| From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
|---|---|
| To: | Bruce Momjian <bruce(at)momjian(dot)us> |
| Cc: | Guillaume Smet <guillaume(dot)smet(at)gmail(dot)com>, pgsql-hackers(at)postgresql(dot)org |
| Subject: | Re: log_duration is redundant, no? |
| Date: | 2006-09-07 22:58:30 |
| Message-ID: | 489.1157669910@sss.pgh.pa.us |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
Bruce Momjian <bruce(at)momjian(dot)us> writes:
> Well, except for bind, all the log output display is zero cost, just a
> printf(), as I remember. The only cost that is significant, I think, is
> the timing of the query, and that is happening for all the setttings
> discussed.
On a machine with slow gettimeofday(), measuring duration at all is
going to hurt, but apparently that is not Guillaume's situation ---
what's costing him is sheer log volume. And remember that the
slow-gettimeofday problem exists mainly on cheap PCs, not server-grade
hardware. Based on his experience I'm prepared to believe that there
is a use-case for logging just the duration for short queries.
It seems like we should either remove the separate log_duration boolean
or make it work as he suggests. I'm leaning to the second answer now.
What's your vote?
regards, tom lane
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Guillaume Smet | 2006-09-07 23:04:48 | Re: log_duration is redundant, no? |
| Previous Message | Bruce Momjian | 2006-09-07 22:52:10 | Re: log_duration is redundant, no? |