At the moment we have a linux server that receives
and send tcp/ip messages.
We noticed that the delay between reception and sending
responses increases if the amount of traffic increases.
We suspect that there is queueing in the tcp/ip buffer,
how can we check the actual size of the tcp/ip buffer.

SUSE linux version 10.

Hi,

welcome to LQ!

The upper limit of (memory) buffer size is defined in
/proc/sys/net/core/wmem_max

I'm not sure that the problem you see is depending
on the buffer size, though. What other steps of fault
finding have you taken, what does the output of ifconfig
have to say about the interface in question?


Cheers,
Tink

The ifconfig does not give any information about the queue size.
The problem that we notice is that the cpuUsage of our server
does not increase above the 60%, if we increase traffic. Instead
we see an increasing delay of our traffic, we suspect that the
tcp/ip queue is increasing. However it is difficult to see the
actual queue size.
(Note the dropped messages referred below are traffic independent 800
dropped per minute with or without traffic, so not relevant).


eth2 Link encap:Ethernet HWaddr 00:19:BB:2B:67:A8
inet addr:164.48.54.49 Bcast:164.48.55.255 Mask:255.255.252.0
inet6 addr: fe80::219:bbff:fe2b:67a8/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:245103453 errors:0 dropped:6707412 overruns:0 frame:0
TX packets:209418323 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:53489347836 (51011.4 Mb) TX bytes:81136844046 (77378.1 Mb)
Interrupt:185 Memory:f8000000-f8012100

lo Link encap:Local Loopback
inet addr:127.0.0.1 Mask:255.0.0.0
inet6 addr: ::1/128 Scope:Host
UP LOOPBACK RUNNING MTU:16436 Metric:1
RX packets:232550030 errors:0 dropped:0 overruns:0 frame:0
TX packets:232550030 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:55020758195 (52471.8 Mb) TX bytes:55020758195 (52471.8 Mb)