RE: Skipping logical replication transactions on subscriber side

On July 19, 2021 2:40 PM Masahiko Sawada  wrote:
> I've attached the updated version patch that incorporated all comments
> I got so far except for the clearing error details part I mentioned
> above. After getting a consensus on those parts, I'll incorporate the
> idea into the patches.

Hi Sawada-san,

I am interested in this feature.
After having a look at the patch, I have a few questions about it.
(Sorry in advance if I missed something)

1) In 0002 patch, it introduces a new view called pg_stat_subscription_errors.
   Since it won't be cleaned automatically after we resolve the conflict, do we
   need a reset function to clean the statistics in it ? Maybe something
   similar to pg_stat_reset_replication_slot which clean the
   pg_stat_replication_slots.

2) For 0003 patch, When I am faced with a conflict, I set skip_xid = xxx, and
   then I resolve the conflict. If I reset skip_xid after resolving the
   conflict, will the change(which cause the conflict before) be applied again ?

3) For 0003 patch, if user set skip_xid to a wrong xid which have not been
   assigned, and then will the change be skipped when the xid is assigned in
   the future even if it doesn't cause any conflicts ?

Besides, It might be better to add some description of patch in each patch's
commit message which will make it easier for new reviewers to follow.


Best regards,
Houzj