fix(worker): cgroup hook now ensure child progresses are killed

worker/cgroup.go

@@ -2,18 +2,20 @@ package worker
 
 import (
 	"bufio"
+	"errors"
 	"fmt"
 	"os"
 	"path/filepath"
 	"strconv"
 	"syscall"
+	"time"
 
 	"golang.org/x/sys/unix"
 
 	"github.com/codeskyblue/go-sh"
 )
 
-var cgSubsystem string = "cpu"
+var cgSubsystem = "cpu"
 
 type cgroupHook struct {
 	emptyHook
@@ -82,24 +84,44 @@ func (c *cgroupHook) killAll() error {
 		return nil
 	}
 	name := c.provider.Name()
+
+	readTaskList := func() ([]int, error) {
+		taskList := []int{}
 		taskFile, err := os.Open(filepath.Join(c.basePath, cgSubsystem, c.baseGroup, name, "tasks"))
 		if err != nil {
-		return err
+			return taskList, err
 		}
 		defer taskFile.Close()
-	taskList := []int{}
+
 		scanner := bufio.NewScanner(taskFile)
 		for scanner.Scan() {
 			pid, err := strconv.Atoi(scanner.Text())
 			if err != nil {
-			return err
+				return taskList, err
 			}
 			taskList = append(taskList, pid)
 		}
+		return taskList, nil
+	}
+
+	for i := 0; i < 4; i++ {
+		if i == 3 {
+			return errors.New("Unable to kill all child tasks")
+		}
+		taskList, err := readTaskList()
+		if err != nil {
+			return err
+		}
+		if len(taskList) == 0 {
+			return nil
+		}
 		for _, pid := range taskList {
 			logger.Debugf("Killing process: %d", pid)
 			unix.Kill(pid, syscall.SIGKILL)
 		}
+		// sleep 10ms for the first round, and 1.01s, 2.01s, 3.01s for the rest
+		time.Sleep(time.Duration(i)*time.Second + 10*time.Millisecond)
+	}
 
 	return nil
 }