On Tue, Mar 9, 2021 at 3:41 PM Markus Wanner <markus.wan...@enterprisedb.com> wrote: > > On 09.03.21 10:37, Amit Kapila wrote: > I guess I don't quite understand the initial motivation for the patch. > It states: "This allows plugins to not allow the enabling of streaming > and two_phase at the same time in logical replication." That's beyond > me ... "allows [..] to not allow"? Why not, an output plugin can still > reasonably request both. And that's a good thing, IMO. What problem > does the patch try to solve? >
AFAIU, Ajin doesn't want to mandate streaming with two_pc option. But, maybe you are right that it doesn't make sense for the user to provide both options but doesn't provide stream_prepare callback, and giving an error in such a case should be fine. I think if we have to follow Ajin's idea then we need to skip 2PC in such a case (both prepare and commit prepare) and make this a regular transaction. -- With Regards, Amit Kapila.
