RFC: using ceph as a backend for an NFSv4.1 pNFS server

[Bram Vandoren]

Hi Rick,

On 04/26/2014 09:47 PM, Rick Macklem wrote:
> Any other comments w.r.t. this would be appreciated, including
> generic stuff like "we couldn't care less about pNFS" or technical
> details/opinions.

I have some experience with Gluster (I suspect Ceph is similar). I don't 
think it's very useful:
- These file systems are implemented in user space (except for the FUSE 
glue). Gluster includes a (user mode) NFS server itself so it can skip 
the kernel VFS layer and FUSE glue to provide NFS.
- Gluster has it own network protocol. You can use this protocol to 
mount the volume on the client instead of NFS.
- You can use a native API to access these filesystems instead of POSIX.

The amount of people waiting for yet another method to access their 
cluster file system is probably limited.

Thanks for your work on the FreeBSD NFS server.

Cheers,
Bram.