13.07.2018 00:00, Peter Geoghegan пишет:
One area that might be worth investigating is retail index tuple deletion performed within the executor in the event of non-HOT updates.
I will try to use this idea. But now I developed a background worker already (experimental state). It clean an index and heap relations with a retail deletion approach. It uses "soft" strategy with conditional buffer locks. Benchmarks on pgbench test demonstrate same execution time as PostgreSQL without the worker, but 90% of dead tuples in index and heap relations cleaned without vacuum.
My main efforts is involved to tuning of the worker. --- Andrey Lepikhov Postgres Professional https://postgrespro.com The Russian Postgres Company
