pg_upgrade test for binary compatibility of core data types

From: Justin Pryzby <pryzby(at)telsasoft(dot)com>
To: Tomas Vondra <tomas(dot)vondra(at)2ndquadrant(dot)com>
Cc: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, Andres Freund <andres(at)anarazel(dot)de>, buschmann(at)nidsa(dot)net, pgsql-hackers(at)lists(dot)postgresql(dot)org
Subject: pg_upgrade test for binary compatibility of core data types
Date: 2020-12-06 18:02:48
Message-ID: 20201206180248.GI24052@telsasoft.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-bugs pgsql-hackers

I'm finally returning to this 14 month old thread:
(was: Re: BUG #16045: vacuum_db crash and illegal memory alloc after pg_upgrade from PG11 to PG12)

On Tue, Oct 15, 2019 at 09:07:25AM +0200, Tomas Vondra wrote:
> On Mon, Oct 14, 2019 at 11:41:18PM -0500, Justin Pryzby wrote:
> >
> > Perhaps it'd be worth creating a test for on-disk format ?
> >
> > Like a table with a column for each core type, which is either SELECTed from
> > after pg_upgrade, or pg_dump output compared before and after.
>
> IMO that would be useful - we now have a couple of these checks for
> different data types (line, unknown, sql_identifier), with a couple of
> combinations each. And I've been looking if we do similar pg_upgrade
> tests, but I haven't found anything. I mean, we do pg_upgrade the
> cluster used for regression tests, but here we need to test a number of
> cases that are meant to abort the pg_upgrade. So we'd need a number of
> pg_upgrade runs, to test that.

I meant to notice if the binary format is accidentally changed again, which was
what happened here:
7c15cef86 Base information_schema.sql_identifier domain on name, not varchar.

I added a table to the regression tests so it's processed by pg_upgrade tests,
run like:

| time make -C src/bin/pg_upgrade check oldsrc=`pwd`/11 oldbindir=`pwd`/11/tmp_install/usr/local/pgsql/bin

I checked that if I cherry-pick 0002 to v11, and comment out
old_11_check_for_sql_identifier_data_type_usage(), then pg_upgrade/test.sh
detects the original problem:
pg_dump: error: Error message from server: ERROR: invalid memory alloc request size 18446744073709551613

I understand the buildfarm has its own cross-version-upgrade test, which I
think would catch this on its own.

These all seem to complicate use of pg_upgrade/test.sh, so 0001 is needed to
allow testing upgrade from older releases.

e78900afd217fa3eaa77c51e23a94c1466af421c Create by default sql/ and expected/ for output directory in pg_regress
40b132c1afbb4b1494aa8e48cc35ec98d2b90777 In the pg_upgrade test suite, don't write to src/test/regress.
fc49e24fa69a15efacd5b8958115ed9c43c48f9a Make WAL segment size configurable at initdb time.
c37b3d08ca6873f9d4eaf24c72a90a550970cbb8 Allow group access on PGDATA
da9b580d89903fee871cf54845ffa2b26bda2e11 Refactor dir/file permissions

--
Justin

Attachment Content-Type Size
0001-WIP-pg_upgrade-test.sh-changes-needed-to-allow-testi.patch text/x-diff 4.5 KB
0002-pg_upgrade-test-to-exercise-binary-compatibility.patch text/x-diff 6.1 KB

In response to

Responses

Browse pgsql-bugs by date

  From Date Subject
Next Message Shinoda, Noriyoshi (PN Japan FSIP) 2020-12-07 08:58:18 RE: BUG #16754: When using LLVM and parallel queries aborted all session by pg_cancel_backend.
Previous Message PG Bug reporting form 2020-12-06 10:12:16 BUG #16764: wal sender exits abnormally which cause wal receiever exits twice

Browse pgsql-hackers by date

  From Date Subject
Next Message Konstantin Knizhnik 2020-12-06 20:50:51 Wrong check in pg_visibility?
Previous Message Andrew Dunstan 2020-12-06 17:16:31 Re: Change definitions of bitmap flags to bit-shifting style