Re: _mdfd_getseg can be expensive

From: Peter Geoghegan <pg(at)heroku(dot)com>
To: Andres Freund <andres(at)anarazel(dot)de>
Cc: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, Kevin Grittner <kgrittn(at)gmail(dot)com>, Robert Haas <robertmhaas(at)gmail(dot)com>, Pg Hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: _mdfd_getseg can be expensive
Date: 2016-08-31 21:09:47
Message-ID: CAM3SWZRfC3idro-ViKYZ23EEcVbpxh3Onhm+7C-b_pcqSdqa_Q@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Thu, Aug 18, 2016 at 5:26 PM, Andres Freund <andres(at)anarazel(dot)de> wrote:
> Rebased version attached. A review would be welcome. Plan to push this
> forward otherwise in the not too far away future.

This looks good.

The only thing that stuck out to any degree is that we don't grow the
"reln->md_seg_fds[forknum]" array within the new _fdvec_resize()
function geometrically. In other words, we don't separately keep track
of the array size and allocated size, and grow the allocated size
using the doubling strategy that you see in many places.

Now, I freely admit that that probably doesn't matter, given what that
array tracks. I'm just pointing out that that aspect did give me
pause. The struct MdfdVec is small enough that that *might* be
worthwhile.

--
Peter Geoghegan

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Peter Geoghegan 2016-08-31 21:25:11 Re: _mdfd_getseg can be expensive
Previous Message Petr Jelinek 2016-08-31 20:51:34 Re: Logical Replication WIP