Running multiple instances of nping results in packet drops

[Dan Brown <dan.brown.vzw () gmail com>]

Hello,

I'm trying to UDP nping from a single VM to 5 other VMs by running 5
separate nping processes.

When I'm UDP pinging to only one VM I see no lost packets (one RCVD for
every SND). The more VMs I UDP ping the larger the lost packet ratio is for
all VMs.

I thought it's might be related to OS buffering so I increased the
following network stack parameters to 25MB, to no avail:

- net.core.rmem_max
- net.core.rmem_default
- net.core.wmem_max
- net.core.wmem_default
- net.core.netdev_max_backlog
- net.ipv4.udp_mem
- net.ipv4.udp_rmem_min
- net.ipv4.udp_wmem_min

Nping cmd I use (x5 times):
# nping --udp <remote VM IP> -g <local port> -p <remote port> --data-length
64 --count 60  --rate 1

UDP fragmentation offloading is enable for the underlying NIC (it's SR-IOV).

Is there anything else I could do to overcome those packet drops?

Thanks,
Dan

_______________________________________________
Sent through the dev mailing list
https://nmap.org/mailman/listinfo/dev
Archived at http://seclists.org/nmap-dev/