diff --git a/br/pkg/lightning/restore/meta_manager.go b/br/pkg/lightning/restore/meta_manager.go
index ace5b91458746..e74fbdb333d54 100644
--- a/br/pkg/lightning/restore/meta_manager.go
+++ b/br/pkg/lightning/restore/meta_manager.go
@@ -655,46 +655,65 @@ func (m *dbTaskMetaMgr) CheckTasksExclusively(ctx context.Context, action func(t
 	if err != nil {
 		return errors.Annotate(err, "enable pessimistic transaction failed")
 	}
-	return exec.Transact(ctx, "check tasks exclusively", func(ctx context.Context, tx *sql.Tx) error {
-		rows, err := tx.QueryContext(
-			ctx,
-			fmt.Sprintf("SELECT task_id, pd_cfgs, status, state, source_bytes, cluster_avail from %s FOR UPDATE", m.tableName),
-		)
-		if err != nil {
-			return errors.Annotate(err, "fetch task metas failed")
-		}
-		defer rows.Close()
 
-		var tasks []taskMeta
-		for rows.Next() {
-			var task taskMeta
-			var statusValue string
-			if err = rows.Scan(&task.taskID, &task.pdCfgs, &statusValue, &task.state, &task.sourceBytes, &task.clusterAvail); err != nil {
+	// there are cases where we have to restart the transaction like error code 1205. We are adding a retry with backoff
+	backOffTime := time.Second
+	retryCnt := 5
+	for i := 0; i < retryCnt; i++ {
+		err = exec.Transact(ctx, "check tasks exclusively", func(ctx context.Context, tx *sql.Tx) error {
+			rows, err := tx.QueryContext(
+				ctx,
+				fmt.Sprintf("SELECT task_id, pd_cfgs, status, state, source_bytes, cluster_avail from %s FOR UPDATE", m.tableName),
+			)
+			if err != nil {
+				return errors.Annotate(err, "fetch task metas failed")
+			}
+			defer rows.Close()
+
+			var tasks []taskMeta
+			for rows.Next() {
+				var task taskMeta
+				var statusValue string
+				if err = rows.Scan(&task.taskID, &task.pdCfgs, &statusValue, &task.state, &task.sourceBytes, &task.clusterAvail); err != nil {
+					return errors.Trace(err)
+				}
+				status, err := parseTaskMetaStatus(statusValue)
+				if err != nil {
+					return err
+				}
+				task.status = status
+				tasks = append(tasks, task)
+			}
+			if err = rows.Err(); err != nil {
 				return errors.Trace(err)
 			}
-			status, err := parseTaskMetaStatus(statusValue)
+			newTasks, err := action(tasks)
 			if err != nil {
-				return err
+				return errors.Trace(err)
 			}
-			task.status = status
-			tasks = append(tasks, task)
-		}
-		if err = rows.Err(); err != nil {
-			return errors.Trace(err)
+			for _, task := range newTasks {
+				// nolint:gosec
+				query := fmt.Sprintf("REPLACE INTO %s (task_id, pd_cfgs, status, state, source_bytes, cluster_avail) VALUES(?, ?, ?, ?, ?, ?)", m.tableName)
+				if _, err = tx.ExecContext(ctx, query, task.taskID, task.pdCfgs, task.status.String(), task.state, task.sourceBytes, task.clusterAvail); err != nil {
+					return errors.Trace(err)
+				}
+			}
+			return nil
+		})
+		if err == nil {
+			break
 		}
-		newTasks, err := action(tasks)
-		if err != nil {
-			return errors.Trace(err)
+		backOffTime *= 2
+		if backOffTime > maxBackoffTime {
+			backOffTime = maxBackoffTime
 		}
-		for _, task := range newTasks {
-			// nolint:gosec
-			query := fmt.Sprintf("REPLACE INTO %s (task_id, pd_cfgs, status, state, source_bytes, cluster_avail) VALUES(?, ?, ?, ?, ?, ?)", m.tableName)
-			if _, err = tx.ExecContext(ctx, query, task.taskID, task.pdCfgs, task.status.String(), task.state, task.sourceBytes, task.clusterAvail); err != nil {
-				return errors.Trace(err)
-			}
+		select {
+		case <-time.After(backOffTime):
+		case <-ctx.Done():
+			return errors.Trace(ctx.Err())
 		}
-		return nil
-	})
+	}
+	return err
 }
 
 func (m *dbTaskMetaMgr) CheckAndPausePdSchedulers(ctx context.Context) (pdutil.UndoFunc, error) {