Re: pageinspect get_raw_page() option to return invalid pages

On 2018-05-04 11:42:35 -0700, Peter Geoghegan wrote:
> On Fri, May 4, 2018 at 11:41 AM, Andres Freund <andres(at)anarazel(dot)de> wrote:
> > Hi,
> >
> > Currently there's no convenient way to get a corrupted page (be it a
> > checksum failure, corrupted pd_upper/lower or whatnot) via pageinspect's
> > get_raw_page(). Given that pageinspect is kind of our tool to do inspect
> > issues on a data level that's not so great.
> >
> > I therefore propose that we add an *option* that bypasses shared buffers
> > and reads the underlying data directly. And then skips over the
> > validation. I'm not 100% sure if this should be done unconditionally,
> > or only if the item wasn't found in cache. There's some coherency
> > differences, obviously.
>
> +1. This would be a good option for amcheck, too.

Could you expand on that? Are you envisioning an option to
ReadBufferExtended()? Because that's certainly not what I'm thinking of
- it seems dangerous to populate shared buffers with an invalid
page. Therefore I was more thinking to falling back to smgrread() or
such.

Greetings,

Andres Freund