Обсуждение: Re: postgresql.org inconsistent (Re: [GENERAL] PG replication across DataCenters)

I should have cross-posted this to pgsql-docs from the beginning, sorry
for the mistake.

For pgsql-docs readers:

The issue is that the official documentation misleadingly omits the
existence of Postgresql-XC:

http://www.postgresql.org/docs/9.3/static/different-replication-solutions.html?

> Synchronous Multimaster Replication

*snip*

> PostgreSQL does not offer this type of replication (...)

Whereas the wiki says in http://wiki.postgresql.org/wiki/Postgres-XC

> Project Overview

*snip*

> Features of PG-XC include:

*snip*

> 2. Synchronous multi-master configuration

Now back to the original thread:

> Knowing the number of forks/projects based on Postgres, maintaining a
> list on a wiki list the one below is just easier for everybody:
> http://wiki.postgresql.org/wiki/Replication,_Clustering,_and_Connection_Pooling

That one doesn't even list PostgreSQL-XC.

For how man years has it been around now...
Can't even remember any more.

Instead it lists Postgres-R, which has been in koma for how long now...
Can't even remember any more.

BTW; No, I don't suffer from that brain disease that makes you lose your
memory (can't remember the name of it any more... ;-).

> Perhaps this list is not completely up-to-date,

To call that an understatement would be an euphemism.

It's simply misleading. And misleading potential users in search of
solutions for their needs is *bad* for the PostgreSQL project.

> but not adding that in the core documentation facilitates the work of
> core maintainers. It gives you all the information you need as well.

Guys, are you really not aware to *that* point how badly you shoot
yourself (and the PostgreSQL project as a whole) in the foot with that
single - wrong - phrase in the "official" documentation:

"PostgreSQL does not offer this type of replication"

Reading that phrase, the average O***** DBA looking for a cheaper
replacement will stop considering PostgreSQL and that's it. You're out
of business. They won't look any further.

Just stop arguing and put *one* *single* *phrase* in the official
documentation instead like:

"PostgreSQL itself does not provide this as a built-in functionality at
the current stage, but there is an open-source "fork" freely available
under the same license as PostgreSQL that does, for details read:

http://wiki.postgresql.org/wiki/Postgres-XC"

Is that really too much work? Heck, give me write-authority on the
documentation and I'll do it for you then. You've already wasted *way*
more brain bandwidth and precious time arguing why that phrase is *not*
there than it would take to put it there once for good. That's the kind
of pointy-haired dysfunctionality I'd expect from a managed corporation,
not from an open-source project.

In fact I would guess that given how closely PostgreSQL-XC follows the
releases of "pure" PostgreSQL and the fact that they use the same
license, at some stage it may be merged entirely.

Sincerely,

Wolfgang

On 12/12/2013 08:18 AM, Wolfgang Keller wrote:

>
>> 2. Synchronous multi-master configuration
>
> Now back to the original thread:
>
>> Knowing the number of forks/projects based on Postgres, maintaining a
>> list on a wiki list the one below is just easier for everybody:
>> http://wiki.postgresql.org/wiki/Replication,_Clustering,_and_Connection_Pooling
>
> That one doesn't even list PostgreSQL-XC.

Then update the page? It is a wiki.

>
> For how man years has it been around now...
> Can't even remember any more.
>
> Instead it lists Postgres-R, which has been in koma for how long now...
> Can't even remember any more.

Nope, it is actively developed and sponsored by Translattice.

>
>> Perhaps this list is not completely up-to-date,
>
> To call that an understatement would be an euphemism.
>
> It's simply misleading. And misleading potential users in search of
> solutions for their needs is *bad* for the PostgreSQL project.

Why are you arguing about this instead of just fixing it?

> Guys, are you really not aware to *that* point how badly you shoot
> yourself (and the PostgreSQL project as a whole) in the foot with that
> single - wrong - phrase in the "official" documentation:
>
> "PostgreSQL does not offer this type of replication"
>
> Reading that phrase, the average O***** DBA looking for a cheaper
> replacement will stop considering PostgreSQL and that's it. You're out
> of business. They won't look any further.

I agree.

> "PostgreSQL itself does not provide this as a built-in functionality at
> the current stage, but there is an open-source "fork" freely available
> under the same license as PostgreSQL that does, for details read:
>
> http://wiki.postgresql.org/wiki/Postgres-XC"
>
> Is that really too much work? Heck, give me write-authority on the
> documentation and I'll do it for you then. You've already wasted *way*

Submit a patch. Is that so hard? I don't understand why you are up in
arms about this.

Sincerely,

Joshua D. Drake


--
Command Prompt, Inc. - http://www.commandprompt.com/  509-416-6579
PostgreSQL Support, Training, Professional Services and Development
High Availability, Oracle Conversion, Postgres-XC, @cmdpromptinc
For my dreams of your image that blossoms
    a rose in the deeps of my heart. - W.B. Yeats

> > Instead it lists Postgres-R, which has been in koma for how long
> > now... Can't even remember any more.
>
> Nope, it is actively developed and sponsored by Translattice.

"Actively developed"?

http://www.postgres-r.org/ lists the last entry in the column "News" on
the right with a date of 2010-07-14.

http://git.postgres-r.org/ lists the "Last Change" to Postgres-R as "2
years ago".

http://www.postgres-r.org/downloads/ lists the last "Snapshot patch"
with a date from 2010-08-29.

The "Postgres-R Live-CD" has a date from 2006-07-04!

Sincerely,

Wolfgang