Re: Catalog_xmin is not advanced when a logical slot is lost

On Mon, Nov 21, 2022 at 9:12 AM Alvaro Herrera <alvherre@alvh.no-ip.org>
wrote:

> On 2022-Nov-21, sirisha chamarthi wrote:
>
> > On Mon, Nov 21, 2022 at 8:05 AM Alvaro Herrera <alvherre@alvh.no-ip.org>
> > wrote:
>
> > > Thank you.  I had pushed mine for CirrusCI to test, and it failed the
> > > assert I added in slot.c:
> > > https://cirrus-ci.com/build/4786354503548928
> > > Not yet sure why, looking into it.
> >
> > Can this be because restart_lsn is not set to InvalidXLogRecPtr for the
> > physical slots?
>
> Hmm, that makes no sense.  Is that yet another bug?  Looking.
>

It appears to be. wal_sender is setting restart_lsn to a valid LSN even
when the slot is invalidated.

postgres=# select pg_Create_physical_replication_slot('s1');
 pg_create_physical_replication_slot
-------------------------------------
 (s1,)
(1 row)

postgres=# select * from pg_replication_slots;
 slot_name | plugin | slot_type | datoid | database | temporary | active |
active_pid | xmin | catalog_xmin | restart_lsn | confirmed_flush_lsn |
wal_status | safe_wal_size | two_phase
-----------+--------+-----------+--------+----------+-----------+--------+------------+------+--------------+-------------+---------------------+------------+---------------+-----------
 s1        |        | physical  |        |          | f         | f      |
           |      |              |             |                     |
       |   -8254390272 | f
(1 row)

postgres=# checkpoint;
CHECKPOINT
postgres=# select * from pg_replication_slots;
 slot_name | plugin | slot_type | datoid | database | temporary | active |
active_pid | xmin | catalog_xmin | restart_lsn | confirmed_flush_lsn |
wal_status | safe_wal_size | two_phase
-----------+--------+-----------+--------+----------+-----------+--------+------------+------+--------------+-------------+---------------------+------------+---------------+-----------
 s1        |        | physical  |        |          | f         | f      |
           |      |              |             |                     |
       |   -8374095064 | f
(1 row)

postgres=# \q
postgres@pgvm:~$ /usr/local/pgsql/bin/pg_receivewal -S s1 -D .
pg_receivewal: error: unexpected termination of replication stream: ERROR:
 requested WAL segment 0000000100000000000000EB has already been removed
pg_receivewal: disconnected; waiting 5 seconds to try again
^Cpostgres@pgvm:~$ /usr/local/pgsql/bin/psql
psql (16devel)
Type "help" for help.

postgres=# select * from pg_replication_slots;
server closed the connection unexpectedly
        This probably means the server terminated abnormally
        before or while processing the request.
The connection to the server was lost. Attempting reset: Failed.
The connection to the server was lost. Attempting reset: Failed.
!?> ^C
!?>


In the log:
2022-11-21 17:31:48.159 UTC [3953664] STATEMENT:  START_REPLICATION SLOT
"s1" 0/EB000000 TIMELINE 1
TRAP: failed Assert("XLogRecPtrIsInvalid(slot_contents.data.restart_lsn)"),
File: "slotfuncs.c", Line: 371, PID: 3953707


>
> --
> Álvaro Herrera               48°01'N 7°57'E  —
> https://www.EnterpriseDB.com/
> "No es bueno caminar con un hombre muerto"
>