Ben Kammer <b_a_kammer@yahoo.co.uk> writes:
> I have a postgresql 9.5.4 cluster and was testing with vacuum freeze. I
> expected the value of xmin to be set to '2' when I vacuum a particular
> table with the VACUUM FREEZE command.
That doesn't happen anymore; freezing just sets an info bit that indicates
that the xmin should be treated as old.
regards, tom lane