From: Tom Lane <tgl@sss.pgh.pa.us>
To: Robert Haas <robertmhaas@gmail.com>
cc: Michael Paquier <michael.paquier@gmail.com>,
        Amit Kapila <amit.kapila16@gmail.com>,
        Dilip Kumar <dilipbalaut@gmail.com>,
        Tomas Vondra <tomas.vondra@2ndquadrant.com>,
        Jim Nasby <Jim.Nasby@bluetreble.com>,
        Andres Freund <andres@anarazel.de>,
        pgsql-hackers <pgsql-hackers@postgresql.org>
Subject: Re: Speed up Clog Access by increasing CLOG buffers
In-reply-to: 
 <CA+TgmobrMF8ALx_7pGM+4G=i-o3NBf+FrB4bh6XHqUF7NuVgDA@mail.gmail.com>
References: 
 <CAA4eK1+8=X9mSNeVeHg_NqMsOR-XKsjuqrYzQf=iCsdh3U4EOA@mail.gmail.com>
 <acf46406-cf6d-41fe-4118-4e1c960b4790@2ndquadrant.com>
 <CAFiTN-tCqDucbikxue0x0dwNOAa6yzjMxJNgXGQm90jVVRSj0Q@mail.gmail.com>
 <CAFiTN-t15PjFTFQH0fBfM5jVSv6rfm5J_8vj=RGuKeFpdLgSoQ@mail.gmail.com>
 <CAA4eK1LyR2A+m=RBSZ6rcPEwJ=rVi1ADPSndXHZdjn56yqO6Vg@mail.gmail.com>
 <4a52a34f-57fa-7bcf-d34c-c15db40f0361@2ndquadrant.com>
 <CAA4eK1KTbNbZSDo=6k4YgaJh_FM20zJCKu2Yt0bxaFMv9QcSXQ@mail.gmail.com>
 <b3586234-6c80-5b64-1261-871e0e852bbb@2ndquadrant.com>
 <956b695d-f6c5-d652-bf0c-cea95981547c@BlueTreble.com>
 <f769dd83-6288-2c37-4958-b7ddad0bc974@2ndquadrant.com>
 <5275bf49-545e-e189-48c5-17b5defc45a2@2ndquadrant.com>
 <CA+TgmobmDLM1dzFvN_tJoBunarMpz+cehx1cVfv-QMKk4rZLPQ@mail.gmail.com>
 <ca714e28-81a2-440c-bd92-79fe1fa0fc15@2ndquadrant.com>
 <CAA4eK1JjatUZu0+HCi=5VM1q-hFgN_OhegPAwEUJqxf-7pESbg@mail.gmail.com>
 <04e39ac9-af28-0fda-8f72-7268197d281f@2ndquadrant.com>
 <84c22fbb-b9c4-a02f-384b-b4feb2c67193@2ndquadrant.com> !
 <CAA4eK1KAysCmBaYGbFXyu0wGTZuMWeNgYBHNd3goNrfD0-G2CA@mail.gmail.com>
 <CAFiTN-tTazysx8GvDFgNvCh17J9AbGM=XgD6PftkoWVCMUSsXQ@mail.gmail.com>
 <CAA4eK1J+67edo_Wnrfx8oJ+rWM_BAr+v6JqvQjKPdLOxR=0d5g@mail.gmail.com>
 <CAFiTN-sSp=UkeXezHnapuaLCNWmmAgzuYVEODbOwj5K2zbgWng@mail.gmail.com>
 <CAFiTN-tZbtsbAx1Wrsmqd7Z8weTOTpCt9sqikj3wXapVEcw-eA@mail.gmail.com>
 <CAA4eK1L9WHYTJY8KUq_+yzVMp1ep1UqB+b3x3Kb6hTVEUWascA@mail.gmail.com>
 <CAB7nPqTYT8oWXxmudwFbLxK7rJWU73LcJz3J+nCfVGvKE31NAQ@mail.gmail.com>
 <CA+TgmobgWHcXDcChX2+BqJDk2dkPVF85ZrJFhUyHHQmw8diTpA@mail.gmail.com>
 <9010.1489112253@sss.pgh.pa.us>
 <CA+TgmobrMF8ALx_7pGM+4G=i-o3NBf+FrB4bh6XHqUF7NuVgDA@mail.gmail.com>
Comments: In-reply-to Robert Haas <robertmhaas@gmail.com>
	message dated "Thu, 09 Mar 2017 21:35:50 -0500"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <20806.1489123260.1@sss.pgh.pa.us>
Content-Transfer-Encoding: quoted-printable
Date: Fri, 10 Mar 2017 00:21:00 -0500
Message-ID: <20807.1489123260@sss.pgh.pa.us>
Precedence: bulk
Sender: pgsql-hackers-owner@postgresql.org

Robert Haas <robertmhaas@gmail.com> writes:
> On Thu, Mar 9, 2017 at 9:17 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
>> Buildfarm thinks eight wasn't enough.
>> https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=3Dclam&dt=3D2017=
-03-10%2002%3A00%3A01

> At first I was confused how you knew that this was the fault of this
> patch, but this seems like a pretty indicator:
> TRAP: FailedAssertion("!(curval =3D=3D 0 || (curval =3D=3D 0x03 && status=
 !=3D
> 0x00) || curval =3D=3D status)", File: "clog.c", Line: 574)

Yeah, that's what led me to blame the clog-group-update patch.

> I'm not sure whether it's related to this problem or not, but now that
> I look at it, this (preexisting) comment looks like entirely wishful
> thinking:
>      * If we update more than one xid on this page while it is being writ=
ten
>      * out, we might find that some of the bits go to disk and others don=
't.
>      * If we are updating commits on the page with the top-level xid that
>      * could break atomicity, so we subcommit the subxids first before we=
 mark
>      * the top-level commit.

Maybe, but that comment dates to 2008 according to git, and clam has
been, er, happy as a clam up to now.  My money is on a newly-introduced
memory-access-ordering bug.

Also, I see clam reported in green just now, so it's not 100%
reproducible :-(

			regards, tom lane


--=20
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers