R: [GENERAL] (solved) R: Very strange problem on index - Mailing list pgsql-general

Hi Raymond,

Thank nyou for your appreciated feedback.

Here is the original message:

i really strange problem, quite near to paranormal, is occurring during a server migration.

We have a table with some millions of record, perfectly working on other Postgresql 9.6.1 machines:

                                 Table "public.webrecord"
  Column   |         Type         |                       Modifiers
-----------+----------------------+-------------------------------------
-----------+----------------------+-------------------
 id        | integer              | not null default nextval('webrecord_id_seq'::regclass)
 categoria | character varying    |
 dominio   | character varying    |
 country   | character varying(5) |
Indexes:
    "keywebrecord" PRIMARY KEY, btree (id)
    "dominio_idx" btree (dominio)

By entering a simple query:

Select * from webrecord where dominio='sito.com' it wait some seconds and cpu raise up.

I tried to:
- delete the tabel
- recreate the schema again
- re-populating it
- reindexing
- vacuum

But nothing happened.
The query planne says it scans lots of rows:

explain analyze select * from webrecord where dominio='crl.microsoft.com';
                                                            QUERY PLAN

----------------------------------------------------------------------------------------------------------------------------------
 Gather  (cost=1000.00..144045.21 rows=62073 width=92) (actual time=1096.202..1096.202 rows=0 loops=1)
   Workers Planned: 4
   Workers Launched: 4
   ->  Parallel Seq Scan on webrecord  (cost=0.00..136837.91 rows=15518 width=92) (actual time=1063.277..1063.277
rows=0loops=5) 
         Filter: ((dominio)::text = 'crl.microsoft.com'::text)
         Rows Removed by Filter: 2482938  Planning time: 0.119 ms  Execution time: 1107.846 ms

On other machine the results are immediate!

I have finished my tries...