View previous topic :: View next topic |
Author |
Message |
mrq n00b
![n00b n00b](/images/ranks/rank_rect_0.gif)
![](images/avatars/74724491240263f72a6d25.jpg)
Joined: 08 Feb 2004 Posts: 15 Location: Warsaw/Poland
|
Posted: Thu Apr 03, 2008 9:20 am Post subject: Network counters reset |
|
|
Hi all,
I have a server with two D-Link DGE-530T Gigabit Ethernet Adapter (rev 11) (driver skge) and one onboard RTL8139. My setup is as follows:
eth0 -> DMZ (D-Link)
eth1 -> LAN (D-Link)
eth2 -> ISP (Realtek)
On the DMZ there is a web server serving high volumes of traffic.
When I look at ifconfig| grep TX I get something like this:
Code: |
TX packets:6926293 errors:0 dropped:0 overruns:0 carrier:0
RX bytes:27656893 (26.3 Mb) TX bytes:17465134 (16.6 Mb)
TX packets:3106715 errors:0 dropped:0 overruns:0 carrier:0
RX bytes:173682343 (165.6 Mb) TX bytes:4107556193 (3917.2 Mb)
TX packets:38144615 errors:0 dropped:0 overruns:6 carrier:0
RX bytes:2664702774 (2541.2 Mb) TX bytes:156594122 (149.3 Mb)
|
And one minute later:
Code: |
TX packets:6933552 errors:0 dropped:0 overruns:0 carrier:0
RX bytes:27453869 (26.1 Mb) TX bytes:17208846 (16.4 Mb)
TX packets:3106821 errors:0 dropped:0 overruns:0 carrier:0
RX bytes:173700155 (165.6 Mb) TX bytes:4107607384 (3917.3 Mb)
TX packets:38158849 errors:0 dropped:0 overruns:6 carrier:0
RX bytes:2665249460 (2541.7 Mb) TX bytes:177902090 (169.6 Mb)
|
The first problem: counters for eth0 reset approximately every second.
I can use iftop to see that the transfers are ok.
Ifstat output looks like this:
Code: |
eth0 eth1 eth2
KB/s in KB/s out KB/s in KB/s out KB/s in KB/s out
ifstat: warning: rollover for interface eth0, reinitialising.
0.00 0.00 0.12 0.12 5.67 230.71
ifstat: warning: rollover for interface eth0, reinitialising.
|
The second problem, which I just realized:
there may be a problem with eth2 counter: the TX bytes is too small for the amount of data that came through that interface during last 4 hours, but RX looks OK.
Any ideas? |
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
BradN Advocate
![Advocate Advocate](/images/ranks/rank-G-1-advocate.gif)
![](images/avatars/gallery/StarCraft/starcraft_protoss_observer.gif)
Joined: 19 Apr 2002 Posts: 2391 Location: Wisconsin (USA)
|
Posted: Thu Apr 03, 2008 12:10 pm Post subject: |
|
|
Any strange messages in dmesg? |
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
mrq n00b
![n00b n00b](/images/ranks/rank_rect_0.gif)
![](images/avatars/74724491240263f72a6d25.jpg)
Joined: 08 Feb 2004 Posts: 15 Location: Warsaw/Poland
|
Posted: Thu Apr 03, 2008 1:13 pm Post subject: |
|
|
No, nothing special. |
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
BradN Advocate
![Advocate Advocate](/images/ranks/rank-G-1-advocate.gif)
![](images/avatars/gallery/StarCraft/starcraft_protoss_observer.gif)
Joined: 19 Apr 2002 Posts: 2391 Location: Wisconsin (USA)
|
Posted: Thu Apr 03, 2008 3:06 pm Post subject: |
|
|
There must be some kind of bug in the kernel relating to the counter code. Perhaps the eth2 counter is rolling over normally, but I don't know what value it does that at, so... *shrug*
Certainly eth0 shouldn't be resetting every second though. I noticed there seems to be no RX/TX polling option in the kernel for this driver, so that can't be causing it. |
|
Back to top |
|
![](templates/gentoo/images/spacer.gif) |
|