The EMC storage is not only residing in the DC for providing storage to one server. At both sites, we have separate EMX SAN storage available. The bricks will be assigned from this EMC SAN storage to each of the gluster nodes. The reason I need gluster is to replicate the data across our DC/DR sites which inturn is connected by a 2Gig data dedicated fiber connection. That?s why I'm repeatedly asking. "What should be the memory/cpu sizing required for these gluster nodes . ?? Thanks & Regards, Bobby Jacob -----Original Message----- From: James [mailto:purpleidea at gmail.com] Sent: Sunday, September 01, 2013 7:41 PM To: Bobby Jacob Cc: Whit Blauvelt; gluster-users at gluster.org Subject: Re: Hardware Sizing Requirments On Sun, Sep 1, 2013 at 12:48 AM, Bobby Jacob <bobby.jacob at alshaya.com> wrote: > Both the gluster nodes are virtual machines. We have EMC storage available on both the sites. The LUN's from EMC storage is presented to both these nodes. I was thinking, we present LUN's from either storage to the gluster nodes present on either sites. These LUN's will be then mounted on the gluster nodes (bricks) using xfs filesystem. Then I create a gluster volume using them and mount that volume on the application servers (2) that's present on either site. Both these sites are connected using a dedicated 2GBps dark fibre. I'm not sure this makes a whole lot of sense... I don't know much about your EMC storage setup, but it seems to me you're just trying to pass it through gluster for no particular reason. If you've got some sort of EMC san, I think it makes sense to use that directly. You can think of gluster as a SAN replacement, not a layer on top of an existing one. But this is all just speculation based on what I'm assuming.
