mirror of
https://github.com/qdrant/qdrant.git
synced 2026-07-30 06:30:57 -05:00
* Propagate HardwareCounterCell through MmapMapIndex::get_values Previously, `MmapMapIndex::get_values` used `ConditionedCounter::never()`, which silently skipped all hardware IO counter tracking for mmap map index value reads. This propagates a real `HardwareCounterCell` through the full call chain so that disk IO from `values_iter` is properly measured. Changes: - `MmapMapIndex::get_values` now accepts `&HardwareCounterCell`, creates a `ConditionedCounter` via `make_conditioned_counter`, and measures the `deleted` bitvec access (matching `check_values_any` behavior). - `MapIndex::get_values` forwards the counter to the `Mmap` variant. - `FacetIndex::get_point_values` trait method now accepts `&HardwareCounterCell`, propagated through `FacetIndexEnum` and both impls (MapIndex, BoolIndex). - `indexed_variable_retriever` accepts and forwards the counter to `IntMapIndex`, `KeywordIndex`, and `UuidMapIndex` calls. - `SegmentBuilder::update` and `_get_ordering_value` accept and propagate the counter instead of creating disposable instances. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * feat: add test --------- Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com> Co-authored-by: Daniel Boros <dancixx@gmail.com>
584 lines
18 KiB
Rust
584 lines
18 KiB
Rust
use std::collections::HashMap;
|
|
use std::str::FromStr;
|
|
use std::sync::Arc;
|
|
use std::sync::atomic::{AtomicBool, Ordering};
|
|
use std::time::{Duration, Instant};
|
|
|
|
use common::budget::ResourcePermit;
|
|
use common::counter::hardware_counter::HardwareCounterCell;
|
|
use common::progress_tracker::ProgressTracker;
|
|
use fs_err as fs;
|
|
use itertools::Itertools;
|
|
use segment::common::operation_error::OperationError;
|
|
use segment::data_types::named_vectors::NamedVectors;
|
|
use segment::data_types::vectors::{DEFAULT_VECTOR_NAME, VectorRef, only_default_vector};
|
|
use segment::entry::entry_point::{NonAppendableSegmentEntry, ReadSegmentEntry, SegmentEntry};
|
|
use segment::id_tracker::IdTracker;
|
|
use segment::index::hnsw_index::get_num_indexing_threads;
|
|
use segment::json_path::JsonPath;
|
|
use segment::segment::Segment;
|
|
use segment::segment_constructor::segment_builder::SegmentBuilder;
|
|
use segment::segment_constructor::simple_segment_constructor::build_simple_segment_with_payload_storage;
|
|
use segment::types::{
|
|
Distance, HnswGlobalConfig, Indexes, PayloadContainer, PayloadFieldSchema, PayloadKeyType,
|
|
PayloadSchemaType, PayloadStorageType, SegmentConfig, VectorDataConfig, VectorStorageType,
|
|
};
|
|
use serde_json::Value;
|
|
use sparse::common::sparse_vector::SparseVector;
|
|
use tempfile::Builder;
|
|
use uuid::Uuid;
|
|
|
|
use crate::fixtures::segment::{
|
|
PAYLOAD_KEY, SPARSE_VECTOR_NAME, build_segment_1, build_segment_2, build_segment_sparse_1,
|
|
build_segment_sparse_2, empty_segment,
|
|
};
|
|
|
|
#[test]
|
|
fn test_building_new_segment() {
|
|
let dir = Builder::new().prefix("segment_dir").tempdir().unwrap();
|
|
let temp_dir = Builder::new().prefix("segment_temp_dir").tempdir().unwrap();
|
|
|
|
let stopped = AtomicBool::new(false);
|
|
|
|
let segment1 = build_segment_1(dir.path());
|
|
let mut segment2 = build_segment_2(dir.path());
|
|
|
|
let mut builder = SegmentBuilder::new(
|
|
temp_dir.path(),
|
|
&segment1.segment_config,
|
|
&HnswGlobalConfig::default(),
|
|
)
|
|
.unwrap();
|
|
|
|
let hw_counter = HardwareCounterCell::new();
|
|
|
|
// Include overlapping with segment1 to check the
|
|
segment2
|
|
.upsert_point(
|
|
100,
|
|
3.into(),
|
|
only_default_vector(&[0., 0., 0., 0.]),
|
|
&hw_counter,
|
|
)
|
|
.unwrap();
|
|
|
|
builder
|
|
.update(&[&segment1, &segment2, &segment2], &stopped, &hw_counter)
|
|
.unwrap();
|
|
|
|
// Check what happens if segment building fails here
|
|
|
|
let segment_count = fs::read_dir(dir.path()).unwrap().count();
|
|
|
|
assert_eq!(segment_count, 2);
|
|
|
|
let temp_segment_count = fs::read_dir(temp_dir.path()).unwrap().count();
|
|
|
|
assert_eq!(temp_segment_count, 1);
|
|
|
|
// Now we finalize building
|
|
|
|
let merged_segment: Segment = builder.build_for_test(dir.path());
|
|
|
|
let new_segment_count = fs::read_dir(dir.path()).unwrap().count();
|
|
|
|
assert_eq!(new_segment_count, 3);
|
|
|
|
assert_eq!(
|
|
merged_segment.iter_points().count(),
|
|
merged_segment.available_point_count(),
|
|
);
|
|
assert_eq!(
|
|
merged_segment.available_point_count(),
|
|
segment1
|
|
.iter_points()
|
|
.chain(segment2.iter_points())
|
|
.unique()
|
|
.count(),
|
|
);
|
|
|
|
assert_eq!(merged_segment.point_version(3.into()), Some(100));
|
|
}
|
|
|
|
#[test]
|
|
fn test_building_new_defragmented_segment() {
|
|
let dir = Builder::new().prefix("segment_dir").tempdir().unwrap();
|
|
let temp_dir = Builder::new().prefix("segment_temp_dir").tempdir().unwrap();
|
|
|
|
let stopped = AtomicBool::new(false);
|
|
|
|
let defragment_key = JsonPath::from_str(PAYLOAD_KEY).unwrap();
|
|
|
|
let hw_counter = HardwareCounterCell::new();
|
|
|
|
let payload_schema = PayloadFieldSchema::FieldType(PayloadSchemaType::Keyword);
|
|
|
|
let mut segment1 = build_segment_1(dir.path());
|
|
segment1
|
|
.create_field_index(7, &defragment_key, Some(&payload_schema), &hw_counter)
|
|
.unwrap();
|
|
|
|
let mut segment2 = build_segment_2(dir.path());
|
|
segment2
|
|
.create_field_index(17, &defragment_key, Some(&payload_schema), &hw_counter)
|
|
.unwrap();
|
|
|
|
let mut builder = SegmentBuilder::new(
|
|
temp_dir.path(),
|
|
&segment1.segment_config,
|
|
&HnswGlobalConfig::default(),
|
|
)
|
|
.unwrap();
|
|
|
|
// Include overlapping with segment1 to check the
|
|
segment2
|
|
.upsert_point(
|
|
100,
|
|
3.into(),
|
|
only_default_vector(&[0., 0., 0., 0.]),
|
|
&hw_counter,
|
|
)
|
|
.unwrap();
|
|
|
|
builder.set_defragment_keys(vec![defragment_key.clone()]);
|
|
|
|
builder
|
|
.update(&[&segment1, &segment2], &stopped, &hw_counter)
|
|
.unwrap();
|
|
|
|
// Check what happens if segment building fails here
|
|
|
|
let segment_count = fs::read_dir(dir.path()).unwrap().count();
|
|
|
|
assert_eq!(segment_count, 2);
|
|
|
|
let temp_segment_count = fs::read_dir(temp_dir.path()).unwrap().count();
|
|
|
|
assert_eq!(temp_segment_count, 1);
|
|
|
|
// Now we finalize building
|
|
|
|
let merged_segment = builder.build_for_test(dir.path());
|
|
|
|
let new_segment_count = fs::read_dir(dir.path()).unwrap().count();
|
|
|
|
assert_eq!(new_segment_count, 3);
|
|
|
|
assert_eq!(
|
|
merged_segment.iter_points().count(),
|
|
merged_segment.available_point_count(),
|
|
);
|
|
assert_eq!(
|
|
merged_segment.available_point_count(),
|
|
segment1
|
|
.iter_points()
|
|
.chain(segment2.iter_points())
|
|
.unique()
|
|
.count(),
|
|
);
|
|
|
|
assert_eq!(merged_segment.point_version(3.into()), Some(100));
|
|
|
|
if let Err(err) = check_points_defragmented(&merged_segment, &defragment_key) {
|
|
panic!("{err}");
|
|
}
|
|
}
|
|
|
|
/// Iterates over the internal point ids of the merged segment and checks that the
|
|
/// points are grouped by the payload value.
|
|
fn check_points_defragmented(
|
|
segment: &Segment,
|
|
defragment_key: &PayloadKeyType,
|
|
) -> Result<(), &'static str> {
|
|
let id_tracker = segment.id_tracker.borrow();
|
|
|
|
// Previously seen group/value.
|
|
let mut previous_value: Option<Value> = None;
|
|
|
|
// keeps track of groups/values that have already been seen while iterating
|
|
let mut seen_values: Vec<Value> = vec![];
|
|
|
|
let hw_counter = HardwareCounterCell::new();
|
|
|
|
for internal_id in id_tracker.point_mappings().iter_internal() {
|
|
let external_id = id_tracker.external_id(internal_id).unwrap();
|
|
let payload = segment.payload(external_id, &hw_counter).unwrap();
|
|
let values = payload.get_value(defragment_key);
|
|
|
|
if values.is_empty() {
|
|
if !seen_values.is_empty() {
|
|
return Err(
|
|
"In a defragmented segment, points without a payload value should come first!",
|
|
);
|
|
}
|
|
|
|
continue;
|
|
}
|
|
|
|
let value = values[0].clone();
|
|
|
|
let Some(prev) = previous_value.as_ref() else {
|
|
previous_value = Some(value);
|
|
continue;
|
|
};
|
|
|
|
if *prev == value {
|
|
continue;
|
|
}
|
|
|
|
if seen_values.contains(&value) {
|
|
return Err("Segment not defragmented");
|
|
}
|
|
|
|
seen_values.push(value.clone());
|
|
previous_value = Some(value);
|
|
}
|
|
|
|
Ok(())
|
|
}
|
|
|
|
#[test]
|
|
fn test_building_new_sparse_segment() {
|
|
let dir = Builder::new().prefix("segment_dir").tempdir().unwrap();
|
|
let temp_dir = Builder::new().prefix("segment_temp_dir").tempdir().unwrap();
|
|
|
|
let stopped = AtomicBool::new(false);
|
|
|
|
let hw_counter = HardwareCounterCell::new();
|
|
|
|
let segment1 = build_segment_sparse_1(dir.path());
|
|
let mut segment2 = build_segment_sparse_2(dir.path());
|
|
|
|
let mut builder = SegmentBuilder::new(
|
|
temp_dir.path(),
|
|
&segment1.segment_config,
|
|
&HnswGlobalConfig::default(),
|
|
)
|
|
.unwrap();
|
|
|
|
// Include overlapping with segment1 to check the
|
|
let vec = SparseVector::new(vec![0, 1, 2, 3], vec![0.0, 0.0, 0.0, 0.0]).unwrap();
|
|
segment2
|
|
.upsert_point(
|
|
100,
|
|
3.into(),
|
|
NamedVectors::from_ref(SPARSE_VECTOR_NAME, VectorRef::Sparse(&vec)),
|
|
&hw_counter,
|
|
)
|
|
.unwrap();
|
|
|
|
builder
|
|
.update(&[&segment1, &segment2, &segment2], &stopped, &hw_counter)
|
|
.unwrap();
|
|
|
|
// Check what happens if segment building fails here
|
|
|
|
let segment_count = fs::read_dir(dir.path()).unwrap().count();
|
|
|
|
assert_eq!(segment_count, 2);
|
|
|
|
let temp_segment_count = fs::read_dir(temp_dir.path()).unwrap().count();
|
|
|
|
assert_eq!(temp_segment_count, 1);
|
|
|
|
// Now we finalize building
|
|
|
|
let merged_segment = builder.build_for_test(dir.path());
|
|
|
|
let new_segment_count = fs::read_dir(dir.path()).unwrap().count();
|
|
|
|
assert_eq!(new_segment_count, 3);
|
|
|
|
assert_eq!(
|
|
merged_segment.iter_points().count(),
|
|
merged_segment.available_point_count(),
|
|
);
|
|
assert_eq!(
|
|
merged_segment.available_point_count(),
|
|
segment1
|
|
.iter_points()
|
|
.chain(segment2.iter_points())
|
|
.unique()
|
|
.count(),
|
|
);
|
|
|
|
assert_eq!(merged_segment.point_version(3.into()), Some(100));
|
|
}
|
|
|
|
fn estimate_build_time(segment: &Segment, stop_delay_millis: Option<u64>) -> (u64, bool) {
|
|
let mut rng = rand::rng();
|
|
let stopped = Arc::new(AtomicBool::new(false));
|
|
|
|
let dir = Builder::new().prefix("segment_dir1").tempdir().unwrap();
|
|
let temp_dir = Builder::new().prefix("segment_temp_dir").tempdir().unwrap();
|
|
|
|
let segment_config = SegmentConfig {
|
|
vector_data: HashMap::from([(
|
|
DEFAULT_VECTOR_NAME.to_owned(),
|
|
VectorDataConfig {
|
|
size: segment.segment_config.vector_data[DEFAULT_VECTOR_NAME].size,
|
|
distance: segment.segment_config.vector_data[DEFAULT_VECTOR_NAME].distance,
|
|
storage_type: VectorStorageType::default(),
|
|
index: Indexes::Hnsw(Default::default()),
|
|
quantization_config: None,
|
|
multivector_config: None,
|
|
datatype: None,
|
|
},
|
|
)]),
|
|
sparse_vector_data: Default::default(),
|
|
payload_storage_type: Default::default(),
|
|
};
|
|
|
|
let mut builder = SegmentBuilder::new(
|
|
temp_dir.path(),
|
|
&segment_config,
|
|
&HnswGlobalConfig::default(),
|
|
)
|
|
.unwrap();
|
|
|
|
let hw_counter = HardwareCounterCell::new();
|
|
builder.update(&[segment], &stopped, &hw_counter).unwrap();
|
|
|
|
let now = Instant::now();
|
|
|
|
if let Some(stop_delay_millis) = stop_delay_millis {
|
|
let stopped_t = stopped.clone();
|
|
|
|
std::thread::Builder::new()
|
|
.name("build_estimator_timeout".to_string())
|
|
.spawn(move || {
|
|
std::thread::sleep(Duration::from_millis(stop_delay_millis));
|
|
stopped_t.store(true, Ordering::Release);
|
|
})
|
|
.unwrap();
|
|
}
|
|
|
|
let permit_cpu_count = get_num_indexing_threads(0);
|
|
let permit = ResourcePermit::dummy(permit_cpu_count as u32);
|
|
let hw_counter = HardwareCounterCell::new();
|
|
let progress = ProgressTracker::new_for_test();
|
|
|
|
let res = builder.build(
|
|
dir.path(),
|
|
Uuid::new_v4(),
|
|
None,
|
|
permit,
|
|
&stopped,
|
|
&mut rng,
|
|
&hw_counter,
|
|
progress,
|
|
);
|
|
|
|
let is_cancelled = match res {
|
|
Ok(_) => false,
|
|
Err(OperationError::Cancelled { .. }) => true,
|
|
Err(err) => {
|
|
eprintln!("Was expecting cancellation signal but got unexpected error: {err:?}");
|
|
false
|
|
}
|
|
};
|
|
|
|
(now.elapsed().as_millis() as u64, is_cancelled)
|
|
}
|
|
|
|
/// Unit test for a specific bug we caught before.
|
|
///
|
|
/// See: <https://github.com/qdrant/qdrant/pull/5614>
|
|
#[test]
|
|
fn test_building_new_segment_bug_5614() {
|
|
let dir = Builder::new().prefix("segment_dir").tempdir().unwrap();
|
|
let temp_dir = Builder::new().prefix("segment_temp_dir").tempdir().unwrap();
|
|
|
|
let stopped = AtomicBool::new(false);
|
|
|
|
let mut segment1 = build_segment_1(dir.path());
|
|
let mut segment2 = build_segment_2(dir.path());
|
|
|
|
let mut builder = SegmentBuilder::new(
|
|
temp_dir.path(),
|
|
&segment1.segment_config,
|
|
&HnswGlobalConfig::default(),
|
|
)
|
|
.unwrap();
|
|
|
|
let vector_100_low = only_default_vector(&[1., 1., 0., 0.]);
|
|
let vector_101_low = only_default_vector(&[2., 2., 0., 0.]);
|
|
let vector_100_high = only_default_vector(&[3., 3., 0., 0.]);
|
|
let vector_101_high = only_default_vector(&[4., 4., 0., 0.]);
|
|
|
|
let hw_counter = HardwareCounterCell::new();
|
|
|
|
// Insert point 100 and 101 in both segments
|
|
// Do this in a specific order so that:
|
|
// - the latter segment has a higher point version
|
|
// - the internal point IDs don't match across segments
|
|
segment1
|
|
.upsert_point(123, 100.into(), vector_100_low, &hw_counter)
|
|
.unwrap();
|
|
segment1
|
|
.upsert_point(123, 101.into(), vector_101_low, &hw_counter)
|
|
.unwrap();
|
|
|
|
segment2
|
|
.upsert_point(124, 101.into(), vector_101_high.clone(), &hw_counter)
|
|
.unwrap();
|
|
segment2
|
|
.upsert_point(124, 100.into(), vector_100_high.clone(), &hw_counter)
|
|
.unwrap();
|
|
|
|
builder
|
|
.update(&[&segment1, &segment2], &stopped, &hw_counter)
|
|
.unwrap();
|
|
|
|
let hw_counter = HardwareCounterCell::new();
|
|
|
|
let merged_segment: Segment = builder.build_for_test(dir.path());
|
|
|
|
// Assert correct point versions - must have latest
|
|
assert_eq!(merged_segment.point_version(100.into()), Some(124));
|
|
assert_eq!(merged_segment.point_version(101.into()), Some(124));
|
|
|
|
// Assert correct vectors still belong to the point
|
|
// This was broken before <https://github.com/qdrant/qdrant/pull/5543>
|
|
assert_eq!(
|
|
merged_segment.all_vectors(100.into(), &hw_counter).unwrap(),
|
|
vector_100_high,
|
|
);
|
|
assert_eq!(
|
|
merged_segment.all_vectors(101.into(), &hw_counter).unwrap(),
|
|
vector_101_high,
|
|
);
|
|
}
|
|
|
|
#[test]
|
|
fn test_building_cancellation() {
|
|
let baseline_dir = Builder::new()
|
|
.prefix("segment_dir_baseline")
|
|
.tempdir()
|
|
.unwrap();
|
|
let dir = Builder::new().prefix("segment_dir").tempdir().unwrap();
|
|
let dir_2 = Builder::new().prefix("segment_dir_2").tempdir().unwrap();
|
|
|
|
let mut baseline_segment = empty_segment(baseline_dir.path());
|
|
let mut segment = empty_segment(dir.path());
|
|
let mut segment_2 = empty_segment(dir_2.path());
|
|
|
|
let hw_counter = HardwareCounterCell::new();
|
|
|
|
for idx in 0..2000 {
|
|
baseline_segment
|
|
.upsert_point(
|
|
1,
|
|
idx.into(),
|
|
only_default_vector(&[0., 0., 0., 0.]),
|
|
&hw_counter,
|
|
)
|
|
.unwrap();
|
|
segment
|
|
.upsert_point(
|
|
1,
|
|
idx.into(),
|
|
only_default_vector(&[0., 0., 0., 0.]),
|
|
&hw_counter,
|
|
)
|
|
.unwrap();
|
|
segment_2
|
|
.upsert_point(
|
|
1,
|
|
idx.into(),
|
|
only_default_vector(&[0., 0., 0., 0.]),
|
|
&hw_counter,
|
|
)
|
|
.unwrap();
|
|
}
|
|
|
|
// Get normal build time
|
|
let (time_baseline, was_cancelled_baseline) = estimate_build_time(&baseline_segment, None);
|
|
assert!(!was_cancelled_baseline);
|
|
eprintln!("baseline time: {time_baseline}");
|
|
|
|
// Checks that optimization with longer cancellation delay will also finish fast
|
|
let early_stop_delay = time_baseline / 20;
|
|
let (time_fast, was_cancelled_early) = estimate_build_time(&segment, Some(early_stop_delay));
|
|
let late_stop_delay = time_baseline / 5;
|
|
let (time_long, was_cancelled_later) = estimate_build_time(&segment_2, Some(late_stop_delay));
|
|
|
|
// Timing on CI (especially Windows) can be noisy due to scheduler delays.
|
|
// Keep a fixed lower bound but scale tolerance for slower baseline runs.
|
|
let acceptable_stopping_delay = std::cmp::max(600, time_baseline / 8); // millis
|
|
|
|
assert!(was_cancelled_early);
|
|
assert!(
|
|
time_fast < early_stop_delay + acceptable_stopping_delay,
|
|
"time_early: {time_fast}, early_stop_delay: {early_stop_delay}"
|
|
);
|
|
|
|
assert!(was_cancelled_later);
|
|
assert!(
|
|
time_long < late_stop_delay + acceptable_stopping_delay,
|
|
"time_later: {time_long}, late_stop_delay: {late_stop_delay}"
|
|
);
|
|
assert!(
|
|
time_long < time_baseline,
|
|
"cancelled build should be faster than baseline: time_later={time_long}, baseline={time_baseline}",
|
|
);
|
|
|
|
assert!(
|
|
time_fast < time_long,
|
|
"time_early: {time_fast}, time_later: {time_long}, was_cancelled_later: {was_cancelled_later}",
|
|
);
|
|
}
|
|
|
|
#[test]
|
|
fn test_building_new_segment_with_mmap_payload() {
|
|
let segment_dir = Builder::new().prefix("segment_dir").tempdir().unwrap();
|
|
let temp_dir = Builder::new().prefix("segment_temp_dir").tempdir().unwrap();
|
|
|
|
let mut segment1 = build_simple_segment_with_payload_storage(
|
|
segment_dir.path(),
|
|
4,
|
|
Distance::Dot,
|
|
PayloadStorageType::Mmap,
|
|
)
|
|
.unwrap();
|
|
|
|
assert_eq!(
|
|
segment1.segment_config.payload_storage_type,
|
|
PayloadStorageType::Mmap
|
|
);
|
|
|
|
let hw_counter = HardwareCounterCell::new();
|
|
|
|
// add one point
|
|
segment1
|
|
.upsert_point(
|
|
1,
|
|
1.into(),
|
|
only_default_vector(&[1.0, 0.0, 1.0, 1.0]),
|
|
&hw_counter,
|
|
)
|
|
.unwrap();
|
|
|
|
let builder = SegmentBuilder::new(
|
|
temp_dir.path(),
|
|
&segment1.segment_config,
|
|
&HnswGlobalConfig::default(),
|
|
)
|
|
.unwrap();
|
|
|
|
let temp_segment_count = fs::read_dir(temp_dir.path()).unwrap().count();
|
|
|
|
assert_eq!(temp_segment_count, 1);
|
|
|
|
// Now we finalize building
|
|
let new_segment = builder.build_for_test(segment_dir.path());
|
|
assert_eq!(
|
|
new_segment.segment_config.payload_storage_type,
|
|
PayloadStorageType::Mmap
|
|
);
|
|
|
|
let new_segment_count = fs::read_dir(segment_dir.path()).unwrap().count();
|
|
|
|
assert_eq!(new_segment_count, 2);
|
|
}
|