netstat -s returning "error parsing /proc/net/snmp" Success

[Bobby_Tables]

Hi,

I have two machine, host1 and host2, that are running Debian 3.1 in an HA setup. AFAIK, they are exactly the same. In the course of tracking down some monitoring bugs, I discovered that on host2 "netstat -s" gives

Bobby_Tables-host2:/etc$ netstat -s
Ip:
    3871931711 total packets received
    5758385 with invalid addresses
    0 forwarded
    0 incoming packets discarded
    3866173326 incoming packets delivered
    2053773355 requests sent out
Icmp:
    79642 ICMP messages received
    2 input ICMP message failed.
    ICMP input histogram:
        destination unreachable: 79623
        echo requests: 7
        echo replies: 12
    2479711 ICMP messages sent
    0 ICMP messages failed
    ICMP output histogram:
        destination unreachable: 2479692
        echo request: 12
        echo replies: 7
IcmpMsg:
        InType0: 12
        InType3: 79623
        InType8: 7
        OutType0: 7
        OutType3: 2479692
        OutType8: 12
Tcp:
    8042 active connections openings
    7950 passive connection openings
    92 failed connection attempts
    1 connection resets received
    3 connections established
    3862988781 segments received
    2050729430 segments send out
    113 segments retransmited
    0 bad segments received.
    4 resets sent
Udp:
    484441 packets received
    2619225 packets to unknown port received.
    0 packet receive errors
    564101 packets sent
UdpLite:
error parsing /proc/net/snmp: Success

"netstat -s" on host1 does not give that error line; following UdpLite is a line for TcpExt (which is absent in the output above).

So far google has turned up a lot of answers dealing with a (RedHat) kernel patch, so I think that's not the issue here.

My version of net-tools is 1.60-17 and netstat is 1.4.2

/proc/net/snmp looks as follows:

Ip: Forwarding DefaultTTL InReceives InHdrErrors InAddrErrors ForwDatagrams InUnknownProtos InDiscards InDelivers OutRequests OutDiscards OutNoRoutes ReasmTimeout ReasmReqds ReasmOKs ReasmFails FragOKs FragFails FragCreates
Ip: 2 64 3871933461 0 5758916 0 0 0 3866174545 2053774491 0 0 0 0 0 0 0 0 0
Icmp: InMsgs InErrors InDestUnreachs InTimeExcds InParmProbs InSrcQuenchs InRedirects InEchos InEchoReps InTimestamps InTimestampReps InAddrMasks InAddrMaskReps OutMsgs OutErrors OutDestUnreachs OutTimeExcds OutParmProbs OutSrcQuenchs OutRedirects OutEchos OutEchoReps OutTimestamps OutTimestampReps OutAddrMasks OutAddrMaskReps
Icmp: 79647 2 79628 0 0 0 0 7 12 0 0 0 0 2479942 0 2479923 0 0 0 0 12 7 0 0 0 0
IcmpMsg: InType0 InType3 InType8 OutType0 OutType3 OutType8
IcmpMsg: 12 79628 7 7 2479923 12
Tcp: RtoAlgorithm RtoMin RtoMax MaxConn ActiveOpens PassiveOpens AttemptFails EstabResets CurrEstab InSegs OutSegs RetransSegs InErrs OutRsts
Tcp: 1 200 120000 -1 8043 7951 92 1 3 3862989689 2050730268 113 0 4
Udp: InDatagrams NoPorts InErrors OutDatagrams RcvbufErrors SndbufErrors
Udp: 484504 2619468 0 564168 0 0
UdpLite: InDatagrams NoPorts InErrors OutDatagrams RcvbufErrors SndbufErrors
UdpLite: 0 0 0 0 0 0

On host1, /proc/net/snmp is similiar except for a TcpExt line after UdpLite.

Does anyone have any ideas on where/how to fix this and/or what's causing it?

Thanks in advance,

Bobby_Tables

[Bobby_Tables]

Well, turns out the two hosts were *not* in exactly the same configuration (and I didn't have access to one of two hosts to verify). Turns out the host giving the netstat -s error was running a slightly older kernel (it shouldn't have been) that has the long since patched bug that produces the /proc/net/snmp error.

Bobby_Tables