On Mon, Mar 7, 2022 at 5:12 PM kuroda.hay...@fujitsu.com <kuroda.hay...@fujitsu.com> wrote: > > Dear Peter, > > > > So, why does the patch use syntax option 1? > > IMU it might be useful for the following case. > > Assuming that multi-master configuration with node1, node2. > Node1 has a publication pub1 and a subscription sub2, node2 has pub2 and sub1. > > From that situation, please consider that new node node3 is added > that subscribe some changes from node2. > > If the feature is introduced as option1, new publication must be defined in > node2. > If that is introduced as option2, however, maybe pub2 can be reused. > i.e. multiple declaration of publications can be avoided. >
Yes. Thanks for the example. I had the same observation in my last post [1] ------ [1] https://www.postgresql.org/message-id/CAHut%2BPtRxiQR_4UFLNThg-NNRV447FvwtcR-BvqMzjyMJXKwfw%40mail.gmail.com Kind Regards, Peter Smith Fujitsu Australia.
