Take the 2-minute tour ×
Server Fault is a question and answer site for system and network administrators. It's 100% free, no registration required.

We have several servers running VMWare ESXi 4 on which are virtual machines running OpenSUSE and Windows 2008r2. We also have a server monitoring these vm's using OpenNMS. For several of the OpenSUSE vm's we get regular High Threshold messages for the percentage of discarded packets.

Inspecting the graphs showed that it are only incoming packages that are being discarded. Peaks are up to 10% discards. It doesn't happen on all Linux VM's. The ones it doesn't happen on are running a slightly older OpenSUSE 11.3 while the others are 11.4 or 12.1 however it is also true that the ones that suffer from this problem have very little load while the heaviest loaded VM has zero discarded packets for the last year.

The windows 2008 servers don't seem to have this issue at all.

So what is a likely cause for this? At what percentage should I start to worry?

Edit: I was wondering if snmp or opennms is showing incorrect data. When I look at the output of netstat -s I see: 0 incoming packets discarded

Here is the complete output (uptime is 137 days):

Ip:
    89907150 total packets received
    5047 with invalid addresses
    0 forwarded
    19748 with unknown protocol
    0 incoming packets discarded
    89882355 incoming packets delivered
    93223508 requests sent out
    133 dropped because of missing route Icmp:
    4121 ICMP messages received
    37 input ICMP message failed.
    ICMP input histogram:
        destination unreachable: 162
        echo requests: 3950
        timestamp request: 9
    24159 ICMP messages sent
    0 ICMP messages failed
    ICMP output histogram:
        destination unreachable: 20200
        echo replies: 3950
        timestamp replies: 9 IcmpMsg:
        InType3: 162
        InType8: 3950
        InType13: 9
        OutType0: 3950
        OutType3: 20200
        OutType14: 9 Tcp:
    309309 active connections openings
    175665 passive connection openings
    582 failed connection attempts
    19123 connection resets received
    27 connections established
    89397500 segments received
    98178773 segments send out
    39585 segments retransmited
    7 bad segments received.
    256696 resets sent Udp:
    448836 packets received
    524 packets to unknown port received.
    0 packet receive errors
    448854 packets sent
    RcvbufErrors: 0
    SndbufErrors: 0 UdpLite:
    InDatagrams: 0
    NoPorts: 0
    InErrors: 0
    OutDatagrams: 0
    RcvbufErrors: 0
    SndbufErrors: 0 TcpExt:
    1897 invalid SYN cookies received
    441 resets received for embryonic SYN_RECV sockets
    ArpFilter: 0
    125024 TCP sockets finished time wait in fast timer
    123 packets rejects in established connections because of timestamp
    2365891 delayed acks sent
    5221 delayed acks further delayed because of locked socket
    Quick ack mode was activated 1517 times
    76762 packets directly queued to recvmsg prequeue.
    32782 packets directly received from backlog
    148216 packets directly received from prequeue
    81827955 packets header predicted
    11453 packets header predicted and directly queued to user
    TCPPureAcks: 2002864
    TCPHPAcks: 81030906
    TCPRenoRecovery: 3
    TCPSackRecovery: 5098
    TCPSACKReneging: 46
    TCPFACKReorder: 31
    TCPSACKReorder: 17
    TCPRenoReorder: 0
    TCPTSReorder: 25
    TCPFullUndo: 31
    TCPPartialUndo: 599
    TCPDSACKUndo: 1712
    TCPLossUndo: 691
    TCPLoss: 5407
    TCPLostRetransmit: 201
    TCPRenoFailures: 0
    TCPSackFailures: 286
    TCPLossFailures: 101
    TCPFastRetrans: 11936
    TCPForwardRetrans: 493
    TCPSlowStartRetrans: 3237
    TCPTimeouts: 15424
    TCPRenoRecoveryFail: 0
    TCPSackRecoveryFail: 165
    TCPSchedulerFailed: 0
    TCPRcvCollapsed: 0
    TCPDSACKOldSent: 1569
    TCPDSACKOfoSent: 5
    TCPDSACKRecv: 7543
    TCPDSACKOfoRecv: 6
    TCPAbortOnSyn: 0
    TCPAbortOnData: 249150
    TCPAbortOnClose: 1391
    TCPAbortOnMemory: 0
    TCPAbortOnTimeout: 160
    TCPAbortOnLinger: 0
    TCPAbortFailed: 0
    TCPMemoryPressures: 0
    TCPSACKDiscard: 15
    TCPDSACKIgnoredOld: 2257
    TCPDSACKIgnoredNoUndo: 739
    TCPSpuriousRTOs: 129
    TCPMD5NotFound: 0
    TCPMD5Unexpected: 0
    TCPSackShifted: 38802
    TCPSackMerged: 24929
    TCPSackShiftFallback: 16027
    TCPBacklogDrop: 0
    TCPMinTTLDrop: 0
    TCPDeferAcceptDrop: 87963
    IPReversePathFilter: 0
    TCPTimeWaitOverflow: 0 IpExt:
    InNoRoutes: 0
    InTruncatedPkts: 0
    InMcastPkts: 2
    OutMcastPkts: 0
    InBcastPkts: 31370
    OutBcastPkts: 0
    InOctets: 17194554126
    OutOctets: 21418612641
    InMcastOctets: 72
    OutMcastOctets: 0
    InBcastOctets: 3492814
    OutBcastOctets: 0
share|improve this question
    
I'd start to worry at anything > 0%. –  Tom O'Connor Oct 3 '12 at 19:14
    
Bump your thresholds up. OpenNMS can be noisy. –  ewwhite Oct 10 '12 at 14:27

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Browse other questions tagged or ask your own question.