Skip to main content

run_flush

Function run_flush 

Source
pub async fn run_flush(engine: &Arc<MeruEngine>) -> Result<()>
Expand description

Run one flush job: takes the oldest immutable memtable and writes it to L0.

Serialized by engine.flush_mutex so that two concurrently-spawned auto-flush tasks don’t both observe the same oldest_immutable(), double-flush it to two L0 Parquet files with identical data, and double-commit competing Iceberg snapshots (Bug G regression).