On Fri, Sep 9, 2022 at 5:36 PM Ashutosh Bapat <ashutosh.bapat....@gmail.com> wrote: > > On Thu, Sep 8, 2022 at 8:32 PM Ashutosh Sharma <ashu.coe...@gmail.com> wrote: > > > > On Thu, Sep 8, 2022 at 6:23 PM Ashutosh Bapat > > <ashutosh.bapat....@gmail.com> wrote: > > > > > > On Thu, Sep 8, 2022 at 4:14 PM Ashutosh Sharma <ashu.coe...@gmail.com> > > > wrote: > Can you please point to the documentation. >
AFAIU there is just one documentation. Here is the link for it: https://www.postgresql.org/docs/current/view-pg-replication-slots.html > It's true that it needs to be clarified. But what you are saying may > not be entirely true in case of streamed transaction. In that case we > might send logically decoded changes of an ongoing transaction as > well. They may even get applied but not necessarily committed. It's a > bit complicated. :) > This can happen in case of big transactions. That's the reason I mentioned that the transaction has a small set of data which is not yet committed but the confirmed_flush_lsn says it has already reached the logical subscriber. And.. lastly sorry for the delayed response. I was not well and couldn't access email for quite some days. The poor dengue had almost killed me :( -- With Regards, Ashutosh Sharma.