Re: Fwd: Query with high planning time compared to execution time - Mailing list pgsql-performance

From Richard Lee
Subject Re: Fwd: Query with high planning time compared to execution time
Date
Msg-id CALhvPTTfzo1xBirRFn+4NEAvGm9-VFBJF1SKvdHtyLM1NROmzA@mail.gmail.com
Whole thread Raw
In response to Re: Fwd: Query with high planning time compared to execution time  (Justin Pryzby <pryzby@telsasoft.com>)
List pgsql-performance
Hi,

Ah, apologize for the mistake. The entire will take several hours to complete and the problem query won't be executed until about halfway through the benchmark. Should I do `perf record` when the query appears? Or one `perf record` at the start of the test and another one when the query appears? I imagine doing a `perf record` of the entire benchmark will fill the storage (only about 100GB of space on the server).

Regards,
Richard Lee

On Mon, Nov 5, 2018 at 9:55 PM Justin Pryzby <pryzby@telsasoft.com> wrote:
On Mon, Nov 05, 2018 at 03:04:29PM +0800, Richard Lee wrote:
> Executed perf-record and perf-report:
> -bash-4.2$ perf record -g -- psql -U sri sri <
> /var/lib/pgsql/10/data/pg_log/1-b10/query.txt
> < ... snipped ... >

That's showing perf output for the psql client.  What you want is output for
the server process (essentially all the client does is move data between the
user to the server).

https://wiki.postgresql.org/wiki/Profiling_with_perf

Justin

pgsql-performance by date:

Previous
From: Justin Pryzby
Date:
Subject: Re: Fwd: Query with high planning time compared to execution time
Next
From: Lincoln Swaine-Moore
Date:
Subject: NOT IN vs. NOT EXISTS performance