First of all, I would like to thank everyone for their input. You all have been incredibly helpful in helping me work through my ignorance.
What exactly does it mean when you say CephFS is not "production ready"? To me, this typically indicates a product that still has business crippling bugs. As I am attempting to build a business based on open source technologies, I have a tendency to shy away from products labeled as "not ready for production". It makes more sense to stay in the Ceph family than to add another layer of complexity.
What I'm looking to accomplish is a virtualization platform based on open source technologies - on a budget... My software stack is basically, Ceph, KVM, OpenVSwitch, Chef, Perceus, and OpenNebula. I've been working with OpenNebula because I like the notion of "Virtual Datacenters" as opposed to "just" virtual machines. What I would like to accomplish with Ceph is a centralized storage space for my Hypervisors in addition to providing storage for the virtual machines and and any other physical machines in the cluster as they need it. The way OpenNebula works is (assume /var/lib/one is ~) it creates virtual machines in ~/datastores/0/<virtualmachine id>/ (0 is the default datastore id). "Non-persistent" storage, such as the base virtual machine image, is copied to ~/datastores/0/<virtualmachine id>/disk.0 for instance. Each Hypervisor would then have a unique ~/datastores/. I'm trying to avoid that. My hypervisors have small disks. Maybe it makes sense to mount each ~/datastores as a unique RBD, at least in my mind it using the hypervisors to copy data from one rbd to another seems like it would be slow.
CephFS loks like it might do exactly what I need, bit I'm certainly open to any suggestions.
Thanks for all your help,
Jon A
On Wed, May 29, 2013 at 12:52 PM, Jens Kristian Søgaard <jens@xxxxxxxxxxxxxxxxxxxx> wrote:
Hi,I think it would be wise if you take a step back and explain what you want to accomplish.
Maybe I'm asking the question wrong because I keep getting the same answer.
If you want to mount a file system on multiple hosts simultaneously, you need a distributed file system such as for example GFS2. This is just like with traditional disk systems where you connect two or more computers to the same storage.
However, if all you want is to be able to store virtual machine images on Ceph and be able to migrate virtual machine from one server to the other using "shared storage" - you do not need to go through all that.
--
Jens Kristian Søgaard, Mermaid Consulting ApS,
jens@xxxxxxxxxxxxxxxxxxxx,
http://www.mermaidconsulting.com/
_______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
