Re: [E1000-devel] Re-2: e1000e detected hardware unit hang problem

[Dave, T. N <tus...@in...>]

>-----Original Message-----
>From: Lars Maschke [mailto:ma...@la...]
>Sent: Monday, March 04, 2013 3:11 PM
>To: Dave, Tushar N
>Cc: e10...@li...
>Subject: Re-2: e1000e detected hardware unit hang problem
>
>Hello Tushar,
>
>first of all. Thanks for Your quick reply.
>
>That's the point. I don't know why this occurs. If I have the chance I see
>a failure of the e1000e driver on the console. The server is completly
>down and I can't logon to get any other information.
>
>The error isn't logged at dmesg.log or syslog on my debian system. There
>is no logging at all after the crash.

We at the least need the dmesg from the kernel to start with. How would you know that you're getting detected hardware unit hang message. Is it possible to take a picture or connect serial console to the server to retrieve the message after crash/hang? 
>
>Only a full reset solves the problem. I get this error every two, three or
>four days. At the crash time no special cron job is running. It occurs
>only an night between 0:00h and 2:30h.
>From now on I try to reset networking with the following bash-script every >night and I hope that it's a good idea:

>---
>#!/bin/sh

>/etc/init.d/networking stop
>/sbin/rmmod e1000e
>/sbin/modprobe e1000e RxIntDelay=0,0 IntMode=1,1 /etc/init.d/networking start >/sbin/ethtool -K eth0 tso off /sbin/shorewall restart
>---
How many e1000e devices you have in the system?

>Do You think that the problem can occur when the other Intel "e1000"
>driver is also loaded on the machine?
I don't think so however if you can, give it a try. And let me know if anything changes! 
Have you tried the in-kernel e1000e driver?

-Tushar