nv-l
[Top] [All Lists]

RE: -- More -- RE: Facilitating discovery of IP Objects

To: nv-l@lists.tivoli.com
Subject: RE: -- More -- RE: Facilitating discovery of IP Objects
From: "Allison, Jason (JALLISON)" <JALLISON@arinc.com>
Date: Wed, 10 Oct 2001 09:39:04 -0400
Stephen,

Thanks for the email.  You make very good points.

The serial interfaces are not un-numbered.  I have dealt with this issue on
a different job with Digital UNIX 4.0D, Netview 5.1.1 and Lucent (Ascend)
PRI/BRI Boundry Routers.  That is not the case here.  The serial interfaces
have IPs.  We have hundreds and hundreds of these links, with more coming.
These links have evolved over time and are now a wopping 9.6KB/sec.
However, I just got done a lenghty discussion with my manager and I beleive
I have some insight into what this issue is.

The polling that is set to '8y' for these interfaces uses a subnet wildcard.
This means that any objects in this subnet will not get polled.  It is my
assesment that with them introducing this some months back, when they tried
to get the object added by pinging the serial interface (valid IP address),
it fell under that subnet and is not getting polled.

Now I feel I have less grasp on the actual problem, so my question is this,
what routines can I use to diagnose this problem in the NOC?  I am going to
try and reproduce this in the development lab, but I think I will see the
same results.

Thanks again for your time,
Jason

-----Original Message-----
From: Stephen Hochstetler [mailto:shochste@us.ibm.com]
Sent: Tuesday, October 09, 2001 6:08 PM
To: IBM NetView Discussion
Subject: Re: -- More -- RE: [NV-L] Facilitating discovery of IP Objects


Jason,

you mentioned having serial interfaces you could not poll/ping which is why
you were adding them manually.   Is this just un-numbered serial interfaces
on routers?   Other interfaces has IP addresses?    If this is the case,
the current NetView release handles those much better than before.   It
recognizes this fact and changes the status poll to the node to be only an
snmp poll since snmp will return the status of that interface.   NetView
will not try and ping those interfaces but will still put them on the map
properly.   You would need to loadhost with an interface that has an IP
address.   This is handled in 6.02.

If you are managing totally non-IP devices, I suggest you do it with an
SNMP proxy or at the least, don't place your items within the IPMAP.
Instead, at the root map, create a new object (beside IPMAP) and put all
your non-IP network devices inside that submap.    You can script this with
the sample of wteuiap6 from an older NetView redbook.

Thanks,
Stephen Hochstetler            shochste@us.ibm.com
ITSO Tivoli Coordinator - Austin
Office - 512-436-8564       FAX - 512-436-1991

ITSO redbooks at  http://www.redbooks.ibm.com


<Prev in Thread] Current Thread [Next in Thread>

Archive operated by Skills 1st Ltd

See also: The NetView Web