[nsp] Setting router ospf passive-interface default

Tue Jun 15 13:37:06 EDT 2004

Hudson, I really thing you didn't get Steinar's point.

Put simply there is a flag attached to each interface. This flag is either ON or OFF. The flag name is "passive interface".
Currently with every configuration line you enter in the running config of an IOS device you get the changes reflected to the underlying configuration ("passive interface" flags) immediately. Even if the configuration is not typed in but loaded from a file/tftp, whatever.
Currently there is no way of switching from
router ospf 1
  passive-interface X
  passive-interface Y
to
router ospf 1
  passive-interface default
  no passive-interface Z
which both have the same underlying values for the "passive interface" flags of the interfaces
without flapping all the flags to OFF and then back to on.

What I think Steinar said is that there should be a (set of) command(s) allowing you to switch from one text configuration to the other atomically. 

E.g. it could be
Router# configure convert passive-interface-default ospf 1
Regardless of what the actual command is, this does not affect the way the routing protocol process works at all. The routing protocol process will "see" the same underlying interface flags, so it will not flap adjacencies.

Cheers,
Boyan Krosnov
for himself

> -----Original Message-----
> From: cisco-nsp-bounces at puck.nether.net 
> [mailto:cisco-nsp-bounces at puck.nether.net] On Behalf Of 
> Hudson Delbert J Contr 61 CS/SCBN
> Sent: 15. júní 2004 16:58
> To: 'sthaug at nethelp.no'; Dan.Wilson at ge.com
> Cc: cisco-nsp at puck.nether.net
> Subject: RE: [nsp] Setting router ospf passive-interface default
> 
> 
> Steinar,
> 
> you are correct in the assessment of the benefits of such a change
> from technical viewpoint,,,no argument there.
> 
> not speaking for anyone but myself but i think a general
> consensus as regards the impractical side is the logistics of 
> implementing the code changes and any anomalies that may arise
> across vendors and enterprise would be monstrous.
> 
> the resources required to accomplish this would not be a good
> business decision for the sake of the convenience of system and
> network admins. the cost to upgrade software, schedulng changes,
> manpower allocation, phased comm-out in business units.
> 
> again....ditto... on impractical...
> 
> btw, i'm a network engineer not a bean counter, so its really
> not a techie bashing, its just the reality of the business 
> environments we work in.
> 
> its all about the benjamins to the suits.
> 
> -----Original Message-----
> From: cisco-nsp-bounces at puck.nether.net
> [mailto:cisco-nsp-bounces at puck.nether.net]On Behalf Of 
> sthaug at nethelp.no
> Sent: Tuesday, June 15, 2004 9:25 AM
> To: Dan.Wilson at ge.com
> Cc: cisco-nsp at puck.nether.net
> Subject: RE: [nsp] Setting router ospf passive-interface default
> 
> 
> > Would you, therefore, increase the amount of code in Cisco's
> implementation
> > of OSPF so that when you make a global change to OSPF, it 
> first checks to
> > see if implementations of your change are currently configured?
> 
> Yes.
> 
> > Sounds like
> > a code nightmare, and I'm not sure I want it to attempt that for me.
> Maybe
> > as a series of questions asked upon implementation of the 
> command, i.e.
> > Passive installed on Serial 1/0.104, is this what you want to do?
> > 
> > Not too practical, eh?
> 
> I don't see why it's so impractical. In fact I would prefer this kind
> of behavior for all configuration changes where the *textual* change
> means that the *actual* configuration for the interface (or whatever)
> in question remains unchanged.
> 
> Steinar Haug, Nethelp consulting, sthaug at nethelp.no
> 
> > 
> > -----Original Message-----
> > From: sthaug at nethelp.no [mailto:sthaug at nethelp.no] 
> > Sent: Tuesday, June 15, 2004 11:14 AM
> > To: Dan.Wilson at ge.com
> > Cc: cisco-nsp at puck.nether.net
> > Subject: RE: [nsp] Setting router ospf passive-interface default
> > 
> > > Well, seriously.   You change whether OSPF talks on all 
> interfaces, and
> > > *don't* think it should run the spf algorithm?  You've 
> made a global
> > change
> > > to OSPF, and it *should* re run its whole table!
> > 
> > If the desired change is: all the interfaces that were 
> passive before the 
> > change should be passive afterwards, and similarly for the 
> all the active
> > interfaces - then no interface changes state with respect 
> to OSPF. Why
> > should it have to bounce adjacencies?
> > 
> > Steinar Haug, Nethelp consulting, sthaug at nethelp.no
> > 
> > 
> > > 
> > > The method you use to change the config isn't going to 
> matter, either.
> > > 
> > > 
> > > 
> > > -----Original Message-----
> > > From: John Kristoff [mailto:jtk at northwestern.edu] 
> > > Sent: Tuesday, June 15, 2004 10:44 AM
> > > To: cisco-nsp at puck.nether.net
> > > Subject: [nsp] Setting router ospf passive-interface default
> > > 
> > > Perhaps I've missed something, but is there a way to 
> implement on a
> > > router already running OSPF, where routing interfaces are 
> not going to
> > > be change the following:
> > > 
> > >   router ospf [process-id]
> > >    passive-interface default
> > > 
> > > Without having adjacency changes occur?  In my limited 
> testing, either
> > > with a quick copy and paste or tftp upload to the running 
> config, it
> > > seems that either will result in OSPF dropping all 
> interfaces causing
> > > neighbor adjacencies to have to be reestablished.  Since 
> in my example,
> > > the routing interfaces will have 'no passive-interface', 
> routing really
> > > doesn't change so it would be nice if reconvergence 
> didn't have to occur
> > > either.
> > > 
> > > John
> > > _______________________________________________
> > > cisco-nsp mailing list  cisco-nsp at puck.nether.net
> > > https://puck.nether.net/mailman/listinfo/cisco-nsp
> > > archive at http://puck.nether.net/pipermail/cisco-nsp/
> > > _______________________________________________
> > > cisco-nsp mailing list  cisco-nsp at puck.nether.net
> > > https://puck.nether.net/mailman/listinfo/cisco-nsp
> > > archive at http://puck.nether.net/pipermail/cisco-nsp/
> _______________________________________________
> cisco-nsp mailing list  cisco-nsp at puck.nether.net
> https://puck.nether.net/mailman/listinfo/cisco-nsp
> archive at http://puck.nether.net/pipermail/cisco-nsp/
> _______________________________________________
> cisco-nsp mailing list  cisco-nsp at puck.nether.net
> https://puck.nether.net/mailman/listinfo/cisco-nsp
> archive at http://puck.nether.net/pipermail/cisco-nsp/
> 
> 