Re: Replication using WAL files - Mailing list pgsql-general

From Josh Harrison
Subject Re: Replication using WAL files
Date
Msg-id 8d89ea1d0712070949g5571a929k786604df4f19ffd7@mail.gmail.com
Whole thread Raw
In response to Re: Replication using WAL files  ("Josh Harrison" <joshques@gmail.com>)
Responses Re: Replication using WAL files
List pgsql-general
 



On 12/7/07, Josh Harrison < joshques@gmail.com> wrote:
> I have 2 servers on which I need to have data replicated. The master server
> should serve for read/write queries and the 2nd server is used mainly for
> research queries(read-only queries) and so it doesn't have to be up-to-date.
...
> Is it possible to implement this thro' WAL shipping?

No. At the moment [1] the WAL shipping system does not permit you to
query the slave. There are a few options:

 http://pgfoundry.org/search/?type_of_search=soft&words=replication&Search=Search

As well as this:

 http://bucardo.org/

[1] But someone is working on this for 8.4. Don't hold your breath, though.


Thanks for the info. Just to clarify, So at the moment for WAL shipping to work the 2nd server should  only be a stand-by server and not a slave(cannot be queried) ?
In your experience, which other replication system (slony,pgpool etc) is better suited for my requirement?

Thanks again
josh


Hi,
I read all the documentations and got a bit confused.
1. I have a primary server which archives the WAL files to the secondary servers's directory(some directory).
2. The recovery.conf   in the secondary server is set to copy these files from this directory.
3. I take a base backup and feed it to the secondary server and start the secondary server.
4. The secondary server now contains all the base backuped data.
5. I set up a trigger (as in this demo http://archives.postgresql.org/sydpug/2006-10/msg00001.php)
  I created/inserted some table/data in the primary server, 'touch' trigger in sec server and  the archived files are reflected in the  sec server's database now and the sec database is up (recovery.conf becomes recovery.done) and I can query the secondary as normal.


My question is that will any future changes in the primary server gets reflected in the secondary server or is it just a 1-time show? If no, is it possible to have a set-up like that (secondary server getting asychronous updates from master thro'  WAL files......I guess "Hot standby using WAL files") ?

Sorry if my question sounds silly.
Thanks for all your help
josh

   


My question is

pgsql-general by date:

Previous
From: Bob Pawley
Date:
Subject: pg_dump
Next
From: Robert Treat
Date:
Subject: Re: Re-partitioning huge schema