Re: pg15b2: large objects lost on upgrade

On 8/3/22 4:19 PM, Tom Lane wrote:
> "Jonathan S. Katz" <jkatz(at)postgresql(dot)org> writes:
>> I did rule out wanting to do the "xid + $X" check after reviewing some
>> of the output. I think that both $X could end up varying, and it really
>> feels like a bandaid.
>
> It is that. I wouldn't feel comfortable with $X less than 100 or so,
> which is probably sloppy enough to draw Robert's ire. Still, realizing
> that what we want right now is a band-aid for 15beta3, I don't think
> it's an unreasonable short-term option.

Attached is the "band-aid / sloppy" version of the patch. Given from the
test examples I kept seeing deltas over 100 for relfrozenxid, I chose
1000. The mxid delta was less, but I kept it at 1000 for consistency
(and because I hope this test is short lived in this state), but can be
talked into otherwise.

>> Andres suggested upthread using "txid_current()" -- for the comparison,
>> that's one thing I looked at. Would any of the XID info from
>> "pg_control_checkpoint()" also serve for this test?
>
> I like the idea of txid_current(), but we have no comparable
> function for mxid do we? While you could get both numbers from
> pg_control_checkpoint(), I doubt that's sufficiently up-to-date.

...unless we force a checkpoint in the test?

Jonathan