Please note - My comments about SNMP going to wrong interface has been
determined to *NOT* be a name resolution issue. Its being looked at by level
2 support right now.
-----Original Message-----
From: Leslie Clark [mailto:lclark@us.ibm.com]
Sent: Friday, September 07, 2001 9:00 AM
To: NV-L@tkg.com
Cc: Peter Raig
Subject: RE: [NV-L] Demand poll times out, but SNMP walk works.
Yeah, that's what it looks like to me, too, except that the
'Monitor...' functions work, and those do real name resolution
before checking xnmsnmpconf. Since it is an SNMP-polled
node, I suspect that the node was discovered by an
interface that is not up or something. For a node that is
not polled via SNMP, it would normally try all interfaces
before giving up (I think it used to work that way, anyway). It
might work differently for snmp-polled nodes.
At any rate, there is NO penalty for sharing ideas on the
listserver, Peter. I do it all the time. We are even allowed
to be wrong - someone will correct us, count on it!
Take note: the building concensus is that it probably
has something to do with name resolution....
Cordially,
Leslie A. Clark
IBM Global Services - Systems Mgmt & Networking
(248) 552-4968 Voicemail, Fax, Pager
============================================================================
===============
Leslie, I don't want to respond to this since I'm internal IBM and it could
be misconstrued as Gospel. But I had this problem once as well. Turned out
the problem was with name resolution. That is, the servers name would
resolve to the unreacheable ip address and Netview would try to use it.
You could ask him to try to play games with name resolution.
Peter Raig
IBM Global Services, Service Delivery Center - South
Consulting I/T Specialist - Senior
----- Forwarded by Peter Raig/Raleigh/IBM on 09/07/2001 09:09 AM -----
"Barr, Scott"
<Scott_Barr@csgsy To: "'IBM NetView
Discussion'"
stems.com> <nv-l@tkg.com>
Sent by: cc:
owner-nv-l@tkg.co Subject: RE: [NV-L] Demand
poll times
m out, but SNMP walk works.
09/07/2001 08:57
AM
Please respond to
IBM NetView
Discussion
We have a problem open with support for SNMP status polls that fail because
*some* of the SNMP requests are flowing to the wrong interface. For
instance, we have 3 routers in our network that get marked down and the
next
status poll marks them back up.
After exhaustive troubleshooting, we found that the packets were being
directed to an interface that was not routable from the NetView server. We
use a loop back interface in each router and the SNMP requests were flowing
to an address in a subnet that NetView could not reach. On the surface it
looked like NetView was not sending the packets at all but we finally found
the packets were going out just to the wrong address.
I would set up a sniffer/snoop and try and catch the SNMP activity. You may
be having this problem too.
-----Original Message-----
From: Brad Martin [mailto:bmartin@MetLife.com]
Sent: Thursday, September 06, 2001 4:24 PM
To: IBM NetView Discussion
Subject: Re: [NV-L] Demand poll times out, but SNMP walk works.
Leslie,
1. Because the node is setup for SNMP management, I get the Current Polling
Parameters and then it jumps to the first SNMP request (Get system
identifier) which times out.
2. The monitor -> mib values -> system information worked fine.
3. I've deleted all references to the router and any IP addresses from
the xnmsnmpconf. Yet the demand poll still times out on the snmpget system
identifier.
Brad Martin
MetLife
"Leslie Clark" <lclark@us.ibm.com>@tkg.com on 09/06/2001 04:03:39 PM
Please respond to "IBM NetView Discussion" <nv-l@tkg.com>
Sent by: owner-nv-l@tkg.com
To: "IBM NetView Discussion" <nv-l@tkg.com>
cc:
Subject: Re: [NV-L] Demand poll times out, but SNMP walk works.
Is it timing out at the very beginning of the demandpoll, or
is it showing the current status of interfaces, and then
failing on the first snmp request?
Does 'Monitor...MIB Values..System Info' time out, or work?
Regardless, I would do xnmsnmpconf and delete the
entry for that node, check for other entries for that node
and delete them as well, and then demandpoll it. The
demandpoll should update it with the right entry. You are
right, what you say you are seeing does not make sense.
And.... get to 6.0.2, of course:)
Cordially,
Leslie A. Clark
IBM Global Services - Systems Mgmt & Networking
Detroit
"Brad Martin"
<bmartin@METLI To: "IBM NetView Discussion"
FE.COM> <nv-l@tkg.com>
Sent by: cc:
owner-nv-l@tkg Subject: [NV-L] Demand poll
times out,
.com but SNMP walk works.
09/06/01 03:32
PM
Please respond
to IBM NetView
Discussion
We are using 6.01 and have one router that has an SNMP timeout on the
demand poll.
When I do an SNMP walk on the hostname and for all IP address everything is
OK.
When I do an xnmsnmpconf -resolve hostname / IP address --- > I get the
correct community string & IP addresses.
When I to a TEST --> IP/TCP/SNMP the results are OK
I've tried doing an xnmsnmpconf -delete hostname & IP address with
xnmsnmpconf -clearCache.
Any ideas as to why the demand poll would timeout?
Brad Martin
MetLife
_________________________________________________________________________
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
|