On Sun, 8 Nov 2020 at 18:19, Greg Stark <stark@mit.edu> wrote:
>
> We had an outage caused by transaction wraparound. And yes, one of the
> first things I did on this site was check that we didn't have any
> databases that were in danger of wraparound.
Fwiw this patch has been in "Ready for Committer" state since April
and has been moved forward three times including missing the release.
It's a pretty short patch and fixes a problem that caused an outage
for $previous_employer and I've had private discussions from other
people who have been struggling with the same issue. Personally I
consider it pretty close to a bug fix and worth backpatching. I think
it's pretty annoying to have put out a release without this fix.
--
greg