2

Running netstat with -d adds dropped packets count But how does one tell where they were dropped?

  • I know they were not dropped at ingress to protocol handling (the traffic is UDP, and netstat -s does not show udp drops)
  • But I can't tell if the drop is due to some egress or ingress issue
  • On Linux one can see RX drops vs TX drops in ifconfig
  • Any other tools recommended?

Netstat Output:

~ lab$  netstat  -ndi
Name  Mtu   Network       Address            Ipkts Ierrs    Opkts Oerrs  Coll Drop
lo0   16384 <Link#1>                       2490046     0  2490046     0     0   0
lo0   16384 ::1/128     ::1                2490046     -  2490046     -     -   -
lo0   16384 127           127.0.0.1        2490046     -  2490046     -     -   -
lo0   16384 fe80::1%lo0 fe80:1::1          2490046     -  2490046     -     -   -
gif0* 1280  <Link#2>                             0     0        0     0     0   0
stf0* 1280  <Link#3>                             0     0        0     0     0   0
en0   1500  <Link#4>    38:c9:86:2d:de:bf 91802142    28 986818044     0     0 44126800
en0   1500  fe80::3ac9: fe80:4::3ac9:86ff 91802142     - 986818044     -     -   -
en0   1500  10.223.1/29   10.223.1.1      91802142     - 986818044     -     -   -
en7   1500  <Link#5>    38:c9:86:26:76:37  4239414     0 479870651     0     0 32604689
en7   1500  fe80::3ac9: fe80:5::3ac9:86ff  4239414     - 479870651     -     -   -
en7   1500  10.223/16     10.223.1.17      4239414     - 479870651     -     -   -
en6   1500  <Link#6>    38:c9:86:29:3e:25 97255294     1 1014031322     0     0 48927772
en6   1500  fe80::3ac9: fe80:6::3ac9:86ff 97255294     - 1014031322     -     -   -
en6   1500  10.223.1.8/29 10.223.1.9      97255294     - 1014031322     -     -   -
en6   1500  10.223.55/29  10.223.55.1     97255294     - 1014031322     -     -   -
en5   1500  <Link#7>    04:69:f8:eb:e8:e3  1754329     0  1714539     0     0   0
en5   1500  fe80::669:f fe80:7::669:f8ff:  1754329     -  1714539     -     -   -
en5   1500  10.223/24     10.223.0.2       1754329     -  1714539     -     -   -
en1   1500  <Link#8>    bc:54:36:cd:30:04        0     0        0     0     0   8
en2   1500  <Link#9>    2a:00:01:2c:33:30        0     0        0     0     0   0
en3   1500  <Link#10>   2a:00:01:2c:33:31        0     0        0     0     0   0
p2p0  2304  <Link#11>   0e:54:36:cd:30:04        0     0        0     0     0   0
awdl0 1484  <Link#12>   f2:0d:c0:8c:cf:32        0     0      248     0     0   0
awdl0 1484  fe80::f00d: fe80:c::f00d:c0ff        0     -      248     -     -   -
bridg 1500  <Link#13>   3a:c9:86:d2:52:00        0     0        1     0     0   0

BTW - due to external observation in my test scenario I can determine that these are egress failures - still would be good if I could tell from netstat (or other local command) what direction the drops are

  • What is the actual problem you are trying to solve? – andrewmh20 Jun 22 '16 at 19:03
  • You mean a problem more than not having feature parity with Linux networking stack? I'm trying to exert UDP load on a network component (DUT) driving traffic from macminis... I am familiar with different settings to increase buffer pools etc - but hate to go muck around in networking parameters without concrete data – nhed Jun 22 '16 at 19:21

You must log in to answer this question.

Browse other questions tagged .