Re: Best high availability solution ?

> -----Original Message-----
> From: pgsql-general-owner(at)postgresql(dot)org
> [mailto:pgsql-general-owner(at)postgresql(dot)org] On Behalf Of
> Arnaud Lesauvage
> Sent: 31 May 2006 10:39
> To: Magnus Hagander
> Cc: pgsql-general(at)postgresql(dot)org
> Subject: Re: [GENERAL] Best high availability solution ?
>
> Otherwise, is the clustering service really necessary ?
> Aren't there simple "virtual IP addresses" solutions available ?
> I could have a single virtual IP for the 2 servers, and having an
> automatic failover if the master server is down ? The slave would
> never be accessed by clients directly (only after a failover) but
> would by synchronized with the master (slony-I).
>
> Does this kind of software exist in the windows world ?

Slony-I only exists for Windows in a not-even-beta state at the moment,
so even if you get things up and running using a virtual IP solution you
will be hand-holding Slony until it gets properly released.

The 'virtual IP' solution you are looking for is called Network Load
Balancing btw, and is on standard edition. I really don't see how you
would make it work safely though - it's designed more for load
distribution, and though you can weight rules towards one server,
there's no guarantee that it won't send a connection to the backup
server.

You would also need some way of automatically handling the Slony
failover which is definitely not recommended.

If I'm honest, I think your boss is going to be disappointed. You would
add a *lot* of complexity to the system to make it handle failures with
zero intervention, and that extra complexity is probably more likely to
go wrong than a single server. I'd spend your time and money on making
sure your raid & ups are good, that you are running on server grade
hardware with ECC RAM, and that you have good out of band management
facilities so even if you are away from the office you can connect via
VPN/modem or whatever and fix things.

Regards, Dave.