Re: [COMMITTERS] pgsql: Fix freezing of a dead HOT-updated tuple

Alvaro Herrera <alvherre(at)alvh(dot)no-ip(dot)org> wrote:
>He means that the tuple that heap_update moves to page 1 (which will no
>longer be processed by vacuum) will contain a multixact that's older
>than relminmxid -- because it is copied unchanged by heap_update instead
>of properly checking against age limit.

I see. The problem is more or less with this heap_update() code:

/*
* And also prepare an Xmax value for the new copy of the tuple. If there
* was no xmax previously, or there was one but all lockers are now gone,
* then use InvalidXid; otherwise, get the xmax from the old tuple. (In
* rare cases that might also be InvalidXid and yet not have the
* HEAP_XMAX_INVALID bit set; that's fine.)
*/
if ((oldtup.t_data->t_infomask & HEAP_XMAX_INVALID) ||
HEAP_LOCKED_UPGRADED(oldtup.t_data->t_infomask) ||
(checked_lockers && !locker_remains))
xmax_new_tuple = InvalidTransactionId;
else
xmax_new_tuple = HeapTupleHeaderGetRawXmax(oldtup.t_data);

My naive guess is that we have to create a new MultiXactId here in at
least some cases, just like FreezeMultiXactId() sometimes does.

--
Peter Geoghegan