Re: Comments on Partial Locking -01

[Martin Bjorklund <mbj@tail-f.com>]

Hi,

"James Balestriere (jbalestr)" <jbalestr@cisco.com> wrote:
> To second some of Sharons items, the impression I got from the document
> was that
>  configuration seems to be the only way resources on the box are
> created.
> We need to also consider hot insertion and removal of hardware.
> E.g. I lock an ethernet interface to get get exclusivity on the
> interface
> but someone physically removes the card. Is my lock persisted ?
> If that card is pushed back in what happens ? If a similar card is put
> in ?
> (e.g I pull out a 4 port card and put in an 8 port card)

I don't think this is different from how the global lock works.  If I
lock the entire datastore, and someone removes a card, will the config
be altered?

> I also support Phil's concerns. Can I "read" a config when someone has
> locked
> a subtree ? The risk for me is I could extract a config which is in
> an intermediate state. We need some rules on reading such that I can
> only
> read a config which is in a know stable state (i.e. not locked)

The way it works right now is that you get what you ask for.  If you
lock /foo, then read from /bar and change /foo according to what you
just read, then you may end up in an inconsistent state.  If you want
to do this consistently, either grab the global lock, or lock /foo and
/bar.


/martin

--
to unsubscribe send a message to netconf-request@ops.ietf.org with
the word 'unsubscribe' in a single line as the message text body.
archive: <http://ops.ietf.org/lists/netconf/>