Discussion:
pg advisory locks
Arun Gokule
2014-10-12 16:41:27 UTC
Permalink
Hi,

I am executing pg_advisory_locks using the following set of statements:

SELECT pg_advisory_lock(317,2);
UPDATE posts SET dislikers = array_remove(dislikers, 7) WHERE id = 317;
update posts set num_dislikes = icount(dislikers), updated_at = now()
where id = 317;
WITH update_likers AS (SELECT pg_advisory_unlock(317,2)) select num_likes,
num_dislikes, (7 IN (select(unnest(likers)))) as liked, (7 IN
(select(unnest(dislikers)))) as disliked from posts where id = 317 LIMIT 1;

These are issued from a multithreaded app. One in 1000 queries, I get a
deadlock after the execution of the above set of statements. i.e. SELECT
try_pg_advisory_lock(317,2) returns false. Is there something obvious that
I am doing wrong?

Thanks,
Arun
Arun Gokule
2014-10-13 02:57:01 UTC
Permalink
Also I noticed that executing the following query:

SELECT *
FROM
(
SELECT DISTINCT age(now(), query_start) AS age,
pg_stat_activity.pid,pg_locks.granted,pg_stat_activity.application_name,pg_stat_activity.backend_start,
pg_stat_activity.xact_start, pg_stat_activity.state_change,
pg_stat_activity.waiting, pg_stat_activity.state,
pg_stat_activity.query_start, left(pg_stat_activity.query, 60)
FROM pg_stat_activity, pg_locks
WHERE pg_locks.pid = pg_stat_activity.pid
) AS foo
WHERE age > '30 seconds'
ORDER BY age DESC;


I get 2 entries for the pid 78573 with one as granted = true and the other
as granted =false but they have the same timestamps. See below:

age | pid | granted | application_name |
backend_start | xact_start |
state_change | waiting | state | query_start |
left
-----------------+-------+---------+------------------+-------------------------------+-------------------------------+-------------------------------+---------+--------+-------------------------------+--------------------------------------------------------------
10:53:25.394862 | 78261 | t | bin/rails | 2014-10-12
05:20:04.543242-07 | | 2014-10-12
08:57:22.950505-07 | f | idle | 2014-10-12 08:57:22.949693-07 |
SELECT pg_advisory_lock(317,2);UPDATE posts SET dislikers =
10:53:23.50218 | 78573 | t | bin/rails | 2014-10-12
05:37:58.120879-07 | 2014-10-12 08:57:24.842375-07 | 2014-10-12
08:57:24.842379-07 | t | active | 2014-10-12 08:57:24.842375-07 |
SELECT pg_advisory_lock(317,2);UPDATE posts SET likers = arr
10:53:23.50218 | 78573 | f | bin/rails | 2014-10-12
05:37:58.120879-07 | 2014-10-12 08:57:24.842375-07 | 2014-10-12
08:57:24.842379-07 | t | active | 2014-10-12 08:57:24.842375-07 |
SELECT pg_advisory_lock(317,2);UPDATE posts SET likers = arr
00:57:53.495221 | 90360 | t | bin/rails | 2014-10-12
18:43:20.09202-07 | | 2014-10-12
18:52:54.849397-07 | f | idle | 2014-10-12 18:52:54.849334-07 |
SELECT "posts"."id" AS t0_r0, "posts"."content" AS t0_r1, "p
Post by Arun Gokule
Hi,
SELECT pg_advisory_lock(317,2);
UPDATE posts SET dislikers = array_remove(dislikers, 7) WHERE id = 317;
update posts set num_dislikes = icount(dislikers), updated_at = now()
where id = 317;
WITH update_likers AS (SELECT pg_advisory_unlock(317,2)) select
num_likes, num_dislikes, (7 IN (select(unnest(likers)))) as liked, (7 IN
(select(unnest(dislikers)))) as disliked from posts where id = 317 LIMIT 1;
These are issued from a multithreaded app. One in 1000 queries, I get a
deadlock after the execution of the above set of statements. i.e. SELECT
try_pg_advisory_lock(317,2) returns false. Is there something obvious that
I am doing wrong?
Thanks,
Arun
--
Thanks,
Arun
Alban Hertroys
2014-10-13 08:05:40 UTC
Permalink
Post by Arun Gokule
Hi,
SELECT pg_advisory_lock(317,2);
UPDATE posts SET dislikers = array_remove(dislikers, 7) WHERE id = 317;
update posts set num_dislikes = icount(dislikers), updated_at = now() where id = 317;
WITH update_likers AS (SELECT pg_advisory_unlock(317,2)) select num_likes, num_dislikes, (7 IN (select(unnest(likers)))) as liked, (7 IN (select(unnest(dislikers)))) as disliked from posts where id = 317 LIMIT 1;
These are issued from a multithreaded app. One in 1000 queries, I get a deadlock after the execution of the above set of statements. i.e. SELECT try_pg_advisory_lock(317,2) returns false. Is there something obvious that I am doing wrong?
Your last query doesn’t call the unlock statement, that WITH section is going to be ignored as the query isn’t using it.

Alban Hertroys
--
If you can't see the forest for the trees,
cut the trees and you'll find there is no forest.
--
Sent via pgsql-general mailing list (pgsql-***@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general
Loading...