Hi again Alfredo.
Just to clarify, when I said:-
If I leave pfcount running on zc:eth3...
I meant if I leave it running _and_ resend my test transmission.
Regards
A.
PS I'm not sure why my messages are showing as separate threads on the mailing
list. Smart phone email...
--
Andrew Howard
[email protected]
----- Reply message -----
From: "Andrew Howard" <[email protected]>
To: <[email protected]>
Subject: pfcount stats on ZC interface not as expected
Date: Mon, Aug 17, 2015 13:28
.
Hi Alfredo,
Apologies in advance for any formatting issues - I'm writing this on my HTC
smart(ish) phone.
The -a active wait option seems to have no effect on pfcount stats, whether on
eth3 or zc:eth3. What would you be expecting?
I can confirm that ethtool -g reports 32768 for Rx and Tx, for both pre-set
maximums and current hw settings.
In terms of sustained data rates:-
When using interface eth3, pfcount reports 95-100% packet reception when the
transmitting Napatech card reports approx 5800 mbps. Higher tx rates lead to
significant packet drop.
The data rate as reported by pfcount was in the same ballpark - about 5.5gbps.
When using interface zc eth3, the stats reported by pfcount are always the
same, regardless of the transmission rate. A packet recv count much lower than
it should be, and packets dropped of zero.
Under the 'actual stats' area the gbps is reported as 0.0, even though Napatech
tx reports 5800 mbps or so.
If I leave pfcount running on zc:eth3, packets received remains unchanged,
whilst packets dropped increases.
Again, I'm continuing to investigate things here, but any suggestions would be
welcome.
Regards,
A.
--
Andrew Howard
[email protected]
_______________________________________________
Ntop-misc mailing list
[email protected]
http://listgateway.unipi.it/mailman/listinfo/ntop-misc
