in my setup there are two upstream providers for which i count the bytes using netflow. everything works fine, except that i still note some flow loss.
the flow collector is linux 2.4.30, flow-tools is 0.67, cisco flow export v5. the linux kernel parameters are: net.core.rmem_max = 10000000 net.core.wmem_max = 10000000 after a month of runnung 2 flow-capture processes i have: May 23 10:35:00 nhcc flow-capture[2158]: STAT: now=1116837300 startup=1114377863 src_ip=10.100.100.1 dst_ip=10.100.100.17 d_ver=5 pkts=3262957 flows=97888707 lost=1470 reset=0 filter_drops=0 May 23 10:35:00 nhcc flow-capture[2160]: STAT: now=1116837300 startup=1114377863 src_ip=10.100.100.2 dst_ip=10.100.100.17 d_ver=5 pkts=1289273 flows=38662809 lost=570 reset=0 filter_drops=0 i have pretty moderate traffic, very fast collector machine and good network hardware (all routers and switches are cisco). what could be the reasons for the flow loss? thanks, Kresimir _______________________________________________ Flow-tools mailing list [EMAIL PROTECTED] http://mailman.splintered.net/mailman/listinfo/flow-tools
