On Sun, Sep 29, 2024 at 6:50 AM Bruce Momjian <br...@momjian.us> wrote: > > On Thu, Sep 26, 2024 at 03:08:52PM +0530, Amit Kapila wrote: > > On Sat, Sep 21, 2024 at 1:50 AM Bruce Momjian <br...@momjian.us> wrote: > > > > > > On Fri, Sep 20, 2024 at 04:05:11PM -0400, Tom Lane wrote: > > > > Bruce Momjian <br...@momjian.us> writes: > > > > > Patch applied to PG 17. > > > > > > > > I don't see a push? > > > > > > Push was delayed because my test script found some uncommitted files due > > > to earlier testing. Should be fine now. > > > > > > > <para> > > <link > > > > linkend="app-pgcreatesubscriber"><application>pg_createsubscriber</application></link>, > > a utility that creates logical replicas from physical standbys > > </para> > > > > This description is okay but according to me, the more compelling use > > case is that this new utility helps to allow online upgrades of > > physical replication setup as explained in the blog [1]. See the > > section: "Upgrading Streaming (Physical) Replication Setup". > > > > </listitem> > > <listitem> > > <para> > > <link > > linkend="pgupgrade"><application>pg_upgrade</application></link> > > now > > preserves replication slots on both publishers and subscribers > > </para> > > > > It is better to write the above statement as: > > "pg_upgrade</application></link> now preserves replication slots on > > publishers and full subscription's state on subscribers". This is > > because replication slots are preserved on publishers. The subscribers > > preserve the subscription state. > > So, as I understand it, this preservation only happens when the _old_ > Postgres version is 17+. >
Yes. > Do we want to try and explain that in the > Postgres 17 release notes? > It would be good if we can capture that information without bloating the release document. However, this information is already present in pg_upgrade docs, so users have a way to know the same even if we can't mention it in the release notes. -- With Regards, Amit Kapila.
