Re: packet drop with intel gigabit / marwell gigabit

OxY wrote:

what kind of details should i attach? to analyze the problem?

Assume that you have process A to cause process B dropping packets:

Step (1):
Run B only -- what is the maximum through without packet drop?
what is the CPU utilization?
what type of traffic (protocol -- TCP, UDP, etc.) does B use?
does B have other I/O? If so, how much is the I/O bandwidth used?

Step (2):
reRun B and involving A gradually (increasing A's wieght) till B drops packets.
provide all above statistics for A and B at this point (where drop starts).

One thing I missed in your original post what protocol drops packet -- TCP or UDP?

-Jin
_______________________________________________
freebsd-performance@xxxxxxxxxxx mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-performance
To unsubscribe, send any mail to "freebsd-performance-unsubscribe@xxxxxxxxxxx"

Relevant Pages

  • Re: port=1026&reason=ICMPsent
    ... > Actually ICMP is a layered protocol the UDP protocol in question is a ... in the payload of an ICMP ... with a payload indicating it was in response to a UDP packet? ...
    (alt.computer.security)
  • Re: port=1026&reason=ICMPsent
    ... Not sure now about the UDP, here is the .enc file decode for one of the ... Time delta from previous packet: ... Fragment offset: 0 ... Protocol: ICMP ...
    (alt.computer.security)
  • Re: Problem with the NDIS MUX IM driver (decapsulation not working)
    ... If the higher-level protocol and the lower-level miniport have enabled some TCP task offload contract, then the decapsulated packet you are indicating may not provide the necessary task offload information. ... then temporarily disabling the NDIS task offload features of the adapter using the adapter's NCPA advanced property tab should make the behavior "better". ... I slap on my own ethernet header infront of the real ...
    (microsoft.public.development.device.drivers)
  • Re: Sniffer Cant see cluster traffic
    ... > protocol and cluster specific multicast MAC address, ... > node specific source MAC address. ... > is a multi-cast packet or a directed packet. ... Yep and that is solely at the ethernet packet level. ...
    (comp.os.vms)
  • Re: Help Interpreting data from Wireshark
    ... What concerns me is that the packet seemed to have a source address of 192.168.1.1 but later in the packet you see the dest as 84.160.95.226 ... Protocol Info ... DENVER.local ICMP Destination unreachable (Port unreachable) ... Fragment offset: 0 ...
    (comp.os.linux.security)