I understand. I also understand that this is a limitation of the
design chosen for RFI. There is a trade-off here. The design
of the RFI component is, in my opinion, quite elegant in its
simplicity with the additional benefit of requiring no administration
and very little overhead. And it is free. It does require, however,
that you actually manage all of the connecting devices in the
path to the target. Standalone devices are not really part of a
network, in Netview's view of things. Alternative designs, such
as those implemented by add-on or standalone correlation products,
have, I imagine, other trade-offs. The RFI is not suitable for all customer
environments. At some sites I turn it off after a while, and at others
it does the job nicely. You have to make that determination yourself.
This is just my opinion, for what it is worth.
Cordially,
Leslie A. Clark
IBM Global Services - Systems Mgmt & Networking
Detroit
"Andreas.Poehlmann" <ANDREAS.POEHLMANN@Allianz.de>@tkg.com on 07/05/2001
10:27:37 AM
Please respond to IBM NetView Discussion <nv-l@tkg.com>
Sent by: owner-nv-l@tkg.com
To: " - *nv-l@tkg.com" <nv-l@tkg.com>
cc:
Subject: [NV-L] RFI problem
Hi Leslie,
we opened a PMR and after some heavy discussions we were asked to open
an "enhancement request" - RFI works as designed and this is a very
special
case nobody else has had up to now... - and that makes me wonder!
have a look at this more general case:
If discovery is done via seedfile you can't be sure, that the topology is
complete.
Running RFI, it is possible, that routers are going down without noticing
that in the map.
E.g.:
isolated router w.r.t. topology.
If this router goes down completely it turns white instead of red - and
this is not propagated.
To our opinion RFI should be able to handle these cases correctly - from a
management
point of view.
regards
Andy Pöhlmann
AGIS - Allianz Gesellschaft für Informatik Service
---------------------- Weitergeleitet von Andreas Poehlmann/MUC/AGIS on
05.07.2001 15:33 ---------------------------
lclark@us.ibm.com@tkg.com on 22.05.2001 17:53:00
Bitte antworten an nv-l@tkg.com
Gesendet von: owner-nv-l@tkg.com
An: nv-l@tkg.com
Kopie:
Thema: Re: Aw: Re: [NV-L] RFI problem
Andreas, if you are on 6.0.2, I would take this to Support and have them
dig into it.
I believe that the -n option should make it do just what you want it to
do.
Cordially,
Leslie A. Clark
IBM Global Services - Systems Mgmt & Networking
Detroit
"Andreas.Poehlmann" <ANDREAS.POEHLMANN@Allianz.de>@tkg.com on 05/22/2001
09:44:16 AM
Please respond to IBM NetView Discussion <nv-l@tkg.com>
Sent by: owner-nv-l@tkg.com
To: " - *nv-l@tkg.com" <nv-l@tkg.com>
cc:
Subject: Aw: Re: [NV-L] RFI problem
Hi,
I already read that and tried to use the -n option, but I couldn't get the
result I wanted....
Even changing the handling of the ambiguous case didn't help.
The scenario:
one router, having four interfaces in four subnets whith no other device -
completely stand alone.
Three of the four interfaces I can't ping (no route) so I unmanaged them.
The fourth interface is green as long as I can ping it.
-> so far everything is fine.
If I loose the connection to this interface, the interface turns to red,
the router turns to red -
AND THEN rfi starts to work and turns the interface and the node to white
-> and that's the problem! I'd like to have the router in red as white is
not propagated....
I coudn't find any setup that realizes this.
ANY help appreciated - as I think RFI is of great use, but I have to
display the correct status
of this router....
regards
Andy Pöhlmann
AGIS München
lclark@us.ibm.com@tkg.com on 22.05.2001 13:58:46
Bitte antworten an nv-l@tkg.com
Gesendet von: owner-nv-l@tkg.com
An: nv-l@tkg.com
Kopie:
Thema: Re: [NV-L] RFI problem
I have seen this, and I think it boils down to a special case of what is
referred to in the documentation as 'the Ambiguous Case'. The -n option
is provided for netmon to control how this case is handled. See the
/usr/OV/doc/RouterFaultIsolation.html document. The -n flag is one
you have to add manually to netmon.lrf. It changes the handling of the
ambiguous case from unreachable to down. I have not tried it. Want to
let us know if it works for this case?
Cordially,
Leslie A. Clark
IBM Global Services - Systems Mgmt & Networking
Detroit
"Andreas.Poehlmann" <ANDREAS.POEHLMANN@Allianz.de>@tkg.com on 05/10/2001
06:50:40 AM
Please respond to IBM NetView Discussion <nv-l@tkg.com>
Sent by: owner-nv-l@tkg.com
To: " - *nv-l@tkg.com" <nv-l@tkg.com>
cc:
Subject: [NV-L] RFI problem
Hi out there,
we're fighting a RFI problem here...
Situation: We monitor a router of which we
- can only ping one of four interfaces and
- none of the interfaces is in a subnet with an other monitored device.
=> This router is logical "not connected" to our network, but we can ping
this interface and
we have SNMP access.
The unpingable interfaces are set to unmanaged.
In case the last pingable interface goes down, RFI assumes the
router is "unreachble", turns it to white and this status change is not
propagated.
I'd like to see this router turning red instead of white!
Having read the manpages and description of the RFI I assume that there is
no
way - but is there realy no way? Any experiences with situations like
this?
Any help appreciated
regards
Andreas Pöhlmann
_________________________________________________________________________
NV-L List information and Archives: http://www.tkg.com/nv-l
_________________________________________________________________________
NV-L List information and Archives: http://www.tkg.com/nv-l
_________________________________________________________________________
NV-L List information and Archives: http://www.tkg.com/nv-l
_________________________________________________________________________
NV-L List information and Archives: http://www.tkg.com/nv-l
_________________________________________________________________________
NV-L List information and Archives: http://www.tkg.com/nv-l
|