Re: log_destination=file

On Thu, Sep 7, 2017 at 7:02 AM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:

> Dmitry Dolgov <9erthalion6(at)gmail(dot)com> writes:
> >> On 31 August 2017 at 14:49, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> >> pgbench with log_statement = all would be a pretty easy test case.
>
> > It seems that for this particular workload it was about 20-25% slower.
>
> Ouch. That seems like rather a large hit :-(. I actually expected
> it to be close to negligible, but I don't think 20% qualifies.
>

Agreed, that's a lot more than I expected. A few questions though:

1. where was stderr actually sent? To the console, to /dev/null or to a
file?

2. Could you run the same thing (on the same machine) with the logging
collector on and logging to file, without the patch? I'd assume that gives
performance similar to running with the patch, but it would be good to
confirm that.

And thanks for running the benchmark, saved me some time!

--
Magnus Hagander
Me: https://www.hagander.net/ <http://www.hagander.net/>
Work: https://www.redpill-linpro.com/ <http://www.redpill-linpro.com/>