Re: [BUG] pg_basebackup from disconnected standby fails

From: Kyotaro HORIGUCHI <horiguchi(dot)kyotaro(at)lab(dot)ntt(dot)co(dot)jp>
To: michael(dot)paquier(at)gmail(dot)com
Cc: alvherre(at)2ndquadrant(dot)com, pgsql-hackers(at)postgresql(dot)org
Subject: Re: [BUG] pg_basebackup from disconnected standby fails
Date: 2016-07-21 09:52:49
Message-ID: 20160721.185249.91505374.horiguchi.kyotaro@lab.ntt.co.jp
Views: Raw Message | Whole Thread | Download mbox
Thread:
Lists: pgsql-hackers

Hello,

At Mon, 11 Jul 2016 16:47:53 +0900, Michael Paquier <michael(dot)paquier(at)gmail(dot)com> wrote in <CAB7nPqQJqEMqY_caBh0=dPBxpmSfd6_Uf4HXSG=LgB1cNsZL=g(at)mail(dot)gmail(dot)com>
> On Mon, Jul 11, 2016 at 4:40 PM, Kyotaro HORIGUCHI
> <horiguchi(dot)kyotaro(at)lab(dot)ntt(dot)co(dot)jp> wrote:
> > That's true, but we don't always have a perfectly comprehensive
> > test suite, consciously or unconsciously. The sentence was
> > inattentive but the "bug" was just the negative comparable to
> > "feature" in my mind. My point was the comparison between adding
> > a test for a corner-case and its cost. It must be added if the
> > fixed feature is fragile. It can be added it doesn't take a too
> > long time to finish.
>
> I'd just add it to be honest. Taking backups from standbys, with or
> without the master being connected is a supported feature, and we want
> to follow-up to be sure that it does not in the future. Having now the
> infrastructure for more complex scenarios does not mean of course that
> we need to test everything and all kind of things, of course, but here
> the benefits are good compared to the cost that a single call of
> pg_basebackup has in the complete the test suite run.

Yes, totally agreed.

regards,

--
Kyotaro Horiguchi
NTT Open Source Software Center

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message happy times 2016-07-21 10:39:20 Unexpected memory usage for repeated inserts within plpgsql function
Previous Message Kyotaro HORIGUCHI 2016-07-21 09:50:07 Re: asynchronous and vectorized execution