nv-l
[Top] [All Lists]

[nv-l] SNMP polling operation

To: nv-l@lists.us.ibm.com
Subject: [nv-l] SNMP polling operation
From: "Evans, Bill" <Bill.Evans@hq.doe.gov>
Date: Wed, 23 Mar 2005 19:30:40 -0500
Delivery-date: Thu, 24 Mar 2005 00:41:36 +0000
Envelope-to: nv-l-archive@lists.skills-1st.co.uk
Reply-to: nv-l@lists.us.ibm.com
Sender: owner-nv-l@lists.us.ibm.com

I've tracked down a problem with SNMP status polling to a point where I find outbound data packets are diagnosed by TCPDUMP to have bad UDP checksums.  TCPDUMP gives me "ChicagOps-doenet.snmp: bad udp cksum 9aa1!] "

The question is where the checksum is originated.  Assuming the following path for the packet:

      NIC Card ß tcpdump ß IP Layer ß UDP Layer ß SNMP manager ß netmon

I would expect the checksum to originate before it hits the UDP layer.  Can anyone identify if it's Netmon or the net-SNMP code which creates that checksum. 

The result is that we get a false failure for the total ChicagOps-doenet node and cannot recover the node unless we do a demand poll.  Quicktest does not work.  Pinging the interfaces will work but the node will fail on the next poll. 

The system is NetView 7.1.4 FP02 executing on SUSE Linux 9.   It does not happen on a Solaris 2.8 with NetView 7.1.3. 

My gratitude to anyone who can answer this one. 

Bill Evans

Tivoli NetView Support for DOE

301-903-0057

<Prev in Thread] Current Thread [Next in Thread>
  • [nv-l] SNMP polling operation, Evans, Bill <=

Archive operated by Skills 1st Ltd

See also: The NetView Web