Module flush_trigger 

Structs

RegionFlushTicker

RegionFlushTicker is used to trigger region flush trigger periodically.

RegionFlushTrigger

RegionFlushTrigger is used to ensure that the estimated WAL replay size stays below a certain threshold by triggering a region flush when the estimated WAL replay size exceeds that threshold. This helps improve datanode startup speed and reduce the overall startup time.

Enums

Event

Event represents various types of events that can be processed by the region flush ticker.

Constants

RECENT_DURATION 🔒

The duration of the recent period.

Functions

filter_regions_below_topic_pruned_entry_id 🔒

Filters regions whose prunable entry id is behind the topic pruned entry id.

filter_regions_by_replay_size 🔒

Filter regions based on the estimated replay size.

filter_regions_for_periodic_checkpoint 🔒

Filters regions that need periodic checkpoint persistence.

group_regions_by_leader 🔒

Group regions by leader.

is_recent 🔒

Check if the timestamp is recent.

mark_checkpoint_persisted 🔒

Marks checkpoint persistence timestamps for regions.

merge_region_ids 🔒

Merges two region id lists and removes duplicates.

retain_checkpoint_persist_records 🔒

Retains checkpoint persistence records for active regions.

should_persist_region_checkpoint 🔒

Determines whether a region checkpoint should be persisted based on current and persisted state.

Type Aliases

RegionFlushTickerRef 🔒