Re: Do we expect tests to work with default_transaction_isolation=serializable

From: Noah Misch <noah(at)leadboat(dot)com>
To: Andres Freund <andres(at)anarazel(dot)de>
Cc: pgsql-hackers(at)postgresql(dot)org
Subject: Re: Do we expect tests to work with default_transaction_isolation=serializable
Date: 2019-06-15 18:47:39
Message-ID: 20190615184739.GC239428@rfd.leadboat.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Sun, May 19, 2019 at 03:55:06PM -0700, Andres Freund wrote:
> I seem to recall that we expect tests to either work with
> default_transaction_isolation=serializable, or to set it to a different
> level where needed.
>
> Currently that's not the case. When running check-world with PGOPTIONS
> set to -c default_transaction_isolation=serializable I get easy to fix
> failures (isolation, plpgsql) but also some apparently hanging tests
> (003_recovery_targets.pl, 003_standby_2.pl).
>
> Do we expect this to work? If it's desirable I'll set up an animal that
> forces it to on.

I'm +1 for making it a project expectation, with an animal to confirm. It's
not expected to work today.

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Noah Misch 2019-06-15 20:39:11 Re: Draft back-branch release notes are up for review
Previous Message Noah Misch 2019-06-15 18:37:59 Re: pg_upgrade can result in early wraparound on databases with high transaction load