Re: XMAX_LOCK_ONLY and XMAX_COMMITTED (fk/multixact code)

> On Aug 27, 2020, at 4:58 PM, Peter Geoghegan <pg@bowt.ie> wrote:
>
> On Thu, Aug 27, 2020 at 4:57 PM Mark Dilger
> <mark.dilger@enterprisedb.com> wrote:
>> The amcheck patch has Asserts in hio.c that purport to disallow writing invalid header bits to disk.
>
> Can it also be a runtime check for the verification process? I think
> that we can easily afford to be fairly exhaustive about stuff like
> this.

These two are both checked in verify_heapam.c.  The point is that the system will also refuse to write out pages that
havethis corruption.  The Asserts could be converted to panics or whatever, but that has other more serious
consequences. Did you have something else in mind? 

—
Mark Dilger
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company