Thread: BUG #1314: STATEMENT_TIMEOUT DOES NOT WORK PROPERLY

BUG #1314: STATEMENT_TIMEOUT DOES NOT WORK PROPERLY

From
"PostgreSQL Bugs List"
Date:
The following bug has been logged online:

Bug reference:      1314
Logged by:          Adnan DURSUN

Email address:      adursun@hvkk.mil.tr

PostgreSQL version: 8.0 Beta

Operating system:   W2K

Description:        STATEMENT_TIMEOUT DOES NOT WORK PROPERLY

Details:

Hi, i use PostgreSQL 8 Beta4.I write a sf like that :
/////////
CREATE OR REPLACE FUNCTION fn_test()
  RETURNS "varchar" AS
$BODY$
DECLARE
   c refcursor;
   r record;
BEGIN
   SET SESSION STATEMENT_TIMEOUT = 1000;
   OPEN C FOR SELECT * FROM T_KULLANICILAR FOR UPDATE;
   FETCH C INTO R;
   RESET STATEMENT_TIMEOUT;
   RETURN '1';
EXCEPTION
   WHEN QUERY_CANCELED THEN
          RAISE EXCEPTION 'UNABLE TO OBTAIN LOCK... :%';
END;
$BODY$
  LANGUAGE 'plpgsql' VOLATILE;
/////////
1.I RUN A THIS "SELECT * FROM T_KULLANICILAR FOR UPDATE" from another
session to lock T_KULLANICILAR table rows.
2.While T_KULLANICILAR table was locked,i called sf like "SELECT FN_TEST()"
from psql command prompt.But STATEMENT_TIMEOUT doesn't effect.
3.but i enter "SET SESSION STATEMENT_TIMEOUT = 1000;" command before step 2,
STATEMENT_TIMEOUT affects and cancel query.

Re: BUG #1314: STATEMENT_TIMEOUT DOES NOT WORK PROPERLY

From
Tom Lane
Date:
"PostgreSQL Bugs List" <pgsql-bugs@postgresql.org> writes:
> CREATE OR REPLACE FUNCTION fn_test()
>   RETURNS "varchar" AS
> $BODY$
> DECLARE
>    c refcursor;
>    r record;
> BEGIN
>    SET SESSION STATEMENT_TIMEOUT = 1000;
>    OPEN C FOR SELECT * FROM T_KULLANICILAR FOR UPDATE;
>    FETCH C INTO R;
>    RESET STATEMENT_TIMEOUT;
>    RETURN '1';
> EXCEPTION
>    WHEN QUERY_CANCELED THEN
>           RAISE EXCEPTION 'UNABLE TO OBTAIN LOCK... :%';
> END;
> $BODY$

statement_timeout limits the time taken for an *interactive* command,
not the time for an individual command inside a function.  What this
means is you can't change it around inside a function and expect
anything to happen --- the timeout, or lack of it, for the current
interactive command was determined long before your function got called.

            regards, tom lane