Re: log_duration and \timing times repeatably much higher than "Total runtime" from explain analyze - Mailing list pgsql-general

From Greg Stark
Subject Re: log_duration and \timing times repeatably much higher than "Total runtime" from explain analyze
Date
Msg-id 877k3c6408.fsf@stark.dyndns.tv
Whole thread Raw
In response to Re: log_duration and \timing times repeatably much higher  (Bruce Momjian <pgman@candle.pha.pa.us>)
Responses Re: log_duration and \timing times repeatably much higher than "Total runtime" from explain analyze  (Martijn van Oosterhout <kleptog@svana.org>)
Re: log_duration and \timing times repeatably much higher than "Total runtime" from explain analyze  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-general
Bruce Momjian <pgman@candle.pha.pa.us> writes:

> I know \timing counts the time to transfer the data to the client, and I
> think log_duration also might have that timing added too.  It does seem
> like a long time to transfer data, though.

a) Even when executing this query normally there's very little data.
   About 1-20 records.

b) this is on an explain analyze which only returns the 37 records of the
   plan. I can't imagine that takes any real time.

This is on a machine that's otherwise pretty much idle. It's a dual processor
PII-900 and it used to perform fine. What's really strange is that some other
queries perform fine but this one and a few others reliably takes this long
and behaves this way under explain analyze. It's as if there's something
specific about this query that triggers the delay.

> >  Total runtime: 316.19 msec
> > (37 rows)
> >
> > Time: 1333.72 ms

--
greg

pgsql-general by date:

Previous
From: elein
Date:
Subject: Re: int1? types?
Next
From: Joe Conway
Date:
Subject: Re: int1? types?