Re: Adding a concept of TEMPORARY TABLESPACE for the use in temp_tablespaces

From: Mitar <mmitar(at)gmail(dot)com>
To: pgsql-hackers(at)lists(dot)postgresql(dot)org
Subject: Re: Adding a concept of TEMPORARY TABLESPACE for the use in temp_tablespaces
Date: 2019-01-06 19:27:36
Message-ID: CAKLmikMVafYSjkk=SRPxd9r=Q_N40QidZ9se15m5SB2CORnn7A@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Hi!

On Sun, Jan 6, 2019 at 11:01 AM Mitar <mmitar(at)gmail(dot)com> wrote:
> I have experimented a bit and performance really improves if /dev/shm
> is used. I have experimented with creating temporary tables inside a
> regular (SSD backed) tablespace /dev/shm and I have seen at least 2x
> improvement in time it takes for a set of modification+select queries
> to complete.

I also tried just to increase temp_buffers to half the memory, and
things are better, but not to the same degree as using a /dev/shm
tablespace. Why is that? (All my temporary objects in my experiments
are small, few 10k rows, few MBs.)

Mitar

--
http://mitar.tnode.com/
https://twitter.com/mitar_m

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Tom Lane 2019-01-06 19:29:05 Re: reducing the footprint of ScanKeyword (was Re: Large writable variables)
Previous Message Alexander Korotkov 2019-01-06 19:06:18 Re: [PATCH] kNN for btree