snowykami/meilisearch
1
0
Fork 0
mirror of https://github.com/meilisearch/meilisearch.git synced 2025-09-06 04:36:32 +00:00
Code Issues Packages Projects Releases Wiki Activity

Optim L: Delete datetimes by range

Browse Source
This commit is contained in:
Mubelotix
2025-08-12 09:33:38 +02:00
parent 62ad7b345a
commit 60a1188a58
1 changed files with 46 additions and 44 deletions
Download Patch File Download Diff File Expand all files Collapse all files

90
crates/index-scheduler/src/scheduler/process_batch.rs
View File

@ -548,19 +548,40 @@ impl IndexScheduler {
let max = to_remove.keys().max().cloned().unwrap_or(i128::MIN); let max = to_remove.keys().max().cloned().unwrap_or(i128::MIN);
let range = min..=max; let range = min..=max;
let mut iter = db.rev_range_mut(wtxn, &range)?; // We iterate over the time database to see which ranges of timestamps need to be removed
while let Some(i) = iter.next() { let lazy_db = db.lazily_decode_data();
let (timestamp, mut tasks) = i?; let iter = lazy_db.range(wtxn, &range)?;
if let Some(to_remove_tasks) = to_remove.remove(&timestamp) { let mut delete_range_start = None;
tasks -= &to_remove_tasks; let mut delete_ranges = Vec::new();
if tasks.is_empty() { let mut to_put: HashMap<i128, RoaringBitmap> = HashMap::new();
// safety: We don't keep references to the database for i in iter {
unsafe { iter.del_current()? }; let (timestamp, data) = i?;
if let Some(to_remove) = to_remove.remove(&timestamp) {
let mut current = data.decode().unwrap(); // TODO
current -= &to_remove;
if current.is_empty() {
delete_range_start = Some(timestamp);
} else { } else {
// safety: We don't keep references to the database // We could close the deletion range but it's not necessary because the new value will get reinserted anyway
unsafe { iter.put_current(&timestamp, &tasks)? }; to_put.insert(timestamp, current);
}
} else if let Some(delete_range_start) = delete_range_start.take() {
// Current one must not be deleted so we need to skip it
delete_ranges.push(delete_range_start..timestamp);
} }
} }
if let Some(delete_range_start) = delete_range_start.take() {
delete_ranges.push(delete_range_start..i128::MAX);
}
for range in delete_ranges {
db.delete_range(wtxn, &range)?;
}
for (timestamp, data) in to_put {
db.put(wtxn, &timestamp, &data)?;
} }
Ok(()) Ok(())
@ -602,11 +623,17 @@ impl IndexScheduler {
tasks_enqueued_to_remove.entry(enqueued_at).or_default().insert(task_id); tasks_enqueued_to_remove.entry(enqueued_at).or_default().insert(task_id);
if let Some(started_at) = task.started_at { if let Some(started_at) = task.started_at {
tasks_started_to_remove.entry(started_at.unix_timestamp_nanos()).or_default().insert(task_id); tasks_started_to_remove
.entry(started_at.unix_timestamp_nanos())
.or_default()
.insert(task_id);
} }
if let Some(finished_at) = task.finished_at { if let Some(finished_at) = task.finished_at {
tasks_finished_to_remove.entry(finished_at.unix_timestamp_nanos()).or_default().insert(task_id); tasks_finished_to_remove
.entry(finished_at.unix_timestamp_nanos())
.or_default()
.insert(task_id);
} }
if let Some(canceled_by) = task.canceled_by { if let Some(canceled_by) = task.canceled_by {
@ -702,8 +729,7 @@ impl IndexScheduler {
// 5. Read batches metadata // 5. Read batches metadata
progress.update_progress(TaskDeletionProgress::RetrievingBatches); progress.update_progress(TaskDeletionProgress::RetrievingBatches);
let (atomic_progress, task_progress) = let (atomic_progress, task_progress) = AtomicBatchStep::new(to_delete_batches.len() as u32);
AtomicBatchStep::new(to_delete_batches.len() as u32);
progress.update_progress(task_progress); progress.update_progress(task_progress);
for range in consecutive_ranges(to_delete_batches.iter()) { for range in consecutive_ranges(to_delete_batches.iter()) {
let iter = self.queue.batches.all_batches.range(&rtxn, &range)?; let iter = self.queue.batches.all_batches.range(&rtxn, &range)?;
@ -742,23 +768,11 @@ impl IndexScheduler {
// 5. Remove task datetimes // 5. Remove task datetimes
progress.update_progress(TaskDeletionProgress::DeletingTasksDateTime); progress.update_progress(TaskDeletionProgress::DeletingTasksDateTime);
remove_datetimes( remove_datetimes(wtxn, tasks_enqueued_to_remove, self.queue.tasks.enqueued_at)?;
wtxn,
tasks_enqueued_to_remove,
self.queue.tasks.enqueued_at,
)?;
remove_datetimes( remove_datetimes(wtxn, tasks_started_to_remove, self.queue.tasks.started_at)?;
wtxn,
tasks_started_to_remove,
self.queue.tasks.started_at,
)?;
remove_datetimes( remove_datetimes(wtxn, tasks_finished_to_remove, self.queue.tasks.finished_at)?;
wtxn,
tasks_finished_to_remove,
self.queue.tasks.finished_at,
)?;
// 6. Delete batches datetimes // 6. Delete batches datetimes
progress.update_progress(TaskDeletionProgress::DeletingBatchesDateTime); progress.update_progress(TaskDeletionProgress::DeletingBatchesDateTime);
@ -773,23 +787,11 @@ impl IndexScheduler {
)?; )?;
} }
remove_datetimes( remove_datetimes(wtxn, batches_enqueued_to_remove, self.queue.batches.enqueued_at)?;
wtxn,
batches_enqueued_to_remove,
self.queue.batches.enqueued_at,
)?;
remove_datetimes( remove_datetimes(wtxn, batches_started_to_remove, self.queue.batches.started_at)?;
wtxn,
batches_started_to_remove,
self.queue.batches.started_at,
)?;
remove_datetimes( remove_datetimes(wtxn, batches_finished_to_remove, self.queue.batches.finished_at)?;
wtxn,
batches_finished_to_remove,
self.queue.batches.finished_at,
)?;
// 7. Remove tasks from indexes, statuses, and kinds // 7. Remove tasks from indexes, statuses, and kinds
progress.update_progress(TaskDeletionProgress::DeletingTasksMetadata); progress.update_progress(TaskDeletionProgress::DeletingTasksMetadata);