func stage1() int { uuid, err := types.NewUUID(flag.Arg(0)) if err != nil { log.PrintE("UUID is missing or malformed", err) return 1 } root := "." p, err := stage1commontypes.LoadPod(root, uuid) if err != nil { log.PrintE("failed to load pod", err) return 1 } // set close-on-exec flag on RKT_LOCK_FD so it gets correctly closed when invoking // network plugins lfd, err := common.GetRktLockFD() if err != nil { log.PrintE("failed to get rkt lock fd", err) return 1 } if err := sys.CloseOnExec(lfd, true); err != nil { log.PrintE("failed to set FD_CLOEXEC on rkt lock", err) return 1 } mirrorLocalZoneInfo(p.Root) flavor, _, err := stage1initcommon.GetFlavor(p) if err != nil { log.PrintE("failed to get stage1 flavor", err) return 3 } var n *networking.Networking if netList.Contained() { fps, err := forwardedPorts(p) if err != nil { log.Error(err) return 6 } n, err = networking.Setup(root, p.UUID, fps, netList, localConfig, flavor, debug) if err != nil { log.PrintE("failed to setup network", err) return 6 } if err = n.Save(); err != nil { log.PrintE("failed to save networking state", err) n.Teardown(flavor, debug) return 6 } if len(mdsToken) > 0 { hostIP, err := n.GetDefaultHostIP() if err != nil { log.PrintE("failed to get default Host IP", err) return 6 } p.MetadataServiceURL = common.MetadataServicePublicURL(hostIP, mdsToken) } } else { if flavor == "kvm" { log.Print("flavor kvm requires private network configuration (try --net)") return 6 } if len(mdsToken) > 0 { p.MetadataServiceURL = common.MetadataServicePublicURL(localhostIP, mdsToken) } } if err = stage1initcommon.WriteDefaultTarget(p); err != nil { log.PrintE("failed to write default.target", err) return 2 } if err = stage1initcommon.WritePrepareAppTemplate(p); err != nil { log.PrintE("failed to write prepare-app service template", err) return 2 } if err := stage1initcommon.SetJournalPermissions(p); err != nil { log.PrintE("warning: error setting journal ACLs, you'll need root to read the pod journal", err) } if flavor == "kvm" { if err := KvmPodToSystemd(p, n); err != nil { log.PrintE("failed to configure systemd for kvm", err) return 2 } } if err = stage1initcommon.PodToSystemd(p, interactive, flavor, privateUsers); err != nil { log.PrintE("failed to configure systemd", err) return 2 } args, env, err := getArgsEnv(p, flavor, debug, n) if err != nil { log.Error(err) return 3 } // create a separate mount namespace so the cgroup filesystems // are unmounted when exiting the pod if err := syscall.Unshare(syscall.CLONE_NEWNS); err != nil { log.FatalE("error unsharing", err) } // we recursively make / a "shared and slave" so mount events from the // new namespace don't propagate to the host namespace but mount events // from the host propagate to the new namespace and are forwarded to // its peer group // See https://www.kernel.org/doc/Documentation/filesystems/sharedsubtree.txt if err := syscall.Mount("", "/", "none", syscall.MS_REC|syscall.MS_SLAVE, ""); err != nil { log.FatalE("error making / a slave mount", err) } if err := syscall.Mount("", "/", "none", syscall.MS_REC|syscall.MS_SHARED, ""); err != nil { log.FatalE("error making / a shared and slave mount", err) } enabledCgroups, err := cgroup.GetEnabledCgroups() if err != nil { log.FatalE("error getting cgroups", err) return 5 } // mount host cgroups in the rkt mount namespace if err := mountHostCgroups(enabledCgroups); err != nil { log.FatalE("couldn't mount the host cgroups", err) return 5 } var serviceNames []string for _, app := range p.Manifest.Apps { serviceNames = append(serviceNames, stage1initcommon.ServiceUnitName(app.Name)) } s1Root := common.Stage1RootfsPath(p.Root) machineID := stage1initcommon.GetMachineID(p) subcgroup, err := getContainerSubCgroup(machineID) if err == nil { if err := mountContainerCgroups(s1Root, enabledCgroups, subcgroup, serviceNames); err != nil { log.PrintE("couldn't mount the container cgroups", err) return 5 } } else { log.PrintE("continuing with per-app isolators disabled", err) } if err = stage1common.WritePpid(os.Getpid()); err != nil { log.Error(err) return 4 } err = stage1common.WithClearedCloExec(lfd, func() error { return syscall.Exec(args[0], args, env) }) if err != nil { log.PrintE(fmt.Sprintf("failed to execute %q", args[0]), err) return 7 } return 0 }
func stage1() int { uuid, err := types.NewUUID(flag.Arg(0)) if err != nil { fmt.Fprintln(os.Stderr, "UUID is missing or malformed") return 1 } root := "." p, err := LoadPod(root, uuid) if err != nil { fmt.Fprintf(os.Stderr, "Failed to load pod: %v\n", err) return 1 } // set close-on-exec flag on RKT_LOCK_FD so it gets correctly closed when invoking // network plugins lfd, err := common.GetRktLockFD() if err != nil { fmt.Fprintf(os.Stderr, "Failed to get rkt lock fd: %v\n", err) return 1 } if err := sys.CloseOnExec(lfd, true); err != nil { fmt.Fprintf(os.Stderr, "Failed to set FD_CLOEXEC on rkt lock: %v\n", err) return 1 } mirrorLocalZoneInfo(p.Root) flavor, _, err := p.getFlavor() if err != nil { fmt.Fprintf(os.Stderr, "Failed to get stage1 flavor: %v\n", err) return 3 } var n *networking.Networking if privNet.Any() { fps, err := forwardedPorts(p) if err != nil { fmt.Fprintln(os.Stderr, err.Error()) return 6 } n, err = networking.Setup(root, p.UUID, fps, privNet, localConfig, flavor) if err != nil { fmt.Fprintf(os.Stderr, "Failed to setup network: %v\n", err) return 6 } if err = n.Save(); err != nil { fmt.Fprintf(os.Stderr, "Failed to save networking state %v\n", err) n.Teardown(flavor) return 6 } if len(mdsToken) > 0 { hostIP, err := n.GetDefaultHostIP() if err != nil { fmt.Fprintf(os.Stderr, "Failed to get default Host IP: %v\n", err) return 6 } p.MetadataServiceURL = common.MetadataServicePublicURL(hostIP, mdsToken) } } else { if flavor == "kvm" { fmt.Fprintf(os.Stderr, "Flavor kvm requires private network configuration (try --private-net).\n") return 6 } if len(mdsToken) > 0 { p.MetadataServiceURL = common.MetadataServicePublicURL(localhostIP, mdsToken) } } if err = p.WritePrepareAppTemplate(); err != nil { fmt.Fprintf(os.Stderr, "Failed to write prepare-app service template: %v\n", err) return 2 } if err = p.PodToSystemd(interactive, flavor); err != nil { fmt.Fprintf(os.Stderr, "Failed to configure systemd: %v\n", err) return 2 } args, env, err := getArgsEnv(p, flavor, debug, n) if err != nil { fmt.Fprintf(os.Stderr, "%v\n", err) return 3 } // create a separate mount namespace so the cgroup filesystems // are unmounted when exiting the pod if err := syscall.Unshare(syscall.CLONE_NEWNS); err != nil { log.Fatalf("error unsharing: %v", err) } // we recursively make / a "shared and slave" so mount events from the // new namespace don't propagate to the host namespace but mount events // from the host propagate to the new namespace and are forwarded to // its peer group // See https://www.kernel.org/doc/Documentation/filesystems/sharedsubtree.txt if err := syscall.Mount("", "/", "none", syscall.MS_REC|syscall.MS_SLAVE, ""); err != nil { log.Fatalf("error making / a slave mount: %v", err) } if err := syscall.Mount("", "/", "none", syscall.MS_REC|syscall.MS_SHARED, ""); err != nil { log.Fatalf("error making / a shared and slave mount: %v", err) } var serviceNames []string for _, app := range p.Manifest.Apps { serviceNames = append(serviceNames, ServiceUnitName(app.Name)) } s1Root := common.Stage1RootfsPath(p.Root) machineID := p.GetMachineID() subcgroup, err := getContainerSubCgroup(machineID) if err == nil { if err := cgroup.CreateCgroups(s1Root, subcgroup, serviceNames); err != nil { fmt.Fprintf(os.Stderr, "Error creating cgroups: %v\n", err) return 5 } } else { fmt.Fprintf(os.Stderr, "Continuing with per-app isolators disabled: %v\n", err) } if err = writePpid(os.Getpid()); err != nil { fmt.Fprintln(os.Stderr, err.Error()) return 4 } err = withClearedCloExec(lfd, func() error { return syscall.Exec(args[0], args, env) }) if err != nil { fmt.Fprintf(os.Stderr, "Failed to execute %q: %v\n", args[0], err) return 7 } return 0 }