Zhen Ye c7b5c23ff6
enhance: filter the empty timetick from consuming side (#46541)
issue: #46540

Empty timetick is just used to sync up the time clock between different
component in milvus. So empty timetick can be ignored if we achieve the
lsn/mvcc semantic for timetick. Currently, some components need the
empty timetick to trigger some operation, such as flush/tsafe. So we
only slow down the empty time tick for 5 seconds.

<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
- Core invariant: with LSN/MVCC semantics consumers only need (a) the
first timetick that advances the latest-required-MVCC to unblock
MVCC-dependent waits and (b) occasional periodic timeticks (~≤5s) for
clock synchronization—therefore frequent non-persisted empty timeticks
can be suppressed without breaking MVCC correctness.
- Logic removed/simplified: per-message dispatch/consumption of frequent
non-persisted empty timeticks is suppressed — an MVCC-aware filter
emptyTimeTickSlowdowner (internal/util/pipeline/consuming_slowdown.go)
short-circuits frequent empty timeticks in the stream pipeline
(internal/util/pipeline/stream_pipeline.go), and the WAL flusher
rate-limits non-persisted timetick dispatch to one emission per ~5s
(internal/streamingnode/server/flusher/flusherimpl/wal_flusher.go); the
delegator exposes GetLatestRequiredMVCCTimeTick to drive the filter
(internal/querynodev2/delegator/delegator.go).
- Why this does NOT introduce data loss or regressions: the slowdowner
always refreshes latestRequiredMVCCTimeTick via
GetLatestRequiredMVCCTimeTick and (1) never filters timeticks <
latestRequiredMVCCTimeTick (so existing tsafe/flush waits stay
unblocked) and (2) always lets the first timetick ≥
latestRequiredMVCCTimeTick pass to notify pending MVCC waits;
separately, WAL flusher suppression applies only to non-persisted
timeticks and still emits when the 5s threshold elapses, preserving
periodic clock-sync messages used by flush/tsafe.
- Enhancement summary (where it takes effect): adds
GetLatestRequiredMVCCTimeTick on ShardDelegator and
LastestMVCCTimeTickGetter, wires emptyTimeTickSlowdowner into
NewPipelineWithStream (internal/util/pipeline), and adds WAL flusher
rate-limiting + metrics
(internal/streamingnode/server/flusher/flusherimpl/wal_flusher.go,
pkg/metrics) to reduce CPU/dispatch overhead while keeping MVCC
correctness and periodic synchronization.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->

---------

Signed-off-by: chyezh <chyezh@outlook.com>
2026-01-06 20:53:24 +08:00
..

Milvus Go Client Test Framework

Overview

This is a comprehensive test framework for the Milvus Go Client, designed to validate various functionalities of the Milvus vector database client. The framework provides a structured approach to writing tests with reusable components and helper functions.

Framework Architecture

Directory Structure

/go_client/
├── testcases/           # Main test cases
│   ├── helper/          # Helper functions and utilities
│   │   ├── helper.go
│   │   ├── data_helper.go
│   │   └── collection_helper.go
│   ├── search_test.go   # Search functionality tests
│   ├── index_test.go    # Index management tests
│   └── ...
├── common/             # Common utilities and constants
└── base/               # Base infrastructure code

Key Components

  • Collection Preparation: Utilities for creating and managing collections
  • Data Generation: Tools for generating test data
  • Helper Functions: Common operations and validations
  • Test Cases: Organized by functionality

Writing Test Cases

Basic Test Structure

func TestYourFeature(t *testing.T) {
    // 1. Setup context and client
    ctx := hp.CreateContext(t, time.Second*common.DefaultTimeout)
    mc := createDefaultMilvusClient(ctx, t)

    // 2. Prepare collection
    prepare, schema := hp.CollPrepare.CreateCollection(
        ctx, t, mc,
        hp.NewCreateCollectionParams(hp.Int64Vec),
        hp.TNewFieldsOption(),
        hp.TNewSchemaOption(),
    )

    // 3. Insert test data
    prepare.InsertData(ctx, t, mc,
        hp.NewInsertParams(schema),
        hp.TNewDataOption(),
    )

    // 4. Execute test operations
    // ... your test logic here ...

    // 5. Validate results
    require.NoError(t, err)
    require.Equal(t, expected, actual)
}

Using Custom Parameters

  1. Collection Creation Parameters
fieldsOption := hp.TNewFieldsOption().
    TWithEnableAnalyzer(true).
    TWithAnalyzerParams(map[string]any{
        "tokenizer": "standard",
    })

schemaOption := hp.TNewSchemaOption().
    TWithEnableDynamicField(true).
    TWithDescription("Custom schema").
    TWithAutoID(false)
  1. Data Insertion Options
insertOption := hp.TNewDataOption().
    TWithNb(1000).           // Number of records
    TWithDim(128).           // Vector dimension
    TWithStart(100).         // Starting ID
    TWithMaxLen(256).        // Maximum length
    TWithTextLang("en")      // Text language
  1. Index Parameters
indexParams := hp.TNewIndexParams(schema).
    TWithFieldIndex(map[string]index.Index{
        common.DefaultVectorFieldName: index.NewIVFSQIndex(
            &index.IVFSQConfig{
                MetricType: entity.L2,
                NList:     128,
            },
        ),
    })
  1. Search Parameters
searchOpt := client.NewSearchOption(schema.CollectionName, 100, vectors).
    WithOffset(0).
    WithLimit(100).
    WithConsistencyLevel(entity.ClStrong).
    WithFilter("int64 >= 100").
    WithOutputFields([]string{"*"}).
    WithSearchParams(map[string]any{
        "nprobe": 16,
        "ef":     64,
    })

Adding New Parameters

  1. Define New Option Type
// In helper/data_helper.go
type YourNewOption struct {
    newParam1 string
    newParam2 int
}
  1. Add Constructor
func TNewYourOption() *YourNewOption {
    return &YourNewOption{
        newParam1: "default",
        newParam2: 0,
    }
}
  1. Add Parameter Methods
func (opt *YourNewOption) TWithNewParam1(value string) *YourNewOption {
    opt.newParam1 = value
    return opt
}

func (opt *YourNewOption) TWithNewParam2(value int) *YourNewOption {
    opt.newParam2 = value
    return opt
}

Best Practices

  1. Test Organization

    • Group related tests in the same file
    • Use clear and descriptive test names
    • Add comments explaining test purpose
  2. Data Generation

    • Use helper functions for generating test data
    • Ensure data is appropriate for the test case
    • Clean up test data after use
  3. Error Handling

    • Use common.CheckErr for consistent error checking
    • Test both success and failure scenarios
    • Validate error messages when appropriate
  4. Performance Considerations

    • Use appropriate timeouts
    • Clean up resources after tests
    • Consider test execution time

Running Tests

# Run all tests
go test ./testcases/...

# Run specific test
go test -run TestYourFeature ./testcases/

# Run with verbose output
go test -v ./testcases/...

# gotestsum
Recommend you to use gotestsum https://github.com/gotestyourself/gotestsum

# Run all default cases
gotestsum --format testname --hide-summary=output -v ./testcases/... --addr=127.0.0.1:19530 -timeout=30m

# Run a specified file
gotestsum --format testname --hide-summary=output ./testcases/collection_test.go ./testcases/main_test.go --addr=127.0.0.1:19530

# Run L3 rg cases
gotestsum --format testname --hide-summary=output -v ./testcases/advcases/... --addr=127.0.0.1:19530 -timeout=30m -tags=rg

# Run advanced rg cases and default cases
# rg cases conflicts with default cases, so -p=1 is required
gotestsum --format testname --hide-summary=output -v ./testcases/... --addr=127.0.0.1:19530 -timeout=30m -tags=rg -p 1

Contributing

  1. Follow the existing code structure
  2. Add comprehensive test cases
  3. Document new parameters and options
  4. Update this README for significant changes
  5. Ensure code quality standards:
    • Run golangci-lint run to check for style mistakes
    • Use gofmt -w your/code/path to format your code before submitting
    • CI will verify both golint and go format compliance