Re: Add 64-bit XIDs into PostgreSQL 15

Hi all,

On 07.07.2025 11:17, Evgeny Voropaev wrote:
> Do-side:
> 1. Having page ABC with several tuples.
> 2. Starting to perform insertion of new tuple
>    2.1. In the case of an inappropriate xid_base, trying to fit base
>    2.1.1 Freezing and pruning tuples without further repairing
> fragmentation.
>     2.1.2 All tuples have been pruned (no alive tuples on the page
> since this moment)
> 3. Inserting a new tuple and setting XLOG_HEAP_INIT_PAGE, assuming
> that the only tuple located at the bottom of the page (assuming that
> fragmentation has been performed).
>
>
> Result: We have the ABC page with the new tuple inserted somewhere in
> the MIDDLE of the page and surrounded with garbage from dead and
> unused tuples. At the same time we have an xlog record bringing the
> XLOG_HEAP_INIT_PAGE bit.
>
> Redo-side
> 1. Observing XLOG_HEAP_INIT_PAGE
> 2. Creating a new page and inserting the new tuple into the first
> position of the page.

Evgeny, I've tried to reproduce described scenario, but I can't get
XLOG_HEAP_INIT_PAGE bit set.

Short steps description.

1. create test table, insert tuple, then exec few update operations,
then exec delete operation,
all tuples placed on one page and "select * from test_table;" return 0
rows.
2. stop postgres and use pg_resetwal to promote xid. This will help us
to prune tuples.
3. start postgres and insert new tuple on page, old tuples pruning
occurs and here we expect
XLOG_HEAP_INIT_PAGE bit set in xlog rec. As I understand, it is the
reason of further tuple location
inconsistency.
But as I see, XLOG_HEAP_INIT_PAGE bit isn't set. The reason is condition

        if (ItemPointerGetOffsetNumber(&(heaptup->t_self)) ==
FirstOffsetNumber &&
            PageGetMaxOffsetNumber(page) == FirstOffsetNumber)
        {
            info |= XLOG_HEAP_INIT_PAGE;
            bufflags |= REGBUF_WILL_INIT;
        }
As I see, new heaptup offset isn't equal to FirstOffsetNumber and it is
quite enough to skip bit set.
I've used waldump to check bit set

rmgr: Heap        len (rec/tot):     75/    75, tx: 2147485000, lsn:
0/02001E90, prev 0/02001E50, desc: INSERT off: 4, flags: 0x00, blkref
#0: rel 1663/5/16431 blk 0

, so offset is 4 and desc: INSERT, in case XLOG_HEAP_INIT_PAGE is set,
it should be INSERT+INIT.
In attached file you can find detailed test description with results, so
you can repeat it.

Maybe I missed something in your scenario, can you please clarify?
Or maybe it is't reproducable at all.

Best regards,
Maksim Melnikov