Close shard if get workflow execution failed on shardownershiplost error

common/persistence/executionManager.go

@@ -73,6 +73,7 @@ func (m *executionManagerImpl) GetWorkflowExecution(
 	internalRequest := &InternalGetWorkflowExecutionRequest{
 		DomainID:  request.DomainID,
 		Execution: request.Execution,
+		RangeID:   request.RangeID,
 	}
 	response, err := m.persistence.GetWorkflowExecution(ctx, internalRequest)
 	if err != nil {
@@ -642,12 +643,9 @@ func (m *executionManagerImpl) SerializeWorkflowMutation(
 	if err != nil {
 		return nil, err
 	}
-	var checksumData *DataBlob
-	if len(input.Checksum.Value) > 0 {
-		checksumData, err = m.serializer.SerializeChecksum(input.Checksum, common.EncodingTypeJSON)
-		if err != nil {
-			return nil, err
-		}
+	checksumData, err := m.serializer.SerializeChecksum(input.Checksum, common.EncodingTypeJSON)
+	if err != nil {
+		return nil, err
 	}
 
 	return &InternalWorkflowMutation{

common/persistence/serializer.go

@@ -304,6 +304,9 @@ func (t *serializerImpl) DeserializeAsyncWorkflowsConfig(data *DataBlob) (*types
 }
 
 func (t *serializerImpl) SerializeChecksum(sum checksum.Checksum, encodingType common.EncodingType) (*DataBlob, error) {
+	if len(sum.Value) == 0 {
+		return nil, nil
+	}
 	return t.serialize(sum, encodingType)
 }
 

service/history/execution/context.go

@@ -1209,9 +1209,6 @@ func (c *contextImpl) getWorkflowExecutionWithRetry(
 	case *types.EntityNotExistsError:
 		// it is possible that workflow does not exists
 		return nil, err
-	case *persistence.ShardOwnershipLostError:
-		// shard is stolen, should stop processing the workflow
-		return nil, err
 	default:
 		c.logger.Error(
 			"Persistent fetch operation failure",

service/history/shard/context.go

@@ -590,7 +590,24 @@ func (s *contextImpl) GetWorkflowExecution(
 	}
 	currentRangeID := s.getRangeID()
 	request.RangeID = currentRangeID
-	return s.executionManager.GetWorkflowExecution(ctx, request)
+	response, err := s.executionManager.GetWorkflowExecution(ctx, request)
+	switch err.(type) {
+	case nil:
+		return response, nil
+	case *persistence.ShardOwnershipLostError:
+		{
+			// Shard is stolen, trigger shutdown of history engine
+			s.logger.Warn(
+				"Closing shard: GetWorkflowExecution failed due to stolen shard.",
+				tag.Error(err),
+				tag.ShardRangeID(currentRangeID),
+			)
+			s.closeShard()
+			return nil, err
+		}
+	default:
+		return nil, err
+	}
 }
 
 func (s *contextImpl) CreateWorkflowExecution(