Coming back to my problem, today happened again, that search comes slow.
Explaining again, auto vacuum and auto analyse for pg_* tables are not configured one by one, so its using that defaults of scale_factor = 0.2 and threshold = 50.
Today, when that problem of speed came back, the view pg_stat_sys_tables for pg_class table shows that it ran 415 times autovacuum only this week and there are 75187 updates on this table.
How can I have been updated 75000 times this table only in a week ? What kind of statement does an update on this table ? And why did it run 415 times autovacuum and only 1 autoanalyse ?
| relname | seq_scan | seq_tup_read | idx_scan | idx_tup_fetch | n_tup_ins | n_tup_upd | n_tup_del | n_tup_hot_upd | n_live_tup | n_dead_tup | n_mod_since_analyze | last_vacuum | last_autovacuum | last_analyze | last_autoanalyze | vacuum_count | autovacuum_count | analyze_count | autoanalyze_count |
|---|
| pg_class | 5395512 | 38598962913 | 616674711 | 2219978581 | 2243 | 75187 | 625 | 44008 | 169891 | 5645 | 7435 | NULL | 2019-08-22 19:58:48.497317 | NULL | 2019-08-22 13:03:02.770579 | 0 | 415 | 0 | 1 |
Sent from the
PostgreSQL - general mailing list archive at Nabble.com.
