Re: [netmod] kw comments on draft-voit-netmod-yang-mount-requirements

[Juergen Schoenwaelder <[e-mail]>]

Is the essence of your long text that the difference of alias mount
and local peer mount is:

alias mount = read-write
peer mount = read-only

Is it documented somewhere how writing (including locking) works with
alias mount, how NACM works with alias mount, ...?

/js

On Fri, Apr 08, 2016 at 07:02:20PM +0000, Alexander Clemm (alex) wrote:
> Hello Juergen
> 
> to your question: 
> "> So why do we need 'local data mount' (aka 'alias mount')? Which 
> > problem does it solve that 'peer mounting' localhost does not solve?"
> 
> Yes, you could apply peer mounting to a local host as a special case.  Certainly there is nothing that should prevent that.  However, the distinction is still relevant.  For one, in the peer mounting case, the remote host is considered the authoritative owner, the mounting host is _not_ the authoritative owner.  In alias mount, since it is the same owner, it has authority.  
> 
> The fact that in peer mount the local host has no authority means that the use cases primarily targeted involve cases that require visibility of data and support for a data retrieval, read-only view.  One of the concerns that had been raised earlier were the ramifications that transactional behavior and locking etc would have in a distributed scenario.  Given that alias-mount is local, things are simplified and use cases that go beyond pure providing visibility easier to support.  So, alias mounting may be an important special case of peer mounting, but also a simple case.  This is a fact that we would like to leverage.  
> 
> In practical terms, alias mount and peer mount can look very similar.  Peer mount involves an additional parameter (to identify the target system), with corresponding additional mountpoint management ramifications.  So, peer mount can in that sense build on top of, or extend, alias mount.  
> 
> --- Alex
> 
> -----Original Message-----
> From: Eric Voit (evoit) 
> Sent: Friday, April 08, 2016 6:33 AM
> To: Juergen Schoenwaelder <[e-mail]>
> Cc: Martin Bjorklund <[e-mail]>; [e-mail]; Alexander Clemm (alex) <[e-mail]>; [e-mail]
> Subject: RE: [netmod] kw comments on draft-voit-netmod-yang-mount-requirements
> 
> > From: Juergen Schoenwaelder, April 08, 2016 2:58 AM
> > 
> > On Fri, Apr 08, 2016 at 02:20:19AM +0000, Eric Voit (evoit) wrote:
> > > My thinking matches more closely to what Kent lays out above:
> > >
> > > schema-mount
> > > data-mount
> > > remote data mount   (a.k.a. peer mount)
> > > local data mount        (a.k.a. alias mount)
> > >
> > > The value in the term "yang mount" is that it provides a conceptual 
> > > umbrella to
> > ensure a cohesive syntax across the four valid permutations of the 
> > above.  The term itself was never intended to be implementable.
> > >
> > 
> > So why do we need 'local data mount' (aka 'alias mount')? Which 
> > problem does it solve that 'peer mounting' localhost does not solve?
> 
> To me it comes down to ease of use for the developer. If the system only allows mounting localhost, why require that parameter in the syntax?   That parameter shouldn't even be available/visible for entry. 
> 
> Eric
> 
> > /js
> > 
> > --
> > Juergen Schoenwaelder           Jacobs University Bremen gGmbH
> > Phone: [phone]         Campus Ring 1 | 28759 Bremen | Germany
> > Fax:   [phone]         <http://www.jacobs-university.de/>

-- 
Juergen Schoenwaelder           Jacobs University Bremen gGmbH
Phone: [phone]         Campus Ring 1 | 28759 Bremen | Germany
Fax:   [phone]         <http://www.jacobs-university.de/>