Thread: pg_standby supported versions

pg_standby supported versions

From
Mark Kirkwood
Date:
I noticed in the docs for pg_standby (docs/src/sgml/pgstandby.sgml) that
we have supported version >= 8.2. However it does seem to work ok with
earlier versions (e.g 8.1) - or am I missing something?

cheers

Mark

Re: pg_standby supported versions

From
Decibel!
Date:
On Aug 13, 2008, at 4:59 AM, Mark Kirkwood wrote:
> I noticed in the docs for pg_standby (docs/src/sgml/pgstandby.sgml)
> that we have supported version >= 8.2. However it does seem to work
> ok with earlier versions (e.g 8.1) - or am I missing something?

Hrm, AFAIK it works... Simon?
--
Decibel!, aka Jim C. Nasby, Database Architect  decibel@decibel.org
Give your computer some brain candy! www.distributed.net Team #1828



Attachment

Re: pg_standby supported versions

From
Simon Riggs
Date:
On Fri, 2008-08-22 at 00:50 -0500, Decibel! wrote:
> On Aug 13, 2008, at 4:59 AM, Mark Kirkwood wrote:
> > I noticed in the docs for pg_standby (docs/src/sgml/pgstandby.sgml)
> > that we have supported version >= 8.2. However it does seem to work
> > ok with earlier versions (e.g 8.1) - or am I missing something?
>
> Hrm, AFAIK it works... Simon?

It works, yes. But in basic form only.

You'll have problems stopping and starting standby, but the most
important change is that the process is not restartable. So if you run
it for a month and then crash, you'll need to re-run the whole month's
WAL data.

--
 Simon Riggs           www.2ndQuadrant.com
 PostgreSQL Training, Services and Support


Re: pg_standby supported versions

From
Mark Kirkwood
Date:
Simon Riggs wrote:
> On Fri, 2008-08-22 at 00:50 -0500, Decibel! wrote:
>
>> On Aug 13, 2008, at 4:59 AM, Mark Kirkwood wrote:
>>
>>> I noticed in the docs for pg_standby (docs/src/sgml/pgstandby.sgml)
>>> that we have supported version >= 8.2. However it does seem to work
>>> ok with earlier versions (e.g 8.1) - or am I missing something?
>>>
>> Hrm, AFAIK it works... Simon?
>>
>
> It works, yes. But in basic form only.
>
> You'll have problems stopping and starting standby, but the most
> important change is that the process is not restartable. So if you run
> it for a month and then crash, you'll need to re-run the whole month's
> WAL data.
>
>
Exactly, after posting this I discovered the relevant CVS logs for
changes to support restartable recovery. It seems to me that perhaps a
little more detail in the docs would be good here.

At work some folks are using pg_standby with 8.1 and while it does work
quite well, it took me a wee while to figure out that some of the issues
they are having are tied up with inability to continue recovery after
restarting the database or the os!

Best wishes

Mark

Re: pg_standby supported versions

From
Simon Riggs
Date:
On Wed, 2008-11-26 at 22:14 +1300, Mark Kirkwood wrote:
> Simon Riggs wrote:
> >
> > You'll have problems stopping and starting standby, but the most
> > important change is that the process is not restartable. So if you
> run
> > it for a month and then crash, you'll need to re-run the whole
> month's
> > WAL data.
> >
> >
> Exactly, after posting this I discovered the relevant CVS logs for
> changes to support restartable recovery. It seems to me that perhaps a
> little more detail in the docs would be good here.

I would hope that the complete absence of anything would be all that is
required to dissuade people. We don't ever update the docs to say "BTW
feature X is not supported until next release". Not sure we should start
now.

--
 Simon Riggs           www.2ndQuadrant.com
 PostgreSQL Training, Services and Support


Re: pg_standby supported versions

From
Mark Kirkwood
Date:
Simon Riggs wrote:
> On Wed, 2008-11-26 at 22:14 +1300, Mark Kirkwood wrote:
>
>>
>> Exactly, after posting this I discovered the relevant CVS logs for
>> changes to support restartable recovery. It seems to me that perhaps a
>> little more detail in the docs would be good here.
>>
>
> I would hope that the complete absence of anything would be all that is
> required to dissuade people. We don't ever update the docs to say "BTW
> feature X is not supported until next release". Not sure we should start
> now.
>
>

While I understand, I also think it is perfectly reasonable to clarify
why a specific minimum version is being recommended. To me this conforms
to the open source spirit of telling folk why something is true, as
opposed to stating that it is.

I'm thinking of something like:

<para>

pg_standby is designed to work with PostgreSQL 8.2 and later, as it
requires the ability to interrupt and restart recovery (absent in
earlier server versions).

</para>