Horton

Re: slowest tap tests - split or accelerate?

Tom Lane <tgl@sss.pgh.pa.us>

View thread View on postgr.es
From: Tom Lane <tgl@sss.pgh.pa.us>
To: Andres Freund <andres@anarazel.de>
Cc: Robert Haas <robertmhaas@gmail.com>, "pgsql-hackers@postgresql.org" <pgsql-hackers@postgresql.org>
Date: 2022-01-17T20:48:54Z
Lists: pgsql-hackers 
Andres Freund <andres@anarazel.de> writes:
> I've occasionally pondered caching initdb results and reusing them across
> tests - just the locking around it seems a bit nasty, but perhaps that could
> be done as part of the tmp_install step. Of course, it'd need to deal with
> different options etc...

I'd actually built a prototype to do that, based on making a reference
cluster and then "cp -a"'ing it instead of re-running initdb.  I gave
up when I found than on slower, disk-bound machines it was hardly
any faster.  Thinking about it now, I wonder why not just re-use one
cluster for many tests, only dropping and re-creating the database
in which the testing happens.

			regards, tom lane

Commits

  1. Don't clean initdb files on template creation failure

  2. Avoid non-POSIX cp flags

  3. Use "template" data directory in tests

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