Take the 2-minute tour ×
Server Fault is a question and answer site for professional system and network administrators. It's 100% free, no registration required.

I have server with Broadcom Corporation NetXtreme II BCM5706 Gigabit Ethernet chips, and 2.6.18-164.2.1.el5 kernel from Red Hat Enterprise Linux Server release 5.4 (Tikanga).

During normal work day this machine gets ~ 40000 packets per second (it's database server), and ifconfig shows that it drops about 1 packet per second. Which while not perfect, is acceptable.

but sometimes, something strange happens, and we are getting (?) less packets. For example:

  • 13:38:50 43260
  • 13:38:51 42930
  • 13:38:52 38696
  • 13:38:53 33006
  • 13:38:54 23013
  • 13:38:55 49485
  • 13:38:56 37514
  • 13:38:57 4858
  • 13:38:58 1089
  • 13:38:59 31054
  • 13:39:00 36540
  • 13:39:01 47228
  • 13:39:02 35634
  • 13:39:03 35348
  • 13:39:04 32908
  • 13:39:05 33226
  • 13:39:06 32639
  • 13:39:07 21842
  • 13:39:08 38560

This is number of packets per second. As you can see at 13:38:57 and 13:38:58 we had way less packets (this data is from tcpdump).

dropped: stats in ifconfig eth1 output don't change, switch (some cisco stuff) doesn't show any dropped packets.

Anyone knows what it could be?

share|improve this question
    
Well, less traffic during that time comes to mind. –  EEAA Sep 28 '11 at 21:30
    
Nope. While there are fluctuations in traffic, they are not even close to the changes in packet count. –  user13185 Sep 28 '11 at 21:32
add comment

1 Answer 1

Bigger packets during that time, resulting in less overall packets. E.g. some sort of bulk data transfer.

share|improve this answer
    
While theoretically possible, its just strange. What would cause suddenly, for very short time, such dramatic change? it's db plugged to website, and the traffic on it is pretty much stable. –  user13185 Sep 29 '11 at 13:39
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.