Hi Fabrice,

On Fri, 31 Oct 2025 at 17:45, Fabrice Chapuis <[email protected]>
wrote:

> Thanks for your large analyze and explanation Alexander. If we go in the
> direction you propose and leave the option to use a suplementary flag
> allow_overwrite, may I propose you some modifications in the patch v0 you
> have attached:
>
>  Why modifying this function?
>
>  drop_local_obsolete_slots(List *remote_slot_list)
>

Think about the following case:
1. We have node1 (primary) and node2 (standby), with slot foo(failover=true)
2. We stop node1 for maintenance and promote node2
3. DROP replication slot on node2
4. Start node1

In this scenario the logical slot "foo" will be left unattended and it
needs to be dropped, because it doesn't exist on the primary.

As I said earlier, IMO it is only the "failover" flag that effectively
indicates the purpose of the slot. The sync flag is purely informative.

Regards,
--
Alexander Kukushkin

Reply via email to