Experience of centralized workflow with NFS-mounted storage?

[David Timothy Strauss]

I realize that your SAN may dictate treating it as a black NFS box.

As long as you're bridging it through bzr+ssh, you'll have to find a redundancy mechanism for the bridge (say, Red Hat Cluster Services), and that mechanism could also serve storage redundancy purposes.

----- "David Timothy Strauss" <david at fourkitchens.com> wrote:

> I'd just like to understand why you wouldn't do one of the following:
> 
> (1) Run bzr as a smart server directly on the server storing the data,
> eliminating the NFS hop.
> 
> (2) Run the bzr client against the NFS server (as a dumb server) using
> an alternate means of access, like SFTP.
> 
> If possible, I'd prefer to answer the question "How do I achieve end
> goal X given environment Y?" rather than "How do I run Bazaar's X over
> NFS?" Treating the NFS server as a black box kind of limits our
> options.
> 
> David