Thread: pg_rewind confusion

pg_rewind confusion

From
Bruce Momjian
Date:
I was confused by pg_rewind's ability to use a remote source server, and
another core team member was confused as well, so I figured I should
update the pg_rewind docs to be clearer.

Attached is a patch to do that.

--
  Bruce Momjian  <bruce@momjian.us>        http://momjian.us
  EnterpriseDB                             http://enterprisedb.com

+ As you are, so once was I. As I am, so you will be. +
+                     Ancient Roman grave inscription +

Attachment

Re: pg_rewind confusion

From
Michael Paquier
Date:
On Thu, Jul 21, 2016 at 3:07 AM, Bruce Momjian <bruce@momjian.us> wrote:
> I was confused by pg_rewind's ability to use a remote source server, and
> another core team member was confused as well, so I figured I should
> update the pg_rewind docs to be clearer.
>
> Attached is a patch to do that.

This is a clean set of improvements.

-   <xref linkend="guc-full-page-writes"> must also be enabled.  That is the
-   default.
+   <xref linkend="guc-full-page-writes"> must also be on, but it is
+   enabled by default.
Perhaps you should add a <literal> markup around "on"? And say "must
be also set to on"

-        with superuser access. This option requires the server to be running
-        and not in recovery mode.
+        with superuser access. This option requires the local or remote
+        server to be running and not in recovery mode.
I don't think that "local or remote" is an improvement here. You could
just say "source".

+      Copy all other files such as <filename>clog</filename> and
clog is not a term used a lot in the docs. It may be better to refer
to pg_clog, or transaction status data.
--
Michael


Re: pg_rewind confusion

From
Bruce Momjian
Date:
On Thu, Jul 21, 2016 at 10:11:00PM +0900, Michael Paquier wrote:
> On Thu, Jul 21, 2016 at 3:07 AM, Bruce Momjian <bruce@momjian.us> wrote:
> > I was confused by pg_rewind's ability to use a remote source server, and
> > another core team member was confused as well, so I figured I should
> > update the pg_rewind docs to be clearer.
> >
> > Attached is a patch to do that.
>
> This is a clean set of improvements.
>
> -   <xref linkend="guc-full-page-writes"> must also be enabled.  That is the
> -   default.
> +   <xref linkend="guc-full-page-writes"> must also be on, but it is
> +   enabled by default.
> Perhaps you should add a <literal> markup around "on"? And say "must
> be also set to on"

Well, "must also be on" isn't using "on" as a literal, but "must be also
set to on" does use it as a literal, and I changed the text to that.

> -        with superuser access. This option requires the server to be running
> -        and not in recovery mode.
> +        with superuser access. This option requires the local or remote
> +        server to be running and not in recovery mode.
> I don't think that "local or remote" is an improvement here. You could
> just say "source".

I was trying to highlight that the source can be remote with
--source-server.  I have change it to what you suggested.  Is it stil
clear?

> +      Copy all other files such as <filename>clog</filename> and
> clog is not a term used a lot in the docs. It may be better to refer
> to pg_clog, or transaction status data.

Agreed, changed to "pg_clog".

Updated patch attached.

--
  Bruce Momjian  <bruce@momjian.us>        http://momjian.us
  EnterpriseDB                             http://enterprisedb.com

+ As you are, so once was I. As I am, so you will be. +
+                     Ancient Roman grave inscription +

Attachment

Re: pg_rewind confusion

From
Michael Paquier
Date:
On Fri, Jul 22, 2016 at 2:15 AM, Bruce Momjian <bruce@momjian.us> wrote:
> On Thu, Jul 21, 2016 at 10:11:00PM +0900, Michael Paquier wrote:
>> +      Copy all other files such as <filename>clog</filename> and
>> clog is not a term used a lot in the docs. It may be better to refer
>> to pg_clog, or transaction status data.
>
> Agreed, changed to "pg_clog".
>
> Updated patch attached.

Thanks. That looks fine to me. This documentation becomes really clearer.
--
Michael


Re: pg_rewind confusion

From
Bruce Momjian
Date:
On Fri, Jul 22, 2016 at 08:35:32AM +0900, Michael Paquier wrote:
> On Fri, Jul 22, 2016 at 2:15 AM, Bruce Momjian <bruce@momjian.us> wrote:
> > On Thu, Jul 21, 2016 at 10:11:00PM +0900, Michael Paquier wrote:
> >> +      Copy all other files such as <filename>clog</filename> and
> >> clog is not a term used a lot in the docs. It may be better to refer
> >> to pg_clog, or transaction status data.
> >
> > Agreed, changed to "pg_clog".
> >
> > Updated patch attached.
>
> Thanks. That looks fine to me. This documentation becomes really clearer.

Patch applied to head only as this did not fix any documentation bugs.

--
  Bruce Momjian  <bruce@momjian.us>        http://momjian.us
  EnterpriseDB                             http://enterprisedb.com

+ As you are, so once was I. As I am, so you will be. +
+                     Ancient Roman grave inscription +