Re: replication questions - Mailing list pgsql-hackers

From Doug Needham
Subject Re: replication questions
Date
Msg-id 200205030515.g435FMO19000@pbo.net
Whole thread Raw
In response to Re: replication questions  (Darren Johnson <darren@up.hrcoxmail.com>)
List pgsql-hackers
On Thursday 02 May 2002 07:54 pm, Darren Johnson wrote:
> >databases. I am trying to find out more information about how to do
> > automatic replication with postgresql.
>
> We did some research on this several months ago, and published the
> results here
>
> http://gborg.postgresql.org/genpage?replication_research
>
> >    My questions are :
> >
> >    1) Is this the correct forum to ask this question?
>
> This is probably better suited for general, but its a topic I'm
> interested in.
>
> >    2) What is the "recomended" replication solution project?
>
> IMHO, Postres-R has the potential to be great solution, but I'm sure
> others have different
> needs. ;-)
>
> >    3) How are others doing replication?
>
> If you asking for approaches, I'd say the majority are master/slave
> asynchronous
> with either triggers or transaction logs.
>
>
> I would ask what are you trying to solve with replication.  Do you want
> to be able
> update all systems in the replica? 
yes.How much bandwidth do you have
> between the
> servers. 
We will have plenty. 
> Do the systems need to be identical at all times?  
Pretty much. 
>What is
> your time frame?
Possibly as soon as six-nine months. 

Our replication solution is shooting for the moon. 
The scenario pitched to me (I'm the DBA for this application) is potentially 
having multiple web applications accessing multiple databases and having 
replication keep them all in sync. 
I have to decide if in this scenario we have the multiple front end's connect 
to a single master database and then have replication move the data to all of 
the backend databases and then notify the backend processes or to have a 
different scenario. 

The specific scenario that marketing/management has given me is to have two 
instances of the same front-end located in different parts of the country, 
each updating a local database and the database ensures that the updates from 
both locations get to the "peer-database".  So end user in location A is able 
to see all updates done to database A and database B for the data that they 
have premission for. 

I know this sounds confusing, it is. I have been doing DBA things for quite a 
while, but I am a litle new to the replication thing. 


Thanks, 

Doug
>
> Depending on how you answer questions like these, your "recommendation"
> will change.
>
> Good luck,
>
> Darren
>
>
>
>
>
> ---------------------------(end of broadcast)---------------------------
> TIP 4: Don't 'kill -9' the postmaster


pgsql-hackers by date:

Previous
From: Lincoln Yeoh
Date:
Subject: Re: a vulnerability in PostgreSQL
Next
From: "Christopher Kings-Lynne"
Date:
Subject: 3 digit year problem