Обсуждение: pgsql: tests: Consistently use pg_basebackup -cfast --no-sync to accele

tests: Consistently use pg_basebackup -cfast --no-sync to accelerate tests.

Most tests invoking pg_basebackup themselves did not yet use -cfast, which
makes pg_basebackup take considerably longer. The only reason this didn't
cause the tests to take many minutes is that spread checkpoints only throttle
when writing out a buffer and there aren't that many dirty buffers in the
tests...

Discussion: https://postgr.es/m/20220117195711.xx4qbxutrrlmo2dg@alap3.anarazel.de

Branch
------
master

Details
-------
https://git.postgresql.org/pg/commitdiff/839f9636b374d67952fdb2b048f055393bc5a8f4

Modified Files
--------------
src/bin/pg_basebackup/t/010_pg_basebackup.pl | 104 ++++++++++++++-------------
src/bin/pg_verifybackup/t/002_algorithm.pl   |   2 +-
src/bin/pg_verifybackup/t/003_corruption.pl  |   2 +-
src/bin/pg_verifybackup/t/004_options.pl     |   2 +-
src/bin/pg_verifybackup/t/006_encoding.pl    |   2 +-
src/bin/pg_verifybackup/t/007_wal.pl         |   4 +-
6 files changed, 61 insertions(+), 55 deletions(-)

Hi Andres,

On Mon, Jan 17, 2022 at 11:48:02PM +0000, Andres Freund wrote:
> tests: Consistently use pg_basebackup -cfast --no-sync to accelerate tests.
>
> Most tests invoking pg_basebackup themselves did not yet use -cfast, which
> makes pg_basebackup take considerably longer. The only reason this didn't
> cause the tests to take many minutes is that spread checkpoints only throttle
> when writing out a buffer and there aren't that many dirty buffers in the
> tests...

Are you planning to back-patch any of this?  It seems to me that it
would not hurt, and making the tests faster on back-branches saves
time when working on those branches.

Thanks,
--
Michael

Hi,

On 2022-01-18 10:18:36 +0900, Michael Paquier wrote:
> On Mon, Jan 17, 2022 at 11:48:02PM +0000, Andres Freund wrote:
> > tests: Consistently use pg_basebackup -cfast --no-sync to accelerate tests.
> > 
> > Most tests invoking pg_basebackup themselves did not yet use -cfast, which
> > makes pg_basebackup take considerably longer. The only reason this didn't
> > cause the tests to take many minutes is that spread checkpoints only throttle
> > when writing out a buffer and there aren't that many dirty buffers in the
> > tests...
> 
> Are you planning to back-patch any of this?  It seems to me that it
> would not hurt, and making the tests faster on back-branches saves
> time when working on those branches.

I hadn't planned to do so. Back-branch tests don't run even close to as often
and there's enough difference across the branches that it's a bit of
per-branch work to adjust. To me this mostly seems helpful to reduce test
times on buildfarm critters / CI and branches we constantly run the tests for.

Greetings,

Andres Freund