Hubert, :-) Your script was one of the first that I found, thanks to the power of Google.
My issue with your script is that, for one reason or another, when piping the logs through rsyslog, we end up with log lines that your perl code won't help, like: Jul 12 04:02:10 artemis postgres[27803]: [53-1] 2010-07-12 04:02:10.980 CDT f...@scheduler 27803 192.168.1.10(50304)LOG: statement: BEGIN;SET statement_timeout=30000;COMMIT;SELECT last_value, slots, used, ROUND(used/slots*100) AS percent, CASE WHEN slots < used THEN 0 ELSE slots - used END AS numleft FROM (SELECT last_value, CEIL((LEAST(max_value, 2147483647)-min_value::numeric +1)/increment_by::NUMERIC) AS slots, CEIL((last_value-min_value::numeric +1)/increment_by::NUMERIC) AS used FROM public.minor_preference_id_seq) foo Jul 12 04:02:11 artemis postgres[28022]: [24-1] 2010-07-12 04:02:11.167 CDT f...@scheduler 28022 192.168.1.10(47322)LOG: duration: 273.333 ms Jul 12 04:02:11 artemis postgres[27487]: [13097-1] 2010-07-12 04:02:11.337 CDT f...@scheduler 27487 192.168.1.10(39055)LOG: duration: 376.298 ms Jul 12 04:02:11 artemis postgres[27803]: [54-1] 2010-07-12 04:02:11.456 CDT f...@scheduler 27803 192.168.1.10(50304)LOG: duration: 476.522 ms IE, the duration ends up on a different line, and basically none of the statements ever match in your perl script. Any guess here? -- Anthony On Wed, 2010-07-14 at 14:28 +0200, hubert depesz lubaczewski wrote: > On Tue, Jul 13, 2010 at 02:53:25PM -0500, Anthony Presley wrote: > > I'm bordering on insanity, trying to track down an IDLE in transaction > > problem. > > you might find this helpful: > http://www.depesz.com/index.php/2008/08/28/hunting-idle-in-transactions/ > > Pozdrawiam, > > Hubert Lubaczewski > -- Sent via pgsql-general mailing list (pgsql-general@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
