lock: Fix decision when a lock needs to be removed (minio#14095)

The code was not properly deciding if a lock needs to be removed 
when it doesn't have quorum anymore. After this commit, a lock will be
forcefully unlocked if nodes reporting they are not able to find a lock
internally breaks the quorum.

Simplify the code as well.

internal/dsync/drwmutex.go

@@ -234,8 +234,8 @@ func (dm *DRWMutex) startContinousLockRefresh(lockLossCallback func(), id, sourc
 			case <-refreshTimer.C:
 				refreshTimer.Reset(drwMutexRefreshInterval)
 
-				refreshed, err := refresh(ctx, dm.clnt, id, source, quorum)
-				if err == nil && !refreshed {
+				noQuorum, err := refreshLock(ctx, dm.clnt, id, source, quorum)
+				if err == nil && noQuorum {
 					// Clean the lock locally and in remote nodes
 					forceUnlock(ctx, dm.clnt, id)
 					// Execute the caller lock loss callback
@@ -273,10 +273,12 @@ func forceUnlock(ctx context.Context, ds *Dsync, id string) {
 
 type refreshResult struct {
 	offline   bool
-	succeeded bool
+	refreshed bool
 }
 
-func refresh(ctx context.Context, ds *Dsync, id, source string, quorum int) (bool, error) {
+// Refresh the given lock in all nodes, return true to indicate if a lock
+// does not exist in enough quorum nodes.
+func refreshLock(ctx context.Context, ds *Dsync, id, source string, quorum int) (bool, error) {
 	restClnts, _ := ds.GetLockers()
 
 	// Create buffered channel of size equal to total number of nodes.
@@ -302,16 +304,12 @@ func refresh(ctx context.Context, ds *Dsync, id, source string, quorum int) (boo
 			defer cancel()
 
 			refreshed, err := c.Refresh(ctx, args)
-			if refreshed && err == nil {
-				ch <- refreshResult{succeeded: true}
+			if err != nil {
+				ch <- refreshResult{offline: true}
+				log("dsync: Unable to call Refresh failed with %s for %#v at %s\n", err, args, c)
 			} else {
-				if err != nil {
-					ch <- refreshResult{offline: true}
-					log("dsync: Unable to call Refresh failed with %s for %#v at %s\n", err, args, c)
-				} else {
-					ch <- refreshResult{succeeded: false}
-					log("dsync: Refresh returned false for %#v at %s\n", args, c)
-				}
+				ch <- refreshResult{refreshed: refreshed}
+				log("dsync: Refresh returned false for %#v at %s\n", args, c)
 			}
 		}(index, c)
 	}
@@ -322,39 +320,32 @@ func refresh(ctx context.Context, ds *Dsync, id, source string, quorum int) (boo
 	// b) received too many refreshed for quorum to be still possible
 	// c) timed out
 	//
-	i, refreshFailed, refreshSucceeded := 0, 0, 0
+	lockNotFound, lockRefreshed := 0, 0
 	done := false
 
-	for ; i < len(restClnts); i++ {
+	for i := 0; i < len(restClnts); i++ {
 		select {
-		case refresh := <-ch:
-			if refresh.offline {
+		case refreshResult := <-ch:
+			if refreshResult.offline {
 				continue
 			}
-			if refresh.succeeded {
-				refreshSucceeded++
+			if refreshResult.refreshed {
+				lockRefreshed++
 			} else {
-				refreshFailed++
+				lockNotFound++
 			}
-			if refreshFailed > quorum {
-				// We know that we are not going to succeed with refresh
+			if lockRefreshed >= quorum || lockNotFound > len(restClnts)-quorum {
 				done = true
 			}
 		case <-ctx.Done():
 			// Refreshing is canceled
 			return false, ctx.Err()
 		}
-
 		if done {
 			break
 		}
 	}
 
-	refreshQuorum := refreshSucceeded >= quorum
-	if !refreshQuorum {
-		refreshQuorum = refreshFailed < quorum
-	}
-
 	// We may have some unused results in ch, release them async.
 	go func() {
 		wg.Wait()
@@ -363,7 +354,8 @@ func refresh(ctx context.Context, ds *Dsync, id, source string, quorum int) (boo
 		}
 	}()
 
-	return refreshQuorum, nil
+	noQuorum := lockNotFound > len(restClnts)-quorum
+	return noQuorum, nil
 }
 
 // lock tries to acquire the distributed lock, returning true or false.

internal/dsync/dsync_test.go

@@ -236,10 +236,46 @@ func TestTwoSimultaneousLocksForDifferentResources(t *testing.T) {
 	time.Sleep(10 * time.Millisecond)
 }
 
-// Test refreshing lock
+// Test refreshing lock - refresh should always return true
+//
+func TestSuccessfulLockRefresh(t *testing.T) {
+	if testing.Short() {
+		t.Skip("skipping test in short mode.")
+	}
+
+	dm := NewDRWMutex(ds, "aap")
+	contextCanceled := make(chan struct{})
+
+	ctx, cl := context.WithCancel(context.Background())
+	cancel := func() {
+		cl()
+		close(contextCanceled)
+	}
+
+	if !dm.GetLock(ctx, cancel, id, source, Options{Timeout: 5 * time.Minute}) {
+		t.Fatal("GetLock() should be successful")
+	}
+
+	timer := time.NewTimer(drwMutexRefreshInterval * 2)
+
+	select {
+	case <-contextCanceled:
+		t.Fatal("Lock context canceled which is not expected")
+	case <-timer.C:
+	}
+
+	// Should be safe operation in all cases
+	dm.Unlock()
+}
+
+// Test canceling context while quorum servers report lock not found
 func TestFailedRefreshLock(t *testing.T) {
+	if testing.Short() {
+		t.Skip("skipping test in short mode.")
+	}
+
 	// Simulate Refresh RPC response to return no locking found
-	for i := range lockServers {
+	for i := range lockServers[:3] {
 		lockServers[i].setRefreshReply(false)
 		defer lockServers[i].setRefreshReply(true)
 	}
@@ -270,6 +306,10 @@ func TestFailedRefreshLock(t *testing.T) {
 
 // Test Unlock should not timeout
 func TestUnlockShouldNotTimeout(t *testing.T) {
+	if testing.Short() {
+		t.Skip("skipping test in short mode.")
+	}
+
 	dm := NewDRWMutex(ds, "aap")
 
 	if !dm.GetLock(context.Background(), nil, id, source, Options{Timeout: 5 * time.Minute}) {