clear checksum

common/persistence/sql/sql_execution_store.go

@@ -35,7 +35,6 @@ import (
 	"github.com/uber/cadence/common"
 	"github.com/uber/cadence/common/collection"
 	"github.com/uber/cadence/common/log"
-	"github.com/uber/cadence/common/persistence"
 	p "github.com/uber/cadence/common/persistence"
 	"github.com/uber/cadence/common/persistence/serialization"
 	"github.com/uber/cadence/common/persistence/sql/sqlplugin"
@@ -380,17 +379,16 @@ func (m *sqlExecutionStore) GetWorkflowExecution(
 	// if the rangeID changed, it means the shard ownership might have changed
 	// and the workflow might have been updated when we read the data, so the data
 	// we read might not be from a consistent view, the checksum validation might fail
-	// in that case, we need to return an error
+	// in that case, we clear the checksum data so that we will not perform the validation
 	if state.ChecksumData != nil {
 		row, err := m.db.SelectFromShards(ctx, &sqlplugin.ShardsFilter{ShardID: int64(m.shardID)})
 		if err != nil {
 			return nil, convertCommonErrors(m.db, "GetWorkflowExecution", "", err)
 		}
 		if row.RangeID != request.RangeID {
-			return nil, &persistence.ShardOwnershipLostError{
-				ShardID: m.shardID,
-				Msg:     fmt.Sprintf("GetWorkflowExecution failed. Previous rangeID: %v, new rangeID: %v", request.RangeID, row.RangeID),
-			}
+			// The GetWorkflowExecution operation will not be impacted by this. ChecksumData is purely for validation purposes.
+			m.logger.Warn("GetWorkflowExecution's checksum is discarded. The shard might have changed owner.")
+			state.ChecksumData = nil
 		}
 	}
 

common/persistence/sql/sql_execution_store_test.go

@@ -3408,10 +3408,25 @@ func TestGetWorkflowExecution(t *testing.T) {
 					RangeID: 1,
 				}, nil)
 			},
-			wantErr: true,
-			assertErr: func(t *testing.T, err error) {
-				assert.IsType(t, &persistence.ShardOwnershipLostError{}, err)
+			want: &persistence.InternalGetWorkflowExecutionResponse{
+				State: &persistence.InternalWorkflowMutableState{
+					ExecutionInfo: &persistence.InternalWorkflowExecutionInfo{
+						DomainID:               "ff9c8a3f-0e4f-4d3e-a4d2-6f5f8f3f7d9d",
+						WorkflowID:             "test-workflow-id",
+						RunID:                  "ee8d7b6e-876c-4b1e-9b6e-5e3e3c6b6b3f",
+						NextEventID:            101,
+						CompletionEventBatchID: -23,
+					},
+					ActivityInfos:       map[int64]*persistence.InternalActivityInfo{},
+					TimerInfos:          map[string]*persistence.TimerInfo{},
+					ChildExecutionInfos: map[int64]*persistence.InternalChildExecutionInfo{},
+					RequestCancelInfos:  map[int64]*persistence.RequestCancelInfo{},
+					SignalInfos:         map[int64]*persistence.SignalInfo{},
+					SignalRequestedIDs:  map[string]struct{}{},
+					ChecksumData:        nil,
+				},
 			},
+			wantErr: false,
 		},
 		{
 			name: "Error - failed to get shard",

service/history/shard/context.go

@@ -585,29 +585,11 @@ func (s *contextImpl) GetWorkflowExecution(
 	ctx context.Context,
 	request *persistence.GetWorkflowExecutionRequest,
 ) (*persistence.GetWorkflowExecutionResponse, error) {
+	request.RangeID = atomic.LoadInt64(&s.rangeID)
 	if s.isClosed() {
 		return nil, ErrShardClosed
 	}
-	currentRangeID := s.getRangeID()
-	request.RangeID = currentRangeID
-	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
-	}
+	return s.executionManager.GetWorkflowExecution(ctx, request)
 }
 
 func (s *contextImpl) CreateWorkflowExecution(