fix:support config index offsetcache and fix create same index again (#…

…35985) #35971 Signed-off-by: luzhang <[email protected]> Co-authored-by: luzhang <[email protected]>
milvus-io · Sep 8, 2024 · 208c8a2 · 208c8a2
1 parent 91d23ec
commit 208c8a2
Show file tree

Hide file tree

Showing 7 changed files with 70 additions and 8 deletions.
diff --git a/configs/milvus.yaml b/configs/milvus.yaml
@@ -425,6 +425,7 @@ queryNode:
     requestResourceRetryInterval: 2000 # retry interval in milliseconds for waiting request resource for lazy load, 2s by default
     maxRetryTimes: 1 # max retry times for lazy load, 1 by default
     maxEvictPerRetry: 1 # max evict count for lazy load, 1 by default
+  indexOffsetCacheEnabled: false # enable index offset cache for some scalar indexes, now is just for bitmap index, enable this param can improve performance for retrieving raw data from index
   grouping:
     enabled: true
     maxNQ: 1000

diff --git a/internal/datacoord/index_meta.go b/internal/datacoord/index_meta.go
@@ -38,6 +38,7 @@ import (
 	"github.com/milvus-io/milvus/pkg/log"
 	"github.com/milvus-io/milvus/pkg/metrics"
 	"github.com/milvus-io/milvus/pkg/util/indexparamcheck"
+	"github.com/milvus-io/milvus/pkg/util/indexparams"
 	"github.com/milvus-io/milvus/pkg/util/timerecord"
 	"github.com/milvus-io/milvus/pkg/util/typeutil"
 )
@@ -191,21 +192,21 @@ func checkParams(fieldIndex *model.Index, req *indexpb.CreateIndexRequest) bool
 	}
 
 	useAutoIndex := false
-	userIndexParamsWithoutMmapKey := make([]*commonpb.KeyValuePair, 0)
+	userIndexParamsWithoutConfigableKey := make([]*commonpb.KeyValuePair, 0)
 	for _, param := range fieldIndex.UserIndexParams {
-		if param.Key == common.MmapEnabledKey {
+		if indexparams.IsConfigableIndexParam(param.Key) {
 			continue
 		}
 		if param.Key == common.IndexTypeKey && param.Value == common.AutoIndexName {
 			useAutoIndex = true
 		}
-		userIndexParamsWithoutMmapKey = append(userIndexParamsWithoutMmapKey, param)
+		userIndexParamsWithoutConfigableKey = append(userIndexParamsWithoutConfigableKey, param)
 	}
 
-	if len(userIndexParamsWithoutMmapKey) != len(req.GetUserIndexParams()) {
+	if len(userIndexParamsWithoutConfigableKey) != len(req.GetUserIndexParams()) {
 		return false
 	}
-	for _, param1 := range userIndexParamsWithoutMmapKey {
+	for _, param1 := range userIndexParamsWithoutConfigableKey {
 		exist := false
 		for i, param2 := range req.GetUserIndexParams() {
 			if param2.Key == param1.Key && param2.Value == param1.Value {

diff --git a/internal/datacoord/index_service.go b/internal/datacoord/index_service.go
@@ -183,6 +183,7 @@ func (s *Server) CreateIndex(ctx context.Context, req *indexpb.CreateIndexReques
 		zap.String("IndexName", req.GetIndexName()), zap.Int64("fieldID", req.GetFieldID()),
 		zap.Any("TypeParams", req.GetTypeParams()),
 		zap.Any("IndexParams", req.GetIndexParams()),
+		zap.Any("UserIndexParams", req.GetUserIndexParams()),
 	)
 
 	if err := merr.CheckHealthy(s.GetStateCode()); err != nil {
@@ -343,7 +344,7 @@ func (s *Server) AlterIndex(ctx context.Context, req *indexpb.AlterIndexRequest)
 
 		// update index params
 		newIndexParams := UpdateParams(index, index.IndexParams, req.GetParams())
-		log.Info("alter index user index params",
+		log.Info("alter index index params",
 			zap.String("indexName", index.IndexName),
 			zap.Any("params", newIndexParams),
 		)

diff --git a/internal/querynodev2/segments/segment.go b/internal/querynodev2/segments/segment.go
@@ -1116,6 +1116,11 @@ func (s *LocalSegment) LoadIndex(ctx context.Context, indexInfo *querypb.FieldIn
 		}
 	}
 
+	// set whether enable offset cache for bitmap index
+	if indexParams["index_type"] == indexparamcheck.IndexBitmap {
+		indexparams.SetBitmapIndexLoadParams(paramtable.Get(), indexParams)
+	}
+
 	if err := indexparams.AppendPrepareLoadParams(paramtable.Get(), indexParams); err != nil {
 		return err
 	}

diff --git a/pkg/util/indexparams/index_params.go b/pkg/util/indexparams/index_params.go
@@ -299,6 +299,14 @@ func SetDiskIndexBuildParams(indexParams map[string]string, fieldDataSize int64)
 	return nil
 }
 
+func SetBitmapIndexLoadParams(params *paramtable.ComponentParam, indexParams map[string]string) {
+	_, exist := indexParams[common.IndexOffsetCacheEnabledKey]
+	if exist {
+		return
+	}
+	indexParams[common.IndexOffsetCacheEnabledKey] = params.QueryNodeCfg.IndexOffsetCacheEnabled.GetValue()
+}
+
 // SetDiskIndexLoadParams set disk index load params with ratio params on queryNode
 // QueryNode cal load params with ratio params ans cpu count...
 func SetDiskIndexLoadParams(params *paramtable.ComponentParam, indexParams map[string]string, numRows int64) error {

diff --git a/pkg/util/paramtable/component_param.go b/pkg/util/paramtable/component_param.go
@@ -2358,6 +2358,8 @@ type queryNodeConfig struct {
 	LazyLoadMaxRetryTimes                ParamItem `refreshable:"true"`
 	LazyLoadMaxEvictPerRetry             ParamItem `refreshable:"true"`
 
+	IndexOffsetCacheEnabled ParamItem `refreshable:"true"`
+
 	// chunk cache
 	ReadAheadPolicy     ParamItem `refreshable:"false"`
 	ChunkCacheWarmingUp ParamItem `refreshable:"true"`
@@ -2864,6 +2866,16 @@ Max read concurrency must greater than or equal to 1, and less than or equal to
 	}
 	p.EnableDisk.Init(base.mgr)
 
+	p.IndexOffsetCacheEnabled = ParamItem{
+		Key:          "queryNode.indexOffsetCacheEnabled",
+		Version:      "2.5.0",
+		DefaultValue: "false",
+		Doc: "enable index offset cache for some scalar indexes, now is just for bitmap index," +
+			" enable this param can improve performance for retrieving raw data from index",
+		Export: true,
+	}
+	p.IndexOffsetCacheEnabled.Init(base.mgr)
+
 	p.DiskCapacityLimit = ParamItem{
 		Key:     "LOCAL_STORAGE_SIZE",
 		Version: "2.2.0",

diff --git a/tests/python_client/testcases/test_query.py b/tests/python_client/testcases/test_query.py
@@ -2703,6 +2703,40 @@ def test_query_string_expr_with_prefixes(self):
         collection_w.query(expression, output_fields=output_fields,
                            check_task=CheckTasks.check_query_results, check_items={exp_res: res})
 
+    @pytest.mark.tags(CaseLabel.L1)
+    def test_bitmap_alter_offset_cache_param(self):
+        """
+        target: test bitmap index with enable offset cache.
+        expected: verify create index and load successfully
+        """
+        collection_w, vectors = self.init_collection_general(prefix, insert_data=True,is_index=False,
+                                                             primary_field=default_int_field_name)[0:2]
+
+        collection_w.create_index(ct.default_float_vec_field_name, default_index_params, index_name="test_vec")
+        collection_w.create_index("varchar", index_name="bitmap_offset_cache", index_params={"index_type": "BITMAP"})
+        time.sleep(1)
+        collection_w.load()
+        expression = 'varchar like "0%"'
+        result , _ = collection_w.query(expression, output_fields=['varchar'])
+        res_len = len(result)
+        collection_w.release()
+        collection_w.alter_index("bitmap_offset_cache",  {'indexoffsetcache.enabled': True})
+        collection_w.create_index("varchar", index_name="bitmap_offset_cache", index_params={"index_type": "BITMAP"})
+        collection_w.load()
+        expression = 'varchar like "0%"'
+        result , _ = collection_w.query(expression, output_fields=['varchar'])
+        res_len_new = len(result)
+        assert res_len_new == res_len
+        collection_w.release()
+        collection_w.alter_index("bitmap_offset_cache",  {'indexoffsetcache.enabled': False})
+        collection_w.create_index("varchar", index_name="bitmap_offset_cache", index_params={"index_type": "BITMAP"})
+        collection_w.load()
+        expression = 'varchar like "0%"'
+        result , _ = collection_w.query(expression, output_fields=['varchar'])
+        res_len_new = len(result)
+        assert res_len_new == res_len
+        collection_w.release()
+
     @pytest.mark.tags(CaseLabel.L1)
     def test_query_string_expr_with_prefixes_auto_index(self):
         """
@@ -2736,7 +2770,7 @@ def test_query_string_expr_with_prefixes_bitmap(self):
                                                              primary_field=default_int_field_name)[0:2]
 
         collection_w.create_index(ct.default_float_vec_field_name, default_index_params, index_name="query_expr_pre_index")
-        collection_w.create_index("varchar", index_name="bitmap_auto_index")
+        collection_w.create_index("varchar", index_name="bitmap_auto_index", index_params={"index_type": "BITMAP"})
         time.sleep(1)
         collection_w.load()
         expression = 'varchar like "0%"'
@@ -2782,7 +2816,7 @@ def test_query_string_expr_with_match_bitmap(self):
                                                              primary_field=default_int_field_name)[0:2]
 
         collection_w.create_index(ct.default_float_vec_field_name, default_index_params, index_name="query_expr_pre_index")
-        collection_w.create_index("varchar", index_name="bitmap_auto_index")
+        collection_w.create_index("varchar", index_name="bitmap_auto_index",  index_params={"index_type": "BITMAP"})
         time.sleep(1)
         collection_w.load()
         expression = 'varchar like "%0%"'