Re: Parallel Seq Scan vs kernel read ahead - Mailing list pgsql-hackers

From Soumyadeep Chakraborty
Subject Re: Parallel Seq Scan vs kernel read ahead
Date
Msg-id CAE-ML+_ibpDEe6eGbp597GeMH-Z_658HW6WmDVH=XjzASQoH7Q@mail.gmail.com
Whole thread Raw
In response to Re: Parallel Seq Scan vs kernel read ahead  (Thomas Munro <thomas.munro@gmail.com>)
List pgsql-hackers
On Tue, Jul 21, 2020 at 9:33 PM Thomas Munro <thomas.munro@gmail.com> wrote:
>
> On Wed, Jul 22, 2020 at 3:57 PM Amit Kapila <amit.kapila16@gmail.com> wrote:
> > Yeah, that is true but every time before the test the same amount of
> > data should be present in shared buffers (or OS cache) if any which
> > will help in getting consistent results.  However, it is fine to
> > reboot the machine as well if that is a convenient way.
>
> We really should have an extension (pg_prewarm?) that knows how to
> kick stuff out PostgreSQL's shared buffers and the page cache
> (POSIX_FADV_DONTNEED).
>
>
+1. Clearing the OS page cache on FreeBSD is non-trivial during testing.
You can't do this on FreeBSD:
sync; echo 3 > /proc/sys/vm/drop_caches

Also, it would be nice to evict only those pages from the OS page cache
that are Postgres pages instead of having to drop everything.

Regards,
Soumyadeep (VMware)



pgsql-hackers by date:

Previous
From: Soumyadeep Chakraborty
Date:
Subject: Re: Parallel Seq Scan vs kernel read ahead
Next
From: Floris Van Nee
Date:
Subject: RE: [PATCH] Keeps tracking the uniqueness with UniqueKey