Re: Non-reserved replication slots and slot advancing

Hi,

On 2018-06-26 16:13:05 +0900, Michael Paquier wrote:
> I have been chewing for the last couple of days on this email from
> Horiguchi-san:
> https://www.postgresql.org/message-id/20180622.163312.254556300.horiguchi.kyotaro@lab.ntt.co.jp
> 
> As summarized, it is actually strange to be able to advance a slot which
> has a non-reserved restart_lsn.  For example, take that which can happen
> on HEAD:
> =# select pg_create_physical_replication_slot('toto');
>  pg_create_physical_replication_slot
> -------------------------------------
>  (toto,)
> (1 row)
> =# select pg_replication_slot_advance('toto', '0/1');
>  pg_replication_slot_advance
> -----------------------------
>  (toto,0/1)
> (1 row)
> =# select slot_name, restart_lsn from pg_replication_slots ;
>  slot_name | restart_lsn
> -----------+-------------
>  toto      | 0/1
> (1 row)

I'm not clear to why this is a problem? Seems like either behaviour can
be argued for. I don't really have an opinion either way. I'd just
remove the item from the open items list, I don't think we need to hold
up the release for it?

Greetings,

Andres Freund