Re: improving wraparound behavior - Mailing list pgsql-hackers

From Stephen Frost
Subject Re: improving wraparound behavior
Date
Msg-id 20190506110955.GY6197@tamriel.snowman.net
Whole thread Raw
In response to Re: improving wraparound behavior  (Andres Freund <andres@anarazel.de>)
List pgsql-hackers
Greetings,

* Andres Freund (andres@anarazel.de) wrote:
> On 2019-05-03 23:08:44 -0400, Stephen Frost wrote:
> > As mentioned though, it's likely to be a
> > quite rare thing to run into, so you'd have to be extra unlucky to even
> > hit this case and perhaps the extra code complication just isn't worth
> > it.
>
> I think just having an actual reason for the problem would be so much
> better than the current status, that I'd tackle the "oops, just about
> everything is screwed, here's the reasons in order" case separately (or
> just plain never).

I certainly agree with that, but if we really think it's so rare that we
don't feel that it's worth worrying about then I'd say we should remove
the 'else' in the 'else if', as I initially suggested, since the chances
of users getting more than one is quite rare and more than two would
be.. impressive, which negates the concern you raised earlier about
there being a bunch of these messages that make it hard for users to
reason about what is the most critical.  Removing the 'else' would make
it strictly less code (albeit not much, but it certainly isn't adding to
the code) and would be more informative for the user who does end up
having two cases happen at (or nearly) the same time.

Thanks,

Stephen

Attachment

pgsql-hackers by date:

Previous
From: Rafia Sabih
Date:
Subject: Re: Pluggable Storage - Andres's take
Next
From: Thomas Munro
Date:
Subject: Re: Fixing order of resowner cleanup in 12, for Windows