Adile Abbadi wrote:
> Hi all,
>
> Not sure if this has been asked before - it probably has been - but I
> figure its probably just easier if I ask so my question is this:
>
> How easy is it - or rather is it possible to create multiple instances
> of the same db on multiple machines?
>
> We are currently running a very highly used postgres db (Ver 7.2) that
> eats up a lot of cpu time when its heavily used. We recently bought
> some very powerful equipment to accommodate this high use, (currently
> its on a single CPU, 1GB of Ram SCSI server - upgraded to a dual CPU,
> 4GB of ram SCSI server) but my fear is over time this issue is going
> to haunt me again and hardware can only be upgraded so much.
>
> So is it possible to do some level of load balancing in postgres - I
> would love a situation where I could have multiple machines running
> multiple instances of the same db (that are all synced in real-time),
> where there is some smart level of load balancing happening.
>
> So if it is possible is there some instruction info out there to help
> me along my way. Any help or insight would be greatly appreciated.
>
> All the best
>
> Adile
>
Hello,
I read an interesting paper on this a while ago, you can find it here:
http://www.cnds.jhu.edu/pub/papers/cnds-2003-3.pdf
It's called "On the Performance of Consistent Wide-Area Database
Replication", and presents an analysis of a Postgres-based replication
system.
Unfortunately, the implementation wasn't made available, but the ideas
may be worth investigating for someone with more free time than myself.
-Mark.
