netstat -i not accurate on -stable ?

Julian Elischer julian at freebsd.org
Mon Jun 24 04:05:55 UTC 2013


On 6/22/13 4:28 PM, Olivier Cochard-Labbé wrote:
> Hi all,
>
> I was testing my pthread-netblbast (PR bin/179085) when I found a
> problem with netstat counters (tested on 9-stable r252056).
>
> netreceive (src/toots/tools/netrate) display this output:
>      542025 pps   43.362 Mbps - 271554 pkts in 0.500998393 ns
>      541562 pps   43.325 Mbps - 271323 pkts in 0.501000754 ns
>      541517 pps   43.321 Mbps - 271300 pkts in 0.500999700 ns
>      540487 pps   43.239 Mbps - 270784 pkts in 0.500999904 ns
>      541541 pps   43.323 Mbps - 271312 pkts in 0.500999621 ns
>      540624 pps   43.250 Mbps - 270853 pkts in 0.501000361 ns
>      541403 pps   43.312 Mbps - 271243 pkts in 0.500999669 ns
>      541079 pps   43.286 Mbps - 271081 pkts in 0.501000212 ns
>      540887 pps   43.271 Mbps - 270985 pkts in 0.501000258 ns
>      541796 pps   43.344 Mbps - 271440 pkts in 0.500999971 ns
>      541056 pps   43.284 Mbps - 271069 pkts in 0.500999719 ns
>      541224 pps   43.298 Mbps - 271155 pkts in 0.501002856 ns
>      541120 pps   43.290 Mbps - 271100 pkts in 0.500997659 ns
>
> But in the same time a "netstat -ihw 1" display this output:
>              input        (Total)           output
>     packets  errs idrops      bytes    packets  errs      bytes colls
>        521k  508k     0        30M          3     0        562     0
>        521k  511k     0        30M          3     0        514     0
>        521k  526k     0        30M          3     0        514     0
>        521k  538k     0        30M          4     0        616     0
>        521k  495k     0        30M          3     0        514     0
>        521k  502k     0        30M          3     0        514     0
>        521k  531k     0        30M          3     0        514     0
>        521k  502k     0        30M          3     0        514     0
>        521k  504k     0        30M          3     0        514     0
>        521k  495k     0        30M          3     0        514     0
>        521k  498k     0        30M          3     0        514     0
>        521k  516k     0        30M          3     0        514     0
>
> There are about 20Kpps difference between them.

check if you get 60 of each in a 1 minute period.
it's possible that due to overhead one of them is going to only get
55 iterations per second producing a difference...

>
> Regards,
>
> Olivier
> _______________________________________________
> freebsd-net at freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-net
> To unsubscribe, send any mail to "freebsd-net-unsubscribe at freebsd.org"
>



More information about the freebsd-net mailing list