Refactor process.go for platform specific

Signed-off-by: John Howard <jhoward@microsoft.com>

Move process sorter to new file

Signed-off-by: Michael Crosby <crosbymichael@gmail.com>

Sort containers by id

This will not be the most accurate sorting but atleast the list will be
consistent inbetween calls.

Signed-off-by: Michael Crosby <crosbymichael@gmail.com>

Allow runtime to be configurable via daemon start

This allows people to pass an alternate name or location to the runtime
binary to start containers.

Signed-off-by: Michael Crosby <crosbymichael@gmail.com>

Fix state output for containers

Return the proper state/status for a container by checking if the pid is
still alive.  Also fix the cleanup handling in the shim to make sure
containers are not left behind.

Signed-off-by: Michael Crosby <crosbymichael@gmail.com>

Properly wait for container start

Signed-off-by: Michael Crosby <crosbymichael@gmail.com>

supervisor/create.go

@@ -20,7 +20,7 @@ type StartTask struct {
 
 func (s *Supervisor) start(t *StartTask) error {
 	start := time.Now()
-	container, err := runtime.New(s.stateDir, t.ID, t.BundlePath, t.Labels)
+	container, err := runtime.New(s.stateDir, t.ID, t.BundlePath, s.runtime, t.Labels)
 	if err != nil {
 		return err
 	}

supervisor/exit.go

@@ -17,9 +17,19 @@ func (s *Supervisor) exit(t *ExitTask) error {
 	proc := t.Process
 	status, err := proc.ExitStatus()
 	if err != nil {
-		logrus.WithField("error", err).Error("containerd: get exit status")
+		logrus.WithFields(logrus.Fields{
+			"error":     err,
+			"pid":       proc.ID(),
+			"id":        proc.Container().ID(),
+			"systemPid": proc.SystemPid(),
+		}).Error("containerd: get exit status")
 	}
-	logrus.WithFields(logrus.Fields{"pid": proc.ID(), "status": status}).Debug("containerd: process exited")
+	logrus.WithFields(logrus.Fields{
+		"pid":       proc.ID(),
+		"status":    status,
+		"id":        proc.Container().ID(),
+		"systemPid": proc.SystemPid(),
+	}).Debug("containerd: process exited")
 
 	// if the process is the the init process of the container then
 	// fire a separate event for this process

supervisor/sort.go

@@ -0,0 +1,27 @@
+package supervisor
+
+import (
+	"sort"
+
+	"github.com/docker/containerd/runtime"
+)
+
+func sortProcesses(p []runtime.Process) {
+	sort.Sort(&processSorter{p})
+}
+
+type processSorter struct {
+	processes []runtime.Process
+}
+
+func (s *processSorter) Len() int {
+	return len(s.processes)
+}
+
+func (s *processSorter) Swap(i, j int) {
+	s.processes[i], s.processes[j] = s.processes[j], s.processes[i]
+}
+
+func (s *processSorter) Less(i, j int) bool {
+	return s.processes[j].ID() == "init"
+}

supervisor/supervisor.go

@@ -6,7 +6,6 @@ import (
 	"io/ioutil"
 	"os"
 	"path/filepath"
-	"sort"
 	"sync"
 	"time"
 
@@ -20,7 +19,7 @@ const (
 )
 
 // New returns an initialized Process supervisor.
-func New(stateDir string, oom bool) (*Supervisor, error) {
+func New(stateDir string, oom bool, runtimeName string) (*Supervisor, error) {
 	startTasks := make(chan *startTask, 10)
 	if err := os.MkdirAll(stateDir, 0755); err != nil {
 		return nil, err
@@ -41,6 +40,7 @@ func New(stateDir string, oom bool) (*Supervisor, error) {
 		subscribers: make(map[chan Event]struct{}),
 		tasks:       make(chan Task, defaultBufferSize),
 		monitor:     monitor,
+		runtime:     runtimeName,
 	}
 	if err := setupEventLog(s); err != nil {
 		return nil, err
@@ -116,7 +116,9 @@ func readEventLog(s *Supervisor) error {
 
 type Supervisor struct {
 	// stateDir is the directory on the system to store container runtime state information.
-	stateDir   string
+	stateDir string
+	// name of the OCI compatible runtime used to execute containers
+	runtime    string
 	containers map[string]*containerInfo
 	startTasks chan *startTask
 	// we need a lock around the subscribers map only because additions and deletions from
@@ -207,7 +209,12 @@ func (s *Supervisor) notifySubscribers(e Event) {
 // therefore it is save to do operations in the handlers that modify state of the system or
 // state of the Supervisor
 func (s *Supervisor) Start() error {
-	logrus.WithField("stateDir", s.stateDir).Debug("containerd: supervisor running")
+	logrus.WithFields(logrus.Fields{
+		"stateDir": s.stateDir,
+		"runtime":  s.runtime,
+		"memory":   s.machine.Memory,
+		"cpus":     s.machine.Cpus,
+	}).Debug("containerd: supervisor running")
 	go func() {
 		for i := range s.tasks {
 			s.handleTask(i)
@@ -277,7 +284,7 @@ func (s *Supervisor) restore() error {
 		if len(exitedProcesses) > 0 {
 			// sort processes so that init is fired last because that is how the kernel sends the
 			// exit events
-			sort.Sort(&processSorter{exitedProcesses})
+			sortProcesses(exitedProcesses)
 			for _, p := range exitedProcesses {
 				e := &ExitTask{
 					Process: p,
@@ -288,19 +295,3 @@ func (s *Supervisor) restore() error {
 	}
 	return nil
 }
-
-type processSorter struct {
-	processes []runtime.Process
-}
-
-func (s *processSorter) Len() int {
-	return len(s.processes)
-}
-
-func (s *processSorter) Swap(i, j int) {
-	s.processes[i], s.processes[j] = s.processes[j], s.processes[i]
-}
-
-func (s *processSorter) Less(i, j int) bool {
-	return s.processes[j].ID() == "init"
-}

supervisor/worker.go

@@ -40,11 +40,15 @@ func (w *worker) Start() {
 		started := time.Now()
 		process, err := t.Container.Start(t.Checkpoint, runtime.NewStdio(t.Stdin, t.Stdout, t.Stderr))
 		if err != nil {
+			logrus.WithFields(logrus.Fields{
+				"error": err,
+				"id":    t.Container.ID(),
+			}).Error("containerd: start container")
+			t.Err <- err
 			evt := &DeleteTask{
 				ID: t.Container.ID(),
 			}
 			w.s.SendTask(evt)
-			t.Err <- err
 			continue
 		}
 		/*