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.

ifconfig tells for eth0 some RX = 2,8GB, TX = 1,3GB value that cannot be real, since I recently transmitted many 10GB+ files over eth0. I would like to know

  • if that's just some ordinary integer overflow (4GB limit)
  • or if that's an indicator that there is some evil rootkit that lies wrong data

It's a silly question, but the discrepance keeps me bothering.

Thank you, Nils

share|improve this question

2 Answers 2

up vote 5 down vote accepted

I would say it is the 4GB wraparound as you are guessing. I ran into this with fairly recent 32 bit linux kernels.

You can grab the source code for your kernel and see if it is the same in include/linux/netdevice.h and check the data type of net_device_stats->rx_bytes. If you are using a 32 bit system and the time is an unsigned long you will get only 2^32 bytes or 4 GB. More about this in a post of mine here.

Unless of course ifconfig grabs is counters from somewhere besides proc these days...

share|improve this answer
    
ifconfig grabs the counter from /proc/net/dev, which is (of course) the same value. Well - thank you for the answer, but I'm really wondering why they don't use an unsigned long long to get a better resolution - 4GB aren't much stuff for network devices in these days and you can't do any usable statistics any more with these cropped values :-/ –  nils Jul 23 '10 at 14:23

You can set up iptables to manage the counters - they can even be made to survive reboots with a save/restore or manually clearing/setting the counters to specific values.

If you don't already have iptables rules you just need to add at least one rule to the input and output chains that allows everything for example and it will provide what you want:

iptables -A INPUT -j ACCEPT
iptables -A OUTPUT -j ACCEPT

Then you can see the totals:

root@devcloner:~# iptables -n -vL
Chain INPUT (policy ACCEPT 2850K packets, 4183M bytes)
 pkts bytes target     prot opt in     out     source               destination         
 22M   32G ACCEPT     all  --  *      *       0.0.0.0/0            0.0.0.0/0            

Chain FORWARD (policy ACCEPT 0 packets, 0 bytes)
 pkts bytes target     prot opt in     out     source               destination         

Chain OUTPUT (policy ACCEPT 657K packets, 43M bytes)
 pkts bytes target     prot opt in     out     source               destination         
12951  813K ACCEPT     all  --  *      *       0.0.0.0/0            0.0.0.0/0           

-x will show the full byte counter:

root@devcloner:~# iptables -n -vL -x
Chain INPUT (policy ACCEPT 2850263 packets, 4182667884 bytes)
        pkts      bytes target     prot opt in     out     source                   destination         
22285352 32724735127 ACCEPT     all  --  *      *       0.0.0.0/0            0.0.0.0/0        

Chain FORWARD (policy ACCEPT 0 packets, 0 bytes)
    pkts      bytes target     prot opt in     out     source               destination         

Chain OUTPUT (policy ACCEPT 657099 packets, 43320848 bytes)
    pkts      bytes target     prot opt in     out     source               destination         
  102453  6738544 ACCEPT     all  --  *      *       0.0.0.0/0            0.0.0.0/0  

That info is probably parseable somewhere from /proc or /sys too.

share|improve this answer

Your Answer

 
discard

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