Re: Legacy GiST invalid tuples

Andres Freund <andres(at)anarazel(dot)de> writes:
> On 2018-07-04 17:02:01 -0400, Alvaro Herrera wrote:
>> On 2018-Jul-04, Andres Freund wrote:
>>> I don't think we can easily require that everyone pg_upgrading to v13+
>>> upgrades to v12 first?

>> We've never done that, I don't know if we can get away with it.
>> Upgrades with pg_upgrade are not cheap enough for that, methinks.
>> Maybe I'm wrong, but people complain even at a *single* pg_upgrade --
>> seems everybody wants in-place upgrades nowadays ...

> Right. That's why I was wondering about making the "last full scan"
> feature backpatchable...

ISTM this is closely related to the business about on-line checksum
enabling that Magnus et al have been working on. In general, we have a
lot of pent-up demand for changes that require on-disk data modification,
so it seems like it's time to design a generic mechanism for doing that
in a nice way (ie without downtime or huge performance penalties).

regards, tom lane