Re: Re : Very asynchrnous replication system

Oh, sorry, I misread your question and didn't realize the distributed
copies would need to make modifications back. AFAIK, the MusicBrainz
system is master/slave, not multi-master.

On Mon, 8 Oct 2007, Laurent ROCHE wrote:

> Ben,

As far as I can see MusicBrainz only does one way replication (which I already achieved).
Two ways replication (without re-sending the information you have just received) is a bit more complex.

I could not find much info on the MusicBrainz project anyway.

Thanks,
L(at)u
The Computing Froggy

----- Message d'origine ----
De : Ben <bench(at)silentmedia(dot)com>
À : Laurent ROCHE <laurent_roche(at)yahoo(dot)com>
Cc : pgsql-general(at)postgresql(dot)org
Envoyé le : Samedi, 6 Octobre 2007, 2h49mn 59s
Objet : Re: [GENERAL] Very asynchrnous replication system

The MusicBrainz project replicates in a way that sounds like it could fit your needs. It depends on a lot of perl, but if that's not a showstopper, then maybe adapting their replication scheme would work for you.
On Oct 5, 2007, at 5:25 PM, Laurent ROCHE wrote:

Hello,

I must replicate (or synchronise) data between disconnected postgreSQL databases ... hence a replication "very asynchronous"!

Application description
This an application to manage sales forces with a central application (and postgreSQL server) where everybody in the office can connect (this a web application: Tomcat with Spring 2+ Struts 2 + iBatis + PostgreSQL).
But there are also "little itinerant applications": the sales rep have the application on their laptop and a PosgreSQL server with a copy of the data (they are interested in ) from the main server database.
>From time to time, the sales reps connect to the internet and request a data synchronisation: they send their modification and receive the headquarters modification.
(no problem about data overlap: the data owner is clearly identified by the application framework)

My problem:
Does such an existing system (the replication one) exist ?
I search the web I did not find anything !
I have started to do something (with triggers, log table and table export - using DDL-Utils) ... but looking at the spreadth of the task layed out in front of me, I am thinking that I will be better off using or adapting an existing replication system

Have fun,
L(at)u
The Computing Froggy

Ne gardez plus qu'une seule adresse mail ! Copiez vos mails vers Yahoo! Mail

_____________________________________________________________________________
Ne gardez plus qu'une seule adresse mail ! Copiez vos mails vers Yahoo! Mail
>From pgsql-general-owner(at)postgresql(dot)org Mon Oct 8 13:41:29 2007
Received: from localhost (unknown [200.46.204.184])
by postgresql.org (Postfix) with ESMTP id D85DD9F95F8
for <pgsql-general-postgresql(dot)org(at)postgresql(dot)org>; Mon, 8 Oct 2007 10:55:26 -0300 (ADT)
Received: from postgresql.org ([200.46.204.71])
by localhost (mx1.hub.org [200.46.204.184]) (amavisd-maia, port 10024)
with ESMTP id 19652-07 for <pgsql-general-postgresql(dot)org(at)postgresql(dot)org>;
Mon, 8 Oct 2007 10:55:17 -0300 (ADT)
Received: from news.hub.org (news.hub.org [200.46.204.72])
by postgresql.org (Postfix) with ESMTP id 696A69F929D
for <pgsql-general(at)postgresql(dot)org>; Mon, 8 Oct 2007 10:55:26 -0300 (ADT)
Received: from news.hub.org (news.hub.org [200.46.204.72])
by news.hub.org (8.14.1/8.14.1) with ESMTP id l98DtKmn062463
for <pgsql-general(at)postgresql(dot)org>; Mon, 8 Oct 2007 10:55:20 -0300 (ADT)
(envelope-from news(at)news(dot)hub(dot)org)
Received: (from news(at)localhost)
by news.hub.org (8.14.1/8.14.1/Submit) id l98DY8XQ053117
for pgsql-general(at)postgresql(dot)org; Mon, 8 Oct 2007 10:34:08 -0300 (ADT)
(envelope-from news)
From: "Dave" <withheld(at)nospam(dot)thanks>
X-Newsgroups: pgsql.general
References: <bf242$4706a213$453dce02$24167(at)FUSE(dot)NET> <4707FFF6(dot)7010402(at)theendofthetunnel(dot)de>
Subject: Re: Foreign Key for PHP serialized data - possible?
Date: Mon, 8 Oct 2007 09:33:59 -0400
X-Priority: 3
X-MSMail-Priority: Normal
X-Newsreader: Microsoft Outlook Express 6.00.2900.3138
X-RFC2646: Format=Flowed; Original
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.3138
Message-ID: <1d644$470a31d5$453dce02$11656(at)FUSE(dot)NET>
X-Complaints-To: abuse(at)usenetserver(dot)com
Organization: UseNetServer.com
Lines: 59
To: pgsql-general(at)postgresql(dot)org
X-Virus-Scanned: Maia Mailguard 1.0.1
X-Archive-Number: 200710/359
X-Sequence-Number: 121651

"Hannes Dorbath" <light(at)theendofthetunnel(dot)de> wrote in message
news:4707FFF6(dot)7010402(at)theendofthetunnel(dot)de(dot)(dot)(dot)
> Dave wrote:
>> What I'm trying to do is to create a foreign key on tb1_column so that if
>> a
>> number in tb2_column changes or gets deleted, it cascades to the
>> appropriate
>> segment of the serialized data.
>> e.g. if an id of Oranges changes from '5' to '24', the tb1_column rows
>> will
>> get changed in the above example row to:
>> a:5:{i:0;s:1:"9";i:1;s:2:"24";i:2;s:2:"11";i:3;s:2:"100";i:4;s:2:"10";}
>
> No offense, but that sounds like sick application design. Anyway, a
> plphp trigger can do it:
>
> http://www.commandprompt.com/community/plphp
>
>
> --
> Best regards,
> Hannes Dorbath

> No offense, but that sounds like sick application design
Offense taken :)
Well, you don't know the whole picture, so I can see why this would look bad
to you. That's why asked any other ideas.
The problem I need to solve is this:

I have main table with lets say:

11 | Pears
100 | Plums
9 | Apples
5 | Oranges
10 | Cranberries

Now, I need to create another table with options for each of the above. So
for the apples:
ID | colors_available | favourite_color |
kind | favourite kind | Other similar options ...
9 | red, green, yellow, ...infinity | red | Granny smith, Golden delicious,
... infinity | Granny smith | Other similar values

What I'm trying to do is to keep the options for each froot in one row, as
opposed to creating separate table for colors, kind, etc.
I realize that I could create separate tables for colors, kind, etc, and
separate tables for favourite colors and favourite kind, but this would
involve accessing four different tables in this case. I guess, I can always
set up views with joins, but could not come up with a good way to query this
in a one returned row, e.g.:
ID | Name | colors | fav_col | kind
| fav_kind | etc.
11 | Apples | red, green, etc. | red | Granny smith, Golden
delicious.. | Granny smith | etc.