[rescue] clustered file storage

[Patrick Giagnocavo]

On Tue, Jul 31, 2001 at 04:44:31PM -0400, joshua d boyd wrote:
> On Tue, Jul 31, 2001 at 08:00:39PM +0000, koyote at koyote.cx wrote:
> > 
> > It depends. One company's model is to have scaleable access to storage 
> > :and: cpu power based on needs (response time of 10 minutes to add or 
> > remove).
> > 
> > I lot of them are using EMC hardware an pathlights for DAS that is 
> > moveable in near realtime (server failure results in DAS being moved to 
> > another server and rebooting within 1min!)
> 
> So, let me see if I foloow you.  One company wants to be able to increase
> storage and CPU power to Oracle (or what ever else).  So, they have oracle
> using data stored on a file server instead of locally, and they have
> oracle running clustered.  Thus they can add more disks to the file
> server, or they can drop another box on the network and add it to the
> oracle cluster.

You have to understand that these things are more than just big, fast
filesystems on a dedicated box.

They offer other things, like snapshots of a file system, needed to do hot
backup of Oracle (though I think Oracle added this feature later).

They automatically tune themselves for the work being done (ie it will tune
itself differently for a decision support system DB than for an OLTP
system).  They have growable file systems.  They have redundancy out the
wazoo.  More than one system can connect to the same set of drives.

Etc, etc.  Many of them monitor each drive and will phone home if the drive
may be prone to failing in the near future.

Citibank may have a system that costs them $1million per minute of downtime.
This is the kind of system that can really use these expensive beasties.

Cordially

Patrick