[release-15.0] VReplication Last Error: retry error if it happens after timeout (#12114)

go/vt/vttablet/tabletmanager/vreplication/controller.go

@@ -72,13 +72,12 @@ func newController(ctx context.Context, params map[string]string, dbClientFactor
 	}
 
 	ct := &controller{
-		vre:               vre,
-		dbClientFactory:   dbClientFactory,
-		mysqld:            mysqld,
-		blpStats:          blpStats,
-		done:              make(chan struct{}),
-		source:            &binlogdatapb.BinlogSource{},
-		lastWorkflowError: newLastError("VReplication Controller", maxTimeToRetryError),
+		vre:             vre,
+		dbClientFactory: dbClientFactory,
+		mysqld:          mysqld,
+		blpStats:        blpStats,
+		done:            make(chan struct{}),
+		source:          &binlogdatapb.BinlogSource{},
 	}
 	log.Infof("creating controller with cell: %v, tabletTypes: %v, and params: %v", cell, tabletTypesStr, params)
 
@@ -89,6 +88,7 @@ func newController(ctx context.Context, params map[string]string, dbClientFactor
 	}
 	ct.id = uint32(id)
 	ct.workflow = params["workflow"]
+	ct.lastWorkflowError = newLastError(fmt.Sprintf("VReplication controller %d for workflow %q", ct.id, ct.workflow), maxTimeToRetryError)
 
 	state := params["state"]
 	blpStats.State.Set(state)

go/vt/vttablet/tabletmanager/vreplication/last_error.go

@@ -26,18 +26,20 @@ import (
 
 /*
  * lastError tracks the most recent error for any ongoing process and how long it has persisted.
- * The err field should be a vterror so as to ensure we have meaningful error codes, causes, stack
+ * The err field should be a vterror to ensure we have meaningful error codes, causes, stack
  * traces, etc.
  */
 type lastError struct {
 	name           string
 	err            error
 	firstSeen      time.Time
+	lastSeen       time.Time
 	mu             sync.Mutex
 	maxTimeInError time.Duration // if error persists for this long, shouldRetry() will return false
 }
 
 func newLastError(name string, maxTimeInError time.Duration) *lastError {
+	log.Infof("Created last error: %s, with maxTimeInError: %s", name, maxTimeInError)
 	return &lastError{
 		name:           name,
 		maxTimeInError: maxTimeInError,
@@ -48,15 +50,27 @@ func (le *lastError) record(err error) {
 	le.mu.Lock()
 	defer le.mu.Unlock()
 	if err == nil {
+		log.Infof("Resetting last error: %s", le.name)
 		le.err = nil
 		le.firstSeen = time.Time{}
+		le.lastSeen = time.Time{}
 		return
 	}
 	if !vterrors.Equals(err, le.err) {
+		log.Infof("Got new last error %+v for %s, was %+v", err, le.name, le.err)
 		le.firstSeen = time.Now()
+		le.lastSeen = time.Now()
 		le.err = err
+	} else {
+		// same error seen
+		log.Infof("Got the same last error for %q: %+v ; first seen at %s and last seen %dms ago", le.name, le.err, le.firstSeen, int(time.Since(le.lastSeen).Milliseconds()))
+		if time.Since(le.lastSeen) > le.maxTimeInError {
+			// reset firstSeen, since it has been long enough since the last time we saw this error
+			log.Infof("Resetting firstSeen for %s, since it is too long since the last one", le.name)
+			le.firstSeen = time.Now()
+		}
+		le.lastSeen = time.Now()
 	}
-	// The error is unchanged so we don't need to do anything
 }
 
 func (le *lastError) shouldRetry() bool {

go/vt/vttablet/tabletmanager/vreplication/last_error_test.go

@@ -24,32 +24,63 @@ import (
 	"github.com/stretchr/testify/require"
 )
 
-func TestLastError(t *testing.T) {
-	le := newLastError("test", 100*time.Millisecond)
+const shortWait = 1 * time.Millisecond
+const longWait = 150 * time.Millisecond
+const maxTimeInError = 100 * time.Millisecond
 
-	t.Run("long running error", func(t *testing.T) {
-		err1 := fmt.Errorf("test1")
+// TestLastErrorZeroMaxTime tests maxTimeInError = 0, should always retry
+func TestLastErrorZeroMaxTime(t *testing.T) {
+	le := newLastError("test", 0)
+	err1 := fmt.Errorf("error1")
+	le.record(err1)
+	require.True(t, le.shouldRetry())
+	time.Sleep(shortWait)
+	require.True(t, le.shouldRetry())
+	time.Sleep(longWait)
+	require.True(t, le.shouldRetry())
+}
+
+// TestLastErrorNoError ensures that an uninitialized lastError always retries
+func TestLastErrorNoError(t *testing.T) {
+	le := newLastError("test", maxTimeInError)
+	require.True(t, le.shouldRetry())
+	err1 := fmt.Errorf("error1")
+	le.record(err1)
+	require.True(t, le.shouldRetry())
+	le.record(nil)
+	require.True(t, le.shouldRetry())
+}
+
+// TestLastErrorOneError validates that we retry an error if happening within the maxTimeInError, but not after
+func TestLastErrorOneError(t *testing.T) {
+	le := newLastError("test", maxTimeInError)
+	err1 := fmt.Errorf("error1")
+	le.record(err1)
+	require.True(t, le.shouldRetry())
+	time.Sleep(shortWait)
+	require.True(t, le.shouldRetry())
+	time.Sleep(shortWait)
+	require.True(t, le.shouldRetry())
+	time.Sleep(longWait)
+	require.False(t, le.shouldRetry())
+}
+
+// TestLastErrorRepeatedError confirms that if same error is repeated we don't retry
+// unless it happens after maxTimeInError
+func TestLastErrorRepeatedError(t *testing.T) {
+	le := newLastError("test", maxTimeInError)
+	err1 := fmt.Errorf("error1")
+	le.record(err1)
+	require.True(t, le.shouldRetry())
+	for i := 1; i < 10; i++ {
 		le.record(err1)
-		require.True(t, le.shouldRetry())
-		time.Sleep(150 * time.Millisecond)
-		require.False(t, le.shouldRetry())
-	})
-
-	t.Run("new long running error", func(t *testing.T) {
-		err2 := fmt.Errorf("test2")
-		le.record(err2)
-		require.True(t, le.shouldRetry())
-		for i := 1; i < 10; i++ {
-			le.record(err2)
-		}
-		require.True(t, le.shouldRetry())
-		time.Sleep(150 * time.Millisecond)
-		le.record(err2)
-		require.False(t, le.shouldRetry())
-	})
-
-	t.Run("no error", func(t *testing.T) {
-		le.record(nil)
-		require.True(t, le.shouldRetry())
-	})
+		time.Sleep(shortWait)
+	}
+	require.True(t, le.shouldRetry())
+
+	// same error happens after maxTimeInError, so it should retry
+	time.Sleep(longWait)
+	require.False(t, le.shouldRetry())
+	le.record(err1)
+	require.True(t, le.shouldRetry())
 }