diff --git a/syncer/checkpoint.go b/syncer/checkpoint.go
index 576c215153..7c4cee91b1 100644
--- a/syncer/checkpoint.go
+++ b/syncer/checkpoint.go
@@ -15,6 +15,7 @@ package syncer
 
 import (
 	"bytes"
+	"context"
 	"database/sql"
 	"encoding/json"
 	"fmt"
@@ -321,8 +322,11 @@ func (cp *RemoteCheckPoint) Clear(tctx *tcontext.Context) error {
 	defer cp.Unlock()
 
 	// delete all checkpoints
+	// use a new context apart from syncer, to make sure when syncer call `cancel` checkpoint could update
+	tctx2, cancel := tctx.WithContext(context.Background()).WithTimeout(maxDMLConnectionDuration)
+	defer cancel()
 	_, err := cp.dbConn.executeSQL(
-		tctx,
+		tctx2,
 		[]string{`DELETE FROM ` + cp.tableName + ` WHERE id = ?`},
 		[]interface{}{cp.id},
 	)
@@ -390,9 +394,12 @@ func (cp *RemoteCheckPoint) DeleteTablePoint(tctx *tcontext.Context, sourceSchem
 		return nil
 	}
 
+	// use a new context apart from syncer, to make sure when syncer call `cancel` checkpoint could update
+	tctx2, cancel := tctx.WithContext(context.Background()).WithTimeout(maxDMLConnectionDuration)
+	defer cancel()
 	cp.logCtx.L().Info("delete table checkpoint", zap.String("schema", sourceSchema), zap.String("table", sourceTable))
 	_, err := cp.dbConn.executeSQL(
-		tctx,
+		tctx2,
 		[]string{`DELETE FROM ` + cp.tableName + ` WHERE id = ? AND cp_schema = ? AND cp_table = ?`},
 		[]interface{}{cp.id, sourceSchema, sourceTable},
 	)
@@ -412,9 +419,12 @@ func (cp *RemoteCheckPoint) DeleteSchemaPoint(tctx *tcontext.Context, sourceSche
 		return nil
 	}
 
+	// use a new context apart from syncer, to make sure when syncer call `cancel` checkpoint could update
+	tctx2, cancel := tctx.WithContext(context.Background()).WithTimeout(maxDMLConnectionDuration)
+	defer cancel()
 	cp.logCtx.L().Info("delete schema checkpoint", zap.String("schema", sourceSchema))
 	_, err := cp.dbConn.executeSQL(
-		tctx,
+		tctx2,
 		[]string{`DELETE FROM ` + cp.tableName + ` WHERE id = ? AND cp_schema = ?`},
 		[]interface{}{cp.id, sourceSchema},
 	)
@@ -522,7 +532,10 @@ func (cp *RemoteCheckPoint) FlushPointsExcept(tctx *tcontext.Context, exceptTabl
 		args = append(args, extraArgs[i])
 	}
 
-	_, err := cp.dbConn.executeSQL(tctx, sqls, args...)
+	// use a new context apart from syncer, to make sure when syncer call `cancel` checkpoint could update
+	tctx2, cancel := tctx.WithContext(context.Background()).WithTimeout(maxDMLConnectionDuration)
+	defer cancel()
+	_, err := cp.dbConn.executeSQL(tctx2, sqls, args...)
 	if err != nil {
 		return err
 	}
diff --git a/syncer/syncer.go b/syncer/syncer.go
index d6edf73636..d27c7fe0dc 100644
--- a/syncer/syncer.go
+++ b/syncer/syncer.go
@@ -872,9 +872,7 @@ func (s *Syncer) flushCheckPoints() error {
 		s.tctx.L().Info("prepare flush sqls", zap.Strings("shard meta sqls", shardMetaSQLs), zap.Reflect("shard meta arguments", shardMetaArgs))
 	}
 
-	tctx, cancel := s.tctx.WithContext(context.Background()).WithTimeout(maxDMLConnectionDuration)
-	defer cancel()
-	err := s.checkpoint.FlushPointsExcept(tctx, exceptTables, shardMetaSQLs, shardMetaArgs)
+	err := s.checkpoint.FlushPointsExcept(s.tctx, exceptTables, shardMetaSQLs, shardMetaArgs)
 	if err != nil {
 		return terror.Annotatef(err, "flush checkpoint %s", s.checkpoint)
 	}