Re: pg_upgrade —link does it remove table bloat

On 2/13/20 12:45 PM, Jason Ralph wrote:
> @Adrian Klaver,
> I was concerned with the 1.4 value of tbloat and wastedbytes value, then again the last autovacuum was at 2020-02-13
02:25:22.533372-05and I took this snapshot at 3:44PMEST.  So it may be ok, what do you think?
 

What is your concern, storage space or performance of queries?

If performance then an EXPLAIN ANALYZE on a query will help show whether 
there is an issue or not.

> 
> 
> current_database | schemaname |                 tablename                 | tbloat | wastedbytes |
iname                     | ibloat | wastedibytes
 
> notimportant         | public              | members                                   |    1.4   |  3080314880  |
members_cobrid                     |    0.2 |            0
 
> 
> 
> 
> 
> 
> notimportant=# select * from pg_stat_user_tables where relname = 'members';
> -[ RECORD 1 ]-------+------------------------------
> relid               | 2045245
> schemaname          | public
> relname             | members
> seq_scan            | 55065
> seq_tup_read        | 201069350222
> idx_scan            | 5349501175
> idx_tup_fetch       | 7201402647
> n_tup_ins           | 910616
> n_tup_upd           | 46730942
> n_tup_del           | 1
> n_tup_hot_upd       | 41845682
> n_live_tup          | 18262438
> n_dead_tup          | 14740
> n_mod_since_analyze | 2476
> last_vacuum         | 2019-10-13 01:01:40.587534-04
> last_autovacuum     | 2020-02-13 02:25:22.533372-05
> last_analyze        | 2019-10-13 01:01:41.916929-04
> last_autoanalyze    | 2020-02-13 13:44:46.273096-05
> vacuum_count        | 15
> autovacuum_count    | 92
> analyze_count       | 15
> autoanalyze_count   | 243
> 
> Jason Ralph
> 


-- 
Adrian Klaver
adrian.klaver@aklaver.com