[ec2] disk partition configurations

[Darren Govoni]

I think I'm going to have to put the data on an EBS on a separate server
where other transient instances can hit a URL to do lookups.

The problem with making a snapshot EBS for each and mounting (although
it would work), involves manual steps and I need my server instances to
startup and run without human intervention.

Downloading from S3 is also feasible, but doing that on 80GB for each
server at each reboot and dumping it onto the /mnt would be expensive
and time consuming.

But thanks for the tips gang. Maybe in the future, there is such as
thing as a partition "bundle" that can be spawned and attached (by an
EC2 instance) logically without manual steps.

On Sat, 2009-06-13 at 20:03 +0200, Thomas Hervé wrote:
> Le samedi 13 juin 2009 à 19:50 +0200, Thomas Butter a écrit :
> > Hi Darren,
> > 
> > the easiest approach is to put your data in a tgz on S3 and fetch it
> > to a local partition in your startup scripts. Local partitions are  
> > faster
> 
> This is wrong. They may more stable on small instances, but EBS
> definitely performs better on larger instances.
> 
> >  and cheaper.
> 
> Well, yeah :).
> 
> -- 
> Thomas
> 
>