Re: pg_upgrade instructions involving "rsync --size-only" might lead to standby corruption? - Mailing list pgsql-hackers

Thanks for that!

> I agree with Andrey – without it, we don't have any good way to upgrade
> large clusters in short time. Default rsync mode (without "--size-only")
> takes a lot of time too, if the load is heavy.
>
> With these adjustments, can "rsync --size-only" remain in the docs as the
> *fast* and safe method to upgrade standbys, or there are still some
> concerns related to corruption risks?

I hope somebody can answer that definitively, but I read Stephen's mail
to indicate that this procedure should be safe in principle (if you know
what you are doing).

> +++ b/doc/src/sgml/ref/pgupgrade.sgml
> @@ -380,22 +380,28 @@ NET STOP postgresql-&majorversion;
>      </para>
> 
>      <para>
> -     Streaming replication and log-shipping standby servers can
> +     Streaming replication and log-shipping standby servers must
>       remain running until a later step.
>      </para>
>     </step>
> 
> -   <step>
> +   <step id="pgupgrade-step-prepare-standbys">
>
>      <para>
> -     If you are upgrading standby servers using methods outlined in section <xref
> -     linkend="pgupgrade-step-replicas"/>, verify that the old standby
> -     servers are caught up by running <application>pg_controldata</application>
> -     against the old primary and standby clusters.  Verify that the
> -     <quote>Latest checkpoint location</quote> values match in all clusters.
> -     (There will be a mismatch if old standby servers were shut down
> -     before the old primary or if the old standby servers are still running.)
> +     If you are upgrading standby servers using methods outlined in
> +     <xref linkend="pgupgrade-step-replicas"/>,

You dropped the "section" before the xref, I think that should be kept
around.

> +                                                ensure that they were running when
> +     you shut down the primaries in the previous step, so all the latest changes

You talk of primaries in plural here, that is a bit weird for PostgreSQL
documentation.

> +     and the shutdown checkpoint record were received. You can verify this by running
> +     <application>pg_controldata</application> against the old primary and standby
> +     clusters. The <quote>Latest checkpoint location</quote> values must match in all
> +     nodes. A mismatch might occur if old standby servers were shut down before
> +     the old primary. To fix a mismatch, start all old servers and return to the
> +     previous step; proceeding with mismatched
> +     <quote>Latest checkpoint location</quote> may lead to standby corruption.
> +    </para>
> +
> +    <para>
>       Also, make sure <varname>wal_level</varname> is not set to
>       <literal>minimal</literal> in the <filename>postgresql.conf</filename> file on the
>       new primary cluster.
> @@ -497,7 +503,6 @@ pg_upgrade.exe
>       linkend="warm-standby"/>) standby servers, you can follow these steps to
>       quickly upgrade them.  You will not be running <application>pg_upgrade</application> on
>       the standby servers, but rather <application>rsync</application> on the primary.
> -     Do not start any servers yet.
>      </para>
> 
>      <para>
> @@ -508,6 +513,15 @@ pg_upgrade.exe
>       is running.
>      </para>
> 
> +    <para>
> +     Before running rsync, to avoid standby corruption, it is absolutely
> +     critical to ensure that both primaries are shut down and standbys
> +     have received the last changes (see <xref linkend="pgupgrade-step-prepare-standbys"/>).

I think this should be something like "ensure both that the primary is
shut down and that the standbys have received all the changes".

> +     Standbys can be running at this point or fully stopped.

"or be fully stopped." I think.

> +                                                             If they
> +     are still running, you can stop, upgrade, and start them one by one; this
> +     can be useful to keep the cluster open for read-only transactions.
> +    </para>

Maybe this is clear from the context, but "upgrade" in the above should
maybe more explicitly refer to the rsync method or else people might
think one can run pg_upgrade on them after all?