Hi,
On 2022-02-03 16:18:27 -0500, John Naylor wrote:
> I just checked some client case notes where they tried just that
> before getting outside help, and both SELECT and VACUUM FREEZE
> commands were rejected.
What kind of SELECT was that? Any chance it caused a write via functions, a
view, whatnot? And what version? What was the exact error message?
VACUUM FREEZE is a *terrible* idea to run when encountering anti-wraparound
issues. I understand why people thing key might need it, but basically all it
achieves is to make VACUUM do a lot more, none of it helpful to get out of the
wraparound-can't-write situation (those rows will already get frozen).
I'd plus one the addition of a HINT that tells users that FREEZE likely is a
bad idea when in wraparound land. We should allow it, because there are
situation where it might make sense, but the people that can make that
judgement know they can ignore the HINT.
Greetings,
Andres Freund