I got a newly installed database (PG 9.4.6 / RHEL 7) which logfile grows rapidly.
Most entries are like
2016-03-03 10:33:29.259 CET [56868] user@db STATEMENT: SELECT field1 AS a1, field2 AS a2, field3 AS a3, field4 AS a4, field5 AS a5, field6 AS a6, MSISDN AS a7, field8 AS a8, RETRY_COUNT AS a9 FROM table WHERE ((field3 <= $1) AND (field8 IS NULL)) ORDER BY field3 ASC LIMIT $2 OFFSET $3
2016-03-03 10:33:29.259 CET [56868] user@db LOG: EXECUTE MESSAGE STATISTICS
2016-03-03 10:33:29.259 CET [56868] user@db DETAIL: ! system usage stats:
! 0.000006 elapsed 0.000000 user 0.000000 system sec
! [0.526408 user 0.112125 sys total]
! 0/0 [0/0] filesystem blocks in/out
! 0/0 [0/1281] page faults/reclaims, 0 [0] swaps
! 0 [0] signals rcvd, 0/0 [0/0] messages rcvd/sent
! 0/0 [3516/0] voluntary/involuntary context switches
I read through the documentation and found out, that it should be one of the log_*_stats parameters and despite the missing config entry in the postgresql.conf and postgresql.auto.conf
> egrep "log_.*stats" postgresql.*conf
postgresql.conf:#log_parser_stats = off
postgresql.conf:#log_planner_stats = off
postgresql.conf:#log_executor_stats = off
postgresql.conf:#log_statement_stats = off
the running instance shows that the parameter is set to “on”.
> echo "show log_statement_stats" | psql
log_statement_stats
---------------------
on
I tried to change it by applying an “ALTER SYSTEM”, but it doesn’t help. The parameter stays “on”.
I also reloaded the database (pg_ctl reload) and restarted it (its a pacemaker resource, so pcs resource restart postgre_db) – But nothing happens.
How do I turn off this setting?