From 4fae756645cf11a9f8bae87dd845d82dce5e428e Mon Sep 17 00:00:00 2001
From: Nicolas Lacasse <nlacasse@google.com>
Date: Wed, 17 Oct 2018 09:01:42 -0700
Subject: Make removing cgroups retry up to 5 seconds.

Sometimes if we try to remove the cgroup directory too soon after killing the
sandbox we EBUSY. This CL adds a retry (up to 5 seconds) for removing.

Deflakes ChrootTest.

PiperOrigin-RevId: 217526909
Change-Id: I749bb172117e2298c9888ecad094072393b94810
---
 runsc/cgroup/BUILD     |  1 +
 runsc/cgroup/cgroup.go | 18 ++++++++++++++++--
 2 files changed, 17 insertions(+), 2 deletions(-)

(limited to 'runsc/cgroup')

diff --git a/runsc/cgroup/BUILD b/runsc/cgroup/BUILD
index 4a535d230..10a8e5feb 100644
--- a/runsc/cgroup/BUILD
+++ b/runsc/cgroup/BUILD
@@ -12,6 +12,7 @@ go_library(
     deps = [
         "//pkg/log",
         "//runsc/specutils",
+        "@com_github_cenkalti_backoff//:go_default_library",
         "@com_github_opencontainers_runtime-spec//specs-go:go_default_library",
     ],
 )
diff --git a/runsc/cgroup/cgroup.go b/runsc/cgroup/cgroup.go
index af0252bb3..7a75a189a 100644
--- a/runsc/cgroup/cgroup.go
+++ b/runsc/cgroup/cgroup.go
@@ -17,6 +17,7 @@
 package cgroup
 
 import (
+	"context"
 	"fmt"
 	"io/ioutil"
 	"os"
@@ -24,7 +25,9 @@ import (
 	"strconv"
 	"strings"
 	"syscall"
+	"time"
 
+	"github.com/cenkalti/backoff"
 	specs "github.com/opencontainers/runtime-spec/specs-go"
 	"gvisor.googlesource.com/gvisor/pkg/log"
 	"gvisor.googlesource.com/gvisor/runsc/specutils"
@@ -214,8 +217,19 @@ func (c *Cgroup) Uninstall() error {
 	}
 	log.Debugf("Deleting cgroup %q", c.Name)
 	for key := range controllers {
-		if err := syscall.Rmdir(c.makePath(key)); err != nil && !os.IsNotExist(err) {
-			return err
+		path := c.makePath(key)
+		log.Debugf("Removing cgroup controller for key=%q path=%q", key, path)
+
+		// If we try to remove the cgroup too soon after killing the
+		// sandbox we might get EBUSY, so we retry for a few seconds
+		// until it succeeds.
+		ctx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
+		defer cancel()
+		b := backoff.WithContext(backoff.NewConstantBackOff(100*time.Millisecond), ctx)
+		if err := backoff.Retry(func() error {
+			return syscall.Rmdir(path)
+		}, b); err != nil {
+			return fmt.Errorf("error removing cgroup path %q: %v", path, err)
 		}
 	}
 	return nil
-- 
cgit v1.2.3