Hi,
I've been experiencing occassional network time-outs with the Intel Gigabit nic's (e1000) in Poweredge systems.
I'm not sure if it's a hard or software problem but it occurs on different systems with Centos 4.2 and 4.3 and was wondering if there' s a workaround available like compiling an updated e1000 module or something.
kind regards,
Geert
I've been experiencing occassional network time-outs with the Intel Gigabit nic's (e1000) in Poweredge systems. I'm not sure if it's a hard or software problem but it occurs on different systems with Centos 4.2 and 4.3 and was wondering if there' s a workaround available like compiling an updated e1000 module or something.
I'd start by checking to make sure that the system is fully updated (bios, OS etc) and then check out linux.dell.com for driver packages, updates and things of that nature.
On Sun, 25 Jun 2006 at 2:49pm, Geert Batsleer wrote
I've been experiencing occassional network time-outs with the Intel Gigabit nic's (e1000) in Poweredge systems.
I'm not sure if it's a hard or software problem but it occurs on different systems with Centos 4.2 and 4.3 and was wondering if there' s a workaround available like compiling an updated e1000 module or something.
There are a *lot* of tunables in the e1000 module. Look in /usr/share/doc/kernel-doc-2.6.9/Documentation/networking/e1000.txt for a list and descriptions. I'd start there.
Long ago, I managed a bunch of Dell PowerEdge 1U systems running Fedora Core 1. The built-in driver worked for a while, but died after a few hours of heavy burn-in. It was a consistent failure that was fixed by downloading and compiling the driver from Intel (as linked to from within the Dell support boards)
Once that was done, it worked a champ - but then I had to recompile the driver every time a kernel update came out. Big pain, glad I don't work there anymore.
-Ben
On Sunday 25 June 2006 05:49, Geert Batsleer wrote:
Hi,
I've been experiencing occassional network time-outs with the Intel Gigabit nic's (e1000) in Poweredge systems.
I'm not sure if it's a hard or software problem but it occurs on different systems with Centos 4.2 and 4.3 and was wondering if there' s a workaround available like compiling an updated e1000 module or something.
kind regards,
Geert
-- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean.
At Mon, 26 Jun 2006 it looks like Benjamin Smith composed:
Long ago, I managed a bunch of Dell PowerEdge 1U systems running Fedora Core
- The built-in driver worked for a while, but died after a few hours of
heavy burn-in. It was a consistent failure that was fixed by downloading and compiling the driver from Intel (as linked to from within the Dell support boards)
Once that was done, it worked a champ - but then I had to recompile the driver every time a kernel update came out. Big pain, glad I don't work there anymore.
I just had to build a few cheap storage servers, nine drives, eight drives hanging off two promise controllers (mdadm) and went and bought new Intel gig cards, the OS saw the cards, used the e1000 driver and the card would work for a few seconds and die. Bought two card actually, one for each server, tried the other card, got a bad checksum error on the card just from to configure it the first time when the OS did not even see it!
Well, I took the Intel gig card out and it was just the most flimsiest fragile thing when I got to looking at it -- nothing like the old cards of the past. Close to $55.00 each too.
Went and bought Netgear's gig card, the GA-311 and it worked and configured right out of the box, kicks butt, was literally twice as heavy as the Intel card and gave me that "just feels solid" in the hand feeling, like NIC cards of the past.
I have not cranked it to see if it can handle full throttle but it beat the experience I was having with the e1000 card.
YMMV