David can you resync and check if the patch we committed addresses your crash?
Thanks Luca On Aug 9, 2011, at 9:03 AM, Luca Deri wrote: > David > the only thing you have to do is use the device name "dan:ethX" if you want, > otherwise the system detects it for you > > Performance. You are sending and receiving from the same PC using two > different ports, correct? Below you have listed the CPU, but at these speeds > the chipset you have and the way the PCIe slot is connected to your CPU is > more important than the CPU type. Can you let us know about this? I have a > Sandy Bridge PC that in theory should be much faster than the one used in the > blog port tests, and it's not like that as the CPY (E3 family) is faster but > the new chipset is much worse. > > THis said I suggest to add "-a" to pfcount. How did you insmod the ixgbe > driver? > > The 4.7.3 PF_RING is something we have just started to code yesterday, so it > might be there's something to fix still. Can you please send us a trace to > track this issue or explain how to reproduce it? > > Regards Luca > > On Aug 9, 2011, at 6:44 AM, David Kwan wrote: > >> Dear All, >> >> I followed the instructions in UsersGuide to setup a PF_RING environment for >> our project. >> Followed by some testes to verify my setup environment correctness : >> >> sender send packet using pfsend at Packet size = 64byte , rate = 10Gbps >> receiver side experienced a 2% packet drop according to pfcount. >> Base on the info of ethtool, the NIC should have received all the packets >> sent from the sender. >> The CPU utilization of a single core is around 99%-100% >> The result is quit different from the blog post - >> http://www.ntop.org/blog/pf_ring/how-to-sendreceive-26mpps-using-pf_ring-on-commodity-hardware/ >> >> Are there anything that I can do to locate where's the problem ? >> >> btw, example program pfsend in PF_RING-4.7.3 Rev 4769 have some problem and >> got killed by the system. I fall back to use PF_RING 4.7.2 Rev exported in >> sender machine. >> >> Regards, >> David. >> >> Testing environment: >> sender & receiver machine : >> Xeon E7-4870 with Silicom 82599 NIC, fresh install Ubuntu 11.04 x64, kernel >> 2.6.38-10-server. >> PF_RING 4.7.3 ($Revision: 4769$) >> >> Results from pfsend and pfcount >> >> # ./pfcount -i dna:eth6 >> ... >> ========================= >> Absolute Stats: [979128687 pkts rcvd][19096141 pkts dropped] >> Total Pkts=998224828/Dropped=1.9 % >> 979'128'686 pkts - 58'747'721'160 bytes [10'527'768.71 pkt/sec - 5'053.32 >> Mbit/sec] >> ========================= >> Actual Stats: 12610497 pkts [1'000.02 ms][12'610'169.13 pkt/sec] >> ========================= >> >> ========================= >> Absolute Stats: [980883387 pkts rcvd][19116613 pkts dropped] >> Total Pkts=1000000000/Dropped=1.9 % >> 980'883'387 pkts - 58'853'003'220 bytes [10'434'428.24 pkt/sec - 5'008.52 >> Mbit/sec] >> ========================= >> Actual Stats: 1754701 pkts [1'000.12 ms][1'754'474.67 pkt/sec] >> >> ========================= >> Absolute Stats: [980883387 pkts rcvd][19116613 pkts dropped] >> Total Pkts=1000000000/Dropped=1.9 % >> 980'883'387 pkts - 58'853'003'220 bytes [10'324'579.88 pkt/sec - 4'955.79 >> Mbit/sec] >> ========================= >> Actual Stats: 0 pkts [1'000.16 ms][0.00 pkt/sec] >> ========================= >> --------------------------------------------------------------------------------------------------------------------------- >> >> # ./pfsend -i dna:eth8 -l 60 -g 5 -n 1000000000 -r 10 >> ... >> TX rate: [current 12'833'331.66 pps/8.62 Gbps][average 12'826'565.33 >> pps/8.62 Gbps][total 987'655'625.00 pkts] >> TX rate: [current 12'834'005.12 pps/8.62 Gbps][average 12'826'657.11 >> pps/8.62 Gbps][total 1'000'000'000.00 pkts] >> TX rate: [current 0.00 pps/0.00 Gbps][average 12'819'067.22 pps/8.61 >> Gbps][total 1'000'000'000.00 pkts] >> >> _______________________________________________ >> Ntop-misc mailing list >> [email protected] >> http://listgateway.unipi.it/mailman/listinfo/ntop-misc > > --- > We can't solve problems by using the same kind of thinking we used when we > created them - Albert Einstein > --- "Debugging is twice as hard as writing the code in the first place. Therefore, if you write the code as cleverly as possible, you are, by definition, not smart enough to debug it. - Brian W. Kernighan _______________________________________________ Ntop-dev mailing list [email protected] http://listgateway.unipi.it/mailman/listinfo/ntop-dev
