Re: Fwd: upgrade to PG11 on secondary fails (no initdb was launched)

Stephen Frost <[email protected]> Fri, 14 Jun 2019 12:15:38 -0400

Newsgroups	gmane.comp.db.postgresql.performance
Message-ID	<[email protected]>

Greetings,

* Bruce Momjian ([email protected]) wrote:
> On Fri, Jun 14, 2019 at 10:39:40AM -0400, Bruce Momjian wrote:
> > On Fri, Jun 14, 2019 at 03:12:29PM +0200, Fabio Pardi wrote:
> > > Using --size-only, tells rsync to only check the size of the blocks.
> > > That is: if the block is present on the destination, and is the same
> > > size as the origin, then skip.
> > 
> > The files are _exactly_ the same on primary and standby, so we don't
> > need to check anything.  Frankly, it is really only doing hard linking
> > of the files.
> 
> Here is the description from our docs:
> 
>        What this does is to record the links created by pg_upgrade's
>        link mode that connect files in the old and new clusters on the
>        primary server.  It then finds matching files in the standby's old
>        cluster and creates links for them in the standby's new cluster.
>        Files that were not linked on the primary are copied from the
>        primary to the standby.  (They are usually small.)  This provides
>        rapid standby upgrades.  Unfortunately, rsync needlessly copies
>        files associated with temporary and unlogged tables because these
>        files don't normally exist on standby servers.
> 
> The primary and standby have to be binary the same or WAL replay would
> not work on the standby.  (Yes, I sometimes forgot how this worked so I
> wrote it down in the docs.)  :-)

Right- this is *not* a general process for building a replica, this is
specifically *only* for when doing a pg_upgrade and *everything* is shut
down when it runs, and every step is checked to ensure that there are no
errors during the process.

Thanks!

Stephen

signature.asc (application/pgp-signature, 819 B)