Re: Fail to create PK or index for large table in Windows - Mailing list pgsql-bugs

From Thomas Munro
Subject Re: Fail to create PK or index for large table in Windows
Date
Msg-id CAEepm=3hs3oXf2dWDh6pb9_7A0MATXCxK-KQx2ziWK05USuKag@mail.gmail.com
Whole thread Raw
In response to Re: Fail to create PK or index for large table in Windows  (Thomas Munro <thomas.munro@enterprisedb.com>)
List pgsql-bugs
On Tue, Nov 20, 2018 at 8:51 AM Thomas Munro
<thomas.munro@enterprisedb.com> wrote:
> Here's the REL_11_STABLE branch built for Windows as of a couple minutes ago:
>
> https://ci.appveyor.com/project/macdice/postgres/builds/20412803/artifacts

I received an off-list confirmation that a large CREATE INDEX could
complete normally on a database that previously showed the error,
using that snapshot of 11 stable.

I'm really pleased with this experiment: it seems there are plenty of
people willing to test on Windows if we can supply builds, in
particular people who've run into a bug.  This makes me think that I
should figure out how to get cfbot to spit out a builds for all
commitfest entries so that proposed fixes and features are
automatically available in a format usable by anyone with Windows and
no special developer tools.  The only problem is the space
requirement; if I just turn that on naively it'll eat 2-3TB of
appveyor's archive space and they might hate me, so I'll need to come
up with a way to push the zip files to external storage and keep just
the latest one, and make sure it's clear which master commit + patches
went into the build.

-- 
Thomas Munro
http://www.enterprisedb.com


pgsql-bugs by date:

Previous
From: Nurmurat Gurbanliyew
Date:
Subject: PostgreSQL v11 offline (portable) version libwinpthread-1.dll error
Next
From: PG Bug reporting form
Date:
Subject: BUG #15518: intarray index crashes hard