The MPLS WG Archive

[Sasha Vainshtein <Sasha@AXERRA.com>]

Thomas and Venkata,

I may be entirely wrong, but it seems to me that 
Venkata in his question did not ever refer to xxxOperStatus. 

AFAIK, xxx RowStatus can be used for disabling an existing 
managed object by setting its value to 'notInService';
this action can be reversed by setting its value to 'active'. 
As a consequence, some MIBs consider xxxAdminStatus as a read-only
attribute whose "up" and "down" values are derived from that of 
the xxxRowStatus which is, of course, read-write, and which, for an 
existing object, allows to observe only these two values ('notReady'
corresponding to the situation when creation of the object is incomplete).

The LSR MIB provides such an example with mplsXCAdminStatus and a
comment that one more value ('testing') is not required
(draft-ietf-mpls-lsr-mib-07.txt).

The MPLS-TE MIB which defined mplsTunnelAdminStatus 
(draft-ietf-mpls-te-mib-02.txt) allows it to be set to 'testing', 
too. I assume (and I may be wrong here; I am not a MIB expert) that:

o Setting mplsTunnelRowStatus to 'active' results in the 
mplsTunnelAdminStatus becoming 'up' and vice versa
o Setting mplsTunnelRowStatus to 'notInService' results in the 
mplsTunnelAdminStatus becoming 'down' and vice versa
o Setting mplsTunnelAdminStatus to 'testing' results in
mplsTunnelRowStatus becoming 'notInService'
o Similar rules apply to any managed object that which admits 'testing'
as a valid value for its AdminStatus.

Is this correct?

With best regards,
                                   Sasha Vainshtein
email:     sasha@axerra.com <mailto:sasha@axerra.com> 
tel:       +972-3-7659993 (office)
           +972-8-9254948 (res.)
           +972-58-674833 (cell.)
 


> -----Original Message-----
> From: Thomas Nadeau [mailto:tnadeau@cisco.com]
> Sent: Monday, December 31, 2001 4:38 PM
> To: Venkata Prasad Piriya
> Cc: mpls@UU.NET; Shashi
> Subject: Re: Regarding mplsTunnelAdminStatus
> 
> 
> 
> >What is difference between mplsTunnelAdminStatus and  
> mplsTunnelRowStatus??
> 
>          One allows for administrative control by the 
> operator, and one
> shows the actual operational status of the tunnel. In short, 
> the administrative
> status is the desired status of the tunnel while the operational
> status shows the actual status of the tunnel at the time of 
> query. When
> the desired status is set to a particular value, the 
> operational status
> should *eventually* reach that same state if possible (see
> below).
> 
> >Where exactly mplsTunnelAdminStatus  come into the picture for an 
> >administrator??
> 
>          It is used to control the tunnel, and it allows the 
> agent to show the
> actual status of the tunnel.
> 
> >My assumption is as follows, Please correct me if I am wrong.
> >
> >Initially User creates a tunnel using Row Status variable with the 
> >value  CREATEANDGO.
> >This tunnel will be brought up once row successfully created .
> 
>          Not exactly. While the tunnel can be created successfully via
> the row status operation and its administrative (desired) status might
> be up, the tunnel's operational status might not be because RSVP
> could not signal an LSP for the tunnel using the constraint parameters
> you specified.  Only when this happens should the tunnel's operational
> status be up.
> 
> >Here after if user wants to
> >modify any object in tunnel table he has to set 
> mplsTunnelAdminStatus  to 
> >down and set
> >the desired object. the changes will be reflected only on 
> changing the 
> >Admin status
> >back to UP.
> 
>          Yes.  This would also cause any signaled LSPs for 
> that tunnel to be
> torn-down.
> 
> >One more question regarding other tables in TE MIB i.e 
> Resource table and 
> >Hop Table.
> >If user modify some object in HOP table or Resource table 
> will it trigger 
> >modification of
> >Tunnel's which are presently using these resources or 
> Hop's???  In this 
> >case what role this
> >mplsTunnelAdminStatus   will play??
> 
>          No, changes to these tables should not affect a 
> tunnel that has 
> admin and oper
> 'up' at the time. In fact, the agent should not allow 
> modification to any 
> corresponding
> hop/resource entries for tunnels that have their 
> operational/administrative 
> states
> set to up.  Remember, the agent must coordinate the related 
> rows in tables 
> to make
> sure that they are consistent.  Allowing modification while 
> the tunnel is 
> running
> would not accomplish this.
> 
>          --Tom
> 
> 
> >Please clarify above queries...
> >
> >Thanks&Regards
> >Venkata Prasad P.
> >
> >
> >
> >
> 
> 
> 
> --------------------------------------------------------------
> ----------
> Mathematics is the supreme nostalgia of our time. 
>