Hi,
On 2018-06-08 13:30:33 -0500, Jeremy Finzel wrote:
> Avoid spuriously marking pages as all-visible (Dan Wood, Pavan Deolasee,
> Álvaro Herrera)
>
> This could happen if some tuples were locked (but not deleted). While
> queries would still function correctly, vacuum would normally ignore such
> pages, with the long-term effect that the tuples were never frozen. In
> recent releases this would eventually result in errors such as "found
> multixact nnnnn from before relminmxid nnnnn".
Oh, I already had forgotten about that one... It does sound like a
likely explanation of your issue. Hard to tell without investigating a
*lot* more closely than I realistically can do remotely. It seems quite
possible to be the cause - I'd strongly suggest to upgrade to prevent
further occurances, or at least exclude it as a cause.
Greetings,
Andres Freund