On Tue, Jan 31, 2023 at 8:07 AM Dimitrios Apostolou <jimis(at)gmx(dot)net> wrote:
> -> Seq Scan on public.test_runs_raw (cost=0.00..9250235.80
> rows=317603680 width=42) (actual time=745910.672..745910.677 rows=10
> loops=1)
> Output: run_n, test_name_n, workitem_n, started_on, duration_ms,
> test_result_n, test_executable_n, test_function_n, test_datatag_n
> Buffers: shared read=2334526
> I/O Timings: shared/local read=691137.029
>
The system has to return 10 live rows to you. If it needs to search
through that many buffers to find 10 live rows you most likely have a large
bloating problem going on. Seems like it is time to vacuum full.
David J.