODO(tmrts): Extract this logic into a utility function.
vols := make(map[types.ACName]types.Volume)
for _, v := range p.Manifest.Volumes {
vols[v.Name] = v
}
absRoot, err := filepath.Abs(p.Root) // Absolute path to the pod's rootfs.
if err != nil {
return err
}
appRootfs := common.AppRootfsPath(absRoot, appName)
rwDirs := []string{}
imageManifest := p.Images[appName.String()]
for _, m := range GenerateMounts(ra, vols, imageManifest) {
mntPath, err := EvaluateSymlinksInsideApp(appRootfs, m.Path)
if err != nil {
return err
}
if !IsMountReadOnly(vols[m.Volume], app.MountPoints) {
rwDirs = append(rwDirs, filepath.Join(common.RelAppRootfsPath(appName), mntPath))
}
}
opts = append(opts, unit.NewUnitOption("Service", "ReadWriteDirectories", strings.Join(rwDirs, " ")))
if interactive {
opts = append(opts, unit.NewUnitOption("Service", "StandardInput", "tty"))
opts = append(opts, unit.NewUnitOption("Service", "StandardOutput", "tty"))
opts = append(opts, unit.NewUnitOption("Service", "StandardError", "tty"))
} else {
opts = append(opts, unit.NewUnitOption("Service", "StandardOutput", "journal+console"))
opts = append(opts, unit.NewUnitOption("Service", "StandardError", "journal+console"))
}
// When an app fails, we shut down the pod
opts = append(opts, unit.NewUnitOption("Unit", "OnFailure", "halt.target"))
for _, eh := range app.EventHandlers {
var typ string
switch eh.Name {
case "pre-start":
typ = "ExecStartPre"
case "post-stop":
typ = "ExecStopPost"
default:
return fmt.Errorf("unrecognized eventHandler: %v", eh.Name)
}
exec := quoteExec(eh.Exec)
opts = append(opts, unit.NewUnitOption("Service", typ, exec))
}
// Some pre-start jobs take a long time, set the timeout to 0
opts = append(opts, unit.NewUnitOption("Service", "TimeoutStartSec", "0"))
var saPorts []types.Port
for _, p := range app.Ports {
if p.SocketActivated {
saPorts = append(saPorts, p)
}
}
for _, i := range app.Isolators {
switch v := i.Value().(type) {
case *types.ResourceMemory:
opts, err = cgroup.MaybeAddIsolator(opts, "memory", v.Limit())
if err != nil {
return err
}
case *types.ResourceCPU:
opts, err = cgroup.MaybeAddIsolator(opts, "cpu", v.Limit())
if err != nil {
return err
}
}
}
if len(saPorts) > 0 {
sockopts := []*unit.UnitOption{
unit.NewUnitOption("Unit", "Description", fmt.Sprintf("Application=%v Image=%v %s", appName, imgName, "socket-activated ports")),
unit.NewUnitOption("Unit", "DefaultDependencies", "false"),
unit.NewUnitOption("Socket", "BindIPv6Only", "both"),
unit.NewUnitOption("Socket", "Service", ServiceUnitName(appName)),
}
for _, sap := range saPorts {
var proto string
switch sap.Protocol {
case "tcp":
proto = "ListenStream"
case "udp":
proto = "ListenDatagram"
default:
return fmt.Errorf("unrecognized protocol: %v", sap.Protocol)
}
// We find the host port for the pod's port and use that in the
// socket unit file.
// This is so because systemd inside the pod will match based on
// the socket port number, and since the socket was created on the
// host, it will have the host port number.
port := findHostPort(*p.Manifest, sap.Name)
if port == 0 {
log.Printf("warning: no --port option for socket-activated port %q, assuming port %d as specified in the manifest", sap.Name, sap.Port)
port = sap.Port
}
sockopts = append(sockopts, unit.NewUnitOption("Socket", proto, fmt.Sprintf("%v", port)))
}
file, err := os.OpenFile(SocketUnitPath(p.Root, appName), os.O_WRONLY|os.O_CREATE, 0644)
if err != nil {
return errwrap.Wrap(errors.New("failed to create socket file"), err)
}
defer file.Close()
if _, err = io.Copy(file, unit.Serialize(sockopts)); err != nil {
return errwrap.Wrap(errors.New("failed to write socket unit file"), err)
}
if err = os.Symlink(path.Join("..", SocketUnitName(appName)), SocketWantPath(p.Root, appName)); err != nil {
return errwrap.Wrap(errors.New("failed to link socket want"), err)
}
opts = append(opts, unit.NewUnitOption("Unit", "Requires", SocketUnitName(appName)))
}
opts = append(opts, unit.NewUnitOption("Unit", "Requires", InstantiatedPrepareAppUnitName(appName)))
opts = append(opts, unit.NewUnitOption("Unit", "After", InstantiatedPrepareAppUnitName(appName)))
opts = append(opts, unit.NewUnitOption("Unit", "Requires", "sysusers.service"))
opts = append(opts, unit.NewUnitOption("Unit", "After", "sysusers.service"))
file, err := os.OpenFile(ServiceUnitPath(p.Root, appName), os.O_WRONLY|os.O_CREATE, 0644)
if err != nil {
return errwrap.Wrap(errors.New("failed to create service unit file"), err)
}
defer file.Close()
if _, err = io.Copy(file, unit.Serialize(opts)); err != nil {
return errwrap.Wrap(errors.New("failed to write service unit file"), err)
}
if err = os.Symlink(path.Join("..", ServiceUnitName(appName)), ServiceWantPath(p.Root, appName)); err != nil {
return errwrap.Wrap(errors.New("failed to link service want"), err)
}
if err = writeAppReaper(p, appName.String(), common.RelAppRootfsPath(appName), binPath); err != nil {
return errwrap.Wrap(fmt.Errorf("failed to write app %q reaper service", appName), err)
}
return nil
}
// appToSystemd transforms the provided RuntimeApp+ImageManifest into systemd units
func appToSystemd(p *stage1commontypes.Pod, ra *schema.RuntimeApp, interactive bool, flavor string, privateUsers string) error {
app := ra.App
appName := ra.Name
imgName := p.AppNameToImageName(appName)
if len(app.Exec) == 0 {
return fmt.Errorf(`image %q has an empty "exec" (try --exec=BINARY)`, imgName)
}
workDir := "/"
if app.WorkingDirectory != "" {
workDir = app.WorkingDirectory
}
env := app.Environment
env.Set("AC_APP_NAME", appName.String())
if p.MetadataServiceURL != "" {
env.Set("AC_METADATA_URL", p.MetadataServiceURL)
}
if err := writeEnvFile(p, env, appName, privateUsers); err != nil {
return errwrap.Wrap(errors.New("unable to write environment file"), err)
}
// This is a partial implementation for app.User and app.Group:
// For now, only numeric ids (and the string "root") are supported.
var uid, gid int
var err error
if app.User == "root" {
uid = 0
} else {
uid, err = strconv.Atoi(app.User)
if err != nil {
return fmt.Errorf("non-numerical user id not supported yet")
}
}
if app.Group == "root" {
gid = 0
} else {
gid, err = strconv.Atoi(app.Group)
if err != nil {
return fmt.Errorf("non-numerical group id not supported yet")
}
}
execWrap := []string{"/appexec", common.RelAppRootfsPath(appName), workDir, RelEnvFilePath(appName), strconv.Itoa(uid), generateGidArg(gid, app.SupplementaryGIDs)}
execStart := quoteExec(append(execWrap, app.Exec...))
opts := []*unit.UnitOption{
unit.NewUnitOption("Unit", "Description", fmt.Sprintf("Application=%v Image=%v", appName, imgName)),
unit.NewUnitOption("Unit", "DefaultDependencies", "false"),
unit.NewUnitOption("Unit", "Wants", fmt.Sprintf("reaper-%s.service", appName)),
unit.NewUnitOption("Service", "Restart", "no"),
unit.NewUnitOption("Service", "ExecStart", execStart),
unit.NewUnitOption("Service", "User", "0"),
unit.NewUnitOption("Service", "Group", "0"),
}
if interactive {
opts = append(opts, unit.NewUnitOption("Service", "StandardInput", "tty"))
opts = append(opts, unit.NewUnitOption("Service", "StandardOutput", "tty"))
opts = append(opts, unit.NewUnitOption("Service", "StandardError", "tty"))
} else {
opts = append(opts, unit.NewUnitOption("Service", "StandardOutput", "journal+console"))
opts = append(opts, unit.NewUnitOption("Service", "StandardError", "journal+console"))
opts = append(opts, unit.NewUnitOption("Service", "SyslogIdentifier", filepath.Base(app.Exec[0])))
}
// When an app fails, we shut down the pod
opts = append(opts, unit.NewUnitOption("Unit", "OnFailure", "halt.target"))
for _, eh := range app.EventHandlers {
var typ string
switch eh.Name {
case "pre-start":
typ = "ExecStartPre"
case "post-stop":
typ = "ExecStopPost"
default:
return fmt.Errorf("unrecognized eventHandler: %v", eh.Name)
}
exec := quoteExec(append(execWrap, eh.Exec...))
opts = append(opts, unit.NewUnitOption("Service", typ, exec))
}
// Some pre-start jobs take a long time, set the timeout to 0
opts = append(opts, unit.NewUnitOption("Service", "TimeoutStartSec", "0"))
var saPorts []types.Port
for _, p := range app.Ports {
if p.SocketActivated {
saPorts = append(saPorts, p)
}
}
for _, i := range app.Isolators {
switch v := i.Value().(type) {
case *types.ResourceMemory:
opts, err = cgroup.MaybeAddIsolator(opts, "memory", v.Limit())
if err != nil {
return err
}
case *types.ResourceCPU:
opts, err = cgroup.MaybeAddIsolator(opts, "cpu", v.Limit())
if err != nil {
return err
}
}
}
if len(saPorts) > 0 {
sockopts := []*unit.UnitOption{
unit.NewUnitOption("Unit", "Description", fmt.Sprintf("Application=%v Image=%v %s", appName, imgName, "socket-activated ports")),
unit.NewUnitOption("Unit", "DefaultDependencies", "false"),
unit.NewUnitOption("Socket", "BindIPv6Only", "both"),
unit.NewUnitOption("Socket", "Service", ServiceUnitName(appName)),
}
for _, sap := range saPorts {
var proto string
switch sap.Protocol {
case "tcp":
proto = "ListenStream"
case "udp":
proto = "ListenDatagram"
default:
return fmt.Errorf("unrecognized protocol: %v", sap.Protocol)
}
sockopts = append(sockopts, unit.NewUnitOption("Socket", proto, fmt.Sprintf("%v", sap.Port)))
}
file, err := os.OpenFile(SocketUnitPath(p.Root, appName), os.O_WRONLY|os.O_CREATE, 0644)
if err != nil {
return errwrap.Wrap(errors.New("failed to create socket file"), err)
}
defer file.Close()
if _, err = io.Copy(file, unit.Serialize(sockopts)); err != nil {
return errwrap.Wrap(errors.New("failed to write socket unit file"), err)
}
if err = os.Symlink(path.Join("..", SocketUnitName(appName)), SocketWantPath(p.Root, appName)); err != nil {
return errwrap.Wrap(errors.New("failed to link socket want"), err)
}
opts = append(opts, unit.NewUnitOption("Unit", "Requires", SocketUnitName(appName)))
}
opts = append(opts, unit.NewUnitOption("Unit", "Requires", InstantiatedPrepareAppUnitName(appName)))
opts = append(opts, unit.NewUnitOption("Unit", "After", InstantiatedPrepareAppUnitName(appName)))
file, err := os.OpenFile(ServiceUnitPath(p.Root, appName), os.O_WRONLY|os.O_CREATE, 0644)
if err != nil {
return errwrap.Wrap(errors.New("failed to create service unit file"), err)
}
defer file.Close()
if _, err = io.Copy(file, unit.Serialize(opts)); err != nil {
return errwrap.Wrap(errors.New("failed to write service unit file"), err)
}
if err = os.Symlink(path.Join("..", ServiceUnitName(appName)), ServiceWantPath(p.Root, appName)); err != nil {
return errwrap.Wrap(errors.New("failed to link service want"), err)
}
if flavor == "kvm" {
// bind mount all shared volumes from /mnt/volumeName (we don't use mechanism for bind-mounting given by nspawn)
err := AppToSystemdMountUnits(common.Stage1RootfsPath(p.Root), appName, p.Manifest.Volumes, ra, UnitsDir)
if err != nil {
return errwrap.Wrap(errors.New("failed to prepare mount units"), err)
}
}
if err = writeAppReaper(p, appName.String()); err != nil {
return errwrap.Wrap(fmt.Errorf("failed to write app %q reaper service", appName), err)
}
return nil
}
