Slow Bitmap Index Scan

Поиск
Список
Период
Сортировка
От Scott Rankin
Тема Slow Bitmap Index Scan
Дата
Msg-id 05111C93-69B4-4D5B-8AD3-8C9C47E0F01E@motus.com
обсуждение исходный текст
Ответы Re: Slow Bitmap Index Scan
Список pgsql-performance

Hello all,

 

We recently moved our production database systems from a 9.4 running on a self-managed EC2 instance to 9.6.10 on Amazon’s AWS (same RAM, CPU).  After the move, we’re finding that certain queries that we run against a GIN full-text index have some occasionally very slow executions and I’m struggling to figure out what to do about it.   I would be very grateful for any ideas!

 

Thanks,

Scott

 

The setup we have is a 32-core, 244 GB RAM primary with a same-sized read replica.  The queries are running off the replica, but performance is roughly the same between the master and the replica.

 

Here’s a query that’s performing badly:

 

SELECT ls.location AS locationId  FROM location_search ls  WHERE ls.client = 83  AND search_field_tsvector @@ to_tsquery('9000:* &Smith''s:* &Mill:*')  AND ls.favorite = TRUE  LIMIT 1

 

The mean time for this query (and others like it) is about 900ms, but the std deviation is over 1000ms and the max is 11000ms.

 

The explain looks like this:

 

Limit  (cost=1516.25..1520.52 rows=1 width=223) (actual time=4506.482..4506.482 rows=0 loops=1)

  Buffers: shared hit=9073

  ->  Bitmap Heap Scan on location_search ls  (cost=1516.25..1520.52 rows=1 width=223) (actual time=4506.480..4506.480 rows=0 loops=1)

        Recheck Cond: (search_field_tsvector @@ to_tsquery('9000:* &Smith''s:* &Mill:*'::text))

        Filter: (favorite AND (client = 83))

        Rows Removed by Filter: 8

        Heap Blocks: exact=12

        Buffers: shared hit=9073

        ->  Bitmap Index Scan on location_search_tsvector_idx  (cost=0.00..1516.25 rows=1 width=0) (actual time=4506.450..4506.450 rows=12 loops=1)

              Index Cond: (search_field_tsvector @@ to_tsquery('9000:* &Smith''s:* &Mill:*'::text))

              Buffers: shared hit=9061

Planning time: 0.240 ms

Execution time: 4509.995 ms

 

The table has about 30 million rows in it.   The table and index definition are:

 

CREATE TABLE public.location_search

(

    id bigint NOT NULL DEFAULT nextval('location_search_id_seq'::regclass),

    person_location bigint,

    person bigint,

    client_location bigint,

    client bigint,

    location bigint,

    org_unit_id bigint,

    latitude numeric(10,7),

    longitude numeric(10,7),

    geofence numeric(10,7),

    address_line_one text COLLATE pg_catalog."default",

    address_line_two text COLLATE pg_catalog."default",

    city text COLLATE pg_catalog."default",

    state text COLLATE pg_catalog."default",

    postal_code text COLLATE pg_catalog."default",

    country text COLLATE pg_catalog."default",

    full_address text COLLATE pg_catalog."default",

    is_google_verified boolean,

    address_source text COLLATE pg_catalog."default",

    active boolean,

    name character varying(255) COLLATE pg_catalog."default",

    external_client_location_id character varying(500) COLLATE pg_catalog."default",

    custom_field_values hstore,

    location_tags hstore,

    legacy_location_id bigint,

    favorite boolean,

    search_field_tsvector tsvector

)

WITH (

    OIDS = FALSE

)

TABLESPACE pg_default;

 

CREATE INDEX location_search_tsvector_idx

    ON public.location_search USING gin

    (search_field_tsvector)

    TABLESPACE pg_default;

 

Right now the output of pgstatginindex is this:

version pending_pages pending_tuples

2              214         9983

 

Lastly, here are some of the relevant config entries:

 

autovacuum on

autovacuum_analyze_scale_factor 0

autovacuum_analyze_threshold 50

autovacuum_freeze_max_age 400000000

autovacuum_max_workers 3

autovacuum_multixact_freeze_max_age 400000000

autovacuum_naptime 30s

autovacuum_vacuum_cost_delay 20ms

autovacuum_vacuum_cost_limit    -1

autovacuum_vacuum_scale_factor 0

autovacuum_vacuum_threshold 50

cpu_index_tuple_cost 0

cpu_operator_cost 0

cpu_tuple_cost 0

cursor_tuple_fraction 0

effective_cache_size 125777784kB

effective_io_concurrency 1

gin_fuzzy_search_limit 0

gin_pending_list_limit 4MB

maintenance_work_mem 4027MB

seq_page_cost 1

shared_buffers 62888888kB

work_mem 200000kB

 

 

SCOTT RANKIN
VP, Technology

Motus, LLC
Two Financial Center, 60 South Street, Boston, MA 02111 
617.467.1900 (O) | 
srankin@motus.com

 

Follow us on LinkedIn | Visit us at motus.com  

 

This email message contains information that Motus, LLC considers confidential and/or proprietary, or may later designate as confidential and proprietary. It is intended only for use of the individual or entity named above and should not be forwarded to any other persons or entities without the express consent of Motus, LLC, nor should it be used for any purpose other than in the course of any potential or actual business relationship with Motus, LLC. If the reader of this message is not the intended recipient, or the employee or agent responsible to deliver it to the intended recipient, you are hereby notified that any dissemination, distribution, or copying of this communication is strictly prohibited. If you have received this communication in error, please notify sender immediately and destroy the original message.

Internal Revenue Service regulations require that certain types of written advice include a disclaimer. To the extent the preceding message contains advice relating to a Federal tax issue, unless expressly stated otherwise the advice is not intended or written to be used, and it cannot be used by the recipient or any other taxpayer, for the purpose of avoiding Federal tax penalties, and was not written to support the promotion or marketing of any transaction or matter discussed herein.

В списке pgsql-performance по дате отправления:

Предыдущее
От: Justin Pryzby
Дата:
Сообщение: Re: Query with high planning time at version 11.1 compared versions10.5 and 11.0
Следующее
От: Justin Pryzby
Дата:
Сообщение: Re: Slow Bitmap Index Scan