Re: HA solution

Thanks for your support.

We have power full HP servers with lots of CPU cores, I/O bandwidth and memory too.

Actually I will give you the environment details, which will help you to understand. 

It is a huge set-up where we have a DC & DR. There will be lots of daily edit and read hits. Also there would be lots of read hits from reporting perspective too. Therefore, I thought of to keep one node for edit and other one for read (OLTP vs REPORTING) kind of structure for DC & if it goes down DR will take care. However in your suggested structure there will be only one active node at a time in DC. Even standby database would be in recovery mode.

Initially I thought of the set-up slony replication for (OLTP vs REPORTING) & pgpool replication for (OLTP vs OLTP) but there will be very huge GIS database size & they agreed only on cluster set-up.

I really need thoughts/comments/help from experts.


--
Thanks & Regards
Dhaval Jaiswal




-----Original Message-----
From: pgsql-general-owner@postgresql.org on behalf of John R Pierce
Sent: Sat 1/15/2011 12:23 PM
To: pgsql-general@postgresql.org
Subject: Re: [GENERAL] HA solution

On 01/14/11 9:47 PM, Jaiswal Dhaval Sudhirkumar wrote:
>
> Hi,
>
> I am looking for active-active clustering solution.
>

best of luck.    active-active is fraught with complex hard-to-solve
problems.



> I have one SAN box and two separate NODES, where I need to create
> active-active cluster. My data directory would be one and mounted to
> the SAN box for both the nodes. (There will be one sharable data
> directory for both the nodes) So the query which will come to the load
> balancer (pgpool) it will route to the node which has a less load.
> However, it will use the same data directory. It is nothing but the
> RAC kind of structure.  Now, my question is.
>
> 1)      Is it possible above implementation in PostgreSQL?
>

No

> 2)      Has someone implemented cluster in their production environment?
>

Oracle RAC supports configurations like that, however in our limited
testing, we came to the conclusion that you needed at least 4 nodes
before it was faster than 1 simple database server.   And of course,
these 4 nodes required 4 times the Oracle licensing, which can get
really expensive really fast.



Postgres can scale very nicely vertically with the right hardware (lots
of CPU cores, lots of memory, lots of IO bandwidth).    The standard
database server configuration for the hardware you described is an
active/standby cluster, where all requests are processed by the active
server, and the standby server only takes over if the active server has
failed.



--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general

The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. 
Any review, re-transmission, dissemination or other use of or taking of any action in reliance upon,this information by persons or entities other than the intended recipient is prohibited. 
If you received this in error, please contact the sender and delete the material from your computer. 
Microland takes all reasonable steps to ensure that its electronic communications are free from viruses. 
However, given Internet accessibility, the Company cannot accept liability for any virus introduced by this e-mail or any attachment and you are advised to use up-to-date virus checking software.