LAN Interface Traffic Dropped

[Bhushan]

Hi All,

Can anyone help me what actions need to be taken to prevent packets dropping?


eth7      Link encap:Ethernet  HWaddr 00:90:FB:1D:2E:BD
          inet addr:192.168.1.250  Bcast:192.168.1.255  Mask:255.255.255.0
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:761993027 errors:0 dropped:73277815 overruns:0 frame:0
          TX packets:905908928 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:350059445609 (326.0 GiB)  TX bytes:967108848739 (900.6 GiB)
          Interrupt:17 Memory:fd9e0000-fda00000

System health is quite good, throughput is above 300 Mbps. Below commands show some of the information given by ethtool command.

[root@manage]# ethtool -i eth7
driver: e1000e
version: 1.3.10-k2
firmware-version: 5.6-2
bus-info: 0000:07:00.1
[root@manage]# ethtool -a eth7
Pause parameters for eth7:
Autonegotiate:  on
RX:             on
TX:             on

[root@manage]# ethtool -c eth7
Coalesce parameters for eth7:
Adaptive RX: off  TX: off
stats-block-usecs: 0
sample-interval: 0
pkt-rate-low: 0
pkt-rate-high: 0

rx-usecs: 3
rx-frames: 0
rx-usecs-irq: 0
rx-frames-irq: 0

tx-usecs: 0
tx-frames: 0
tx-usecs-irq: 0
tx-frames-irq: 0

rx-usecs-low: 0
rx-frame-low: 0
tx-usecs-low: 0
tx-frame-low: 0

rx-usecs-high: 0
rx-frame-high: 0
tx-usecs-high: 0
tx-frame-high: 0

[root@manage]# ethtool -g eth7
Ring parameters for eth7:
Pre-set maximums:
RX:             4096
RX Mini:        0
RX Jumbo:       0
TX:             4096
Current hardware settings:
RX:             256
RX Mini:        0
RX Jumbo:       0
TX:             256

[root@manage]# ethtool -k eth7
Offload parameters for eth7:
rx-checksumming: on
tx-checksumming: on
scatter-gather: on
tcp-segmentation-offload: on
udp-fragmentation-offload: off
generic-segmentation-offload: on
generic-receive-offload: on
large-receive-offload: off
[root@manage]# ethtool -S eth7
NIC statistics:
     rx_packets: 772538472
     tx_packets: 918761168
     rx_bytes: 354685899158
     tx_bytes: 981404765940
     rx_broadcast: 1334276
     tx_broadcast: 1339509
     rx_multicast: 2755
     tx_multicast: 0
     rx_errors: 0
     tx_errors: 0
     tx_dropped: 0
     multicast: 2755
     collisions: 0
     rx_length_errors: 0
     rx_over_errors: 0
     rx_crc_errors: 0
     rx_frame_errors: 0
     rx_no_buffer_count: 17967
     rx_missed_errors: 37739
     tx_aborted_errors: 0
     tx_carrier_errors: 0
     tx_fifo_errors: 0
     tx_heartbeat_errors: 0
     tx_window_errors: 0
     tx_abort_late_coll: 0
     tx_deferred_ok: 0
     tx_single_coll_ok: 0
     tx_multi_coll_ok: 0
     tx_timeout_count: 0
     tx_restart_queue: 15346748
     rx_long_length_errors: 0
     rx_short_length_errors: 0
     rx_align_errors: 0
     tx_tcp_seg_good: 10955
     tx_tcp_seg_failed: 0
     rx_flow_control_xon: 39740307
     rx_flow_control_xoff: 39740311
     tx_flow_control_xon: 862
     tx_flow_control_xoff: 1671
     rx_long_byte_count: 354685899158
     rx_csum_offload_good: 742388486
     rx_csum_offload_errors: 1302
     rx_header_split: 0
     alloc_rx_buff_failed: 0
     tx_smbus: 0
     rx_smbus: 0
     dropped_smbus: 0
     rx_dma_failed: 0
     tx_dma_failed: 0

Regards,
Bhushan Karia

[Bhushan]

Other strange thing is that whenever i start tcpdump capturing on same interface dropped counters are not increased.

Regards,
Bhushan Karia

[nehal dattani]

Hi


On Thu, Jan 10, 2013 at 6:16 PM, Bhushan <bhusha...@gmail.com> wrote:
> rx_no_buffer_count: 17967
> rx_missed_errors: 37739

This suggests that your current ring buffers are not capable of holding data.

Run following command to fix this.


ethtool -G eth7 rx 4096


Please note that

1) This will not reset current dropped counters. However, drop
counters should not increase.

2) This will change at every reboot/ NIC up-down. I am sure that you
know how to use /etc/rc.local


3) ==> Other strange thing is that whenever i start tcpdump capturing

on same interface dropped counters are not increased.

There is no way tcpdump can detect this. So its not strange, its normal.

--
The production of too many useful things results in too many useless people.

With best regards,
Nehal Dattani

[Bhushan]

I already tried to increase rx and tx parameters with -G flag yesterday but still problem persists.

Regards,
Bhushan Karia

[nehal dattani]

Hi

> I already tried to increase rx and tx parameters with -G flag yesterday but
> still problem persists.

How are you verifying it ?

Also give output of netstat -s collected at intervals of 12 or 24 hours.