Hi list
It's my first message here, but I use CentOS from long time ago (sorry my language, I'm spanish from Colombia)
I have a Intel board with an Integrated Network Card 82578DC. I've sucessfully installed the driver with the RPM kmod-e1000e; I've found it in ElRepo.org, and I've installed it via Yum.
After I've installed the package (see this: http://lists.elrepo.org/pipermail/elrepo/2010-December/000416.html ), I do:
ifdown eth1 modprobe -r e1000e ifup eth1
After these steps, my card works OK and get a valid IP address. However, when I reboot my machine, I must write those commands again. I've tried to put e1000e in /etc/modprobe.d/blachlist; however, it doesn't works.
Does anubody knows what must I do to have my network card working from start?
Thanks!
2011/2/15 Soporte Virtual soportev@gmail.com:
Hi list
It's my first message here, but I use CentOS from long time ago (sorry my language, I'm spanish from Colombia)
I have a Intel board with an Integrated Network Card 82578DC. I've sucessfully installed the driver with the RPM kmod-e1000e; I've found it in ElRepo.org, and I've installed it via Yum.
After I've installed the package (see this: http://lists.elrepo.org/pipermail/elrepo/2010-December/000416.html%C2%A0), I do:
ifdown eth1 modprobe -r e1000e ifup eth1
After these steps, my card works OK and get a valid IP address. However, when I reboot my machine, I must write those commands again. I've tried to put e1000e in /etc/modprobe.d/blachlist; however, it doesn't works.
Does anubody knows what must I do to have my network card working from start?
Thanks!
Hi again, list!
I've solved it! I've checked in /etc/sysconfig/networking/devices/ifcfg-eth1, and it had ONBOOT=no. With ONBOOT=yes issue was solved.
Thanks!
Am 16.02.2011 01:27, schrieb Soporte Virtual:
2011/2/15 Soporte Virtual soportev@gmail.com:
Hi list
It's my first message here, but I use CentOS from long time ago (sorry my language, I'm spanish from Colombia)
I have a Intel board with an Integrated Network Card 82578DC. I've sucessfully installed the driver with the RPM kmod-e1000e; I've found it in ElRepo.org, and I've installed it via Yum.
After I've installed the package (see this: http://lists.elrepo.org/pipermail/elrepo/2010-December/000416.html ), I do:
ifdown eth1 modprobe -r e1000e ifup eth1
After these steps, my card works OK and get a valid IP address. However, when I reboot my machine, I must write those commands again. I've tried to put e1000e in /etc/modprobe.d/blachlist; however, it doesn't works.
Does anubody knows what must I do to have my network card working from start?
Thanks!
Hi again, list!
I've solved it! I've checked in /etc/sysconfig/networking/devices/ifcfg-eth1, and it had ONBOOT=no. With ONBOOT=yes issue was solved.
Glad you were able to solve your problem your own.
Just a note about how the ELrepo package works: It sets the original kernel module (shipping with the CentOS/RHEL) kernel on a blacklist and forces to load the own module instead. Though both modules have the same name.
# cat /etc/depmod.d/kmod-e1000e.conf override e1000e * weak-updates/e1000e
Just curious, any specific reason why to choose the ELrepo module over the one coming with the CentOS kernel?
I have an RHEL 6 system with an "Intel Corporation 82574L Gigabit Network Connection" NIC, where the e1000e module of the RHEL kernel fails to drive the hardware (no traffic possible, strange effects when tcpduming it). With the kmod-e1000e from ELrepo the 2 onboard 82574L NICs now work.
Alexander
On Wed, Feb 16, 2011 at 8:57 AM, Alexander Dalloz ad+lists@uni-x.org wrote:
Just a note about how the ELrepo package works: It sets the original kernel module (shipping with the CentOS/RHEL) kernel on a blacklist and forces to load the own module instead. Though both modules have the same name.
# cat /etc/depmod.d/kmod-e1000e.conf override e1000e * weak-updates/e1000e
Just curious, any specific reason why to choose the ELrepo module over the one coming with the CentOS kernel?
I have an RHEL 6 system with an "Intel Corporation 82574L Gigabit Network Connection" NIC, where the e1000e module of the RHEL kernel fails to drive the hardware (no traffic possible, strange effects when tcpduming it). With the kmod-e1000e from ELrepo the 2 onboard 82574L NICs now work.
You may want to look through this ELRepo web site:
http://elrepo.org/tiki/Driver+Versions
As for the e1000e driver, the version offered by ELRepo is newer than what is in the EL6 kernel. Often times, ELRepo drivers are built from the manufacturer's source files, which would not happen with the upstream kernels because their source is from kernel.org).
But if you'd like more info or questions regarding ELRepo's drivers, you will need to ask on the ELRepo mailing list.
Akemi
Am 16.02.2011 18:09, schrieb Akemi Yagi:
On Wed, Feb 16, 2011 at 8:57 AM, Alexander Dalloz ad+lists@uni-x.org wrote:
Just curious, any specific reason why to choose the ELrepo module over the one coming with the CentOS kernel?
I have an RHEL 6 system with an "Intel Corporation 82574L Gigabit Network Connection" NIC, where the e1000e module of the RHEL kernel fails to drive the hardware (no traffic possible, strange effects when tcpduming it). With the kmod-e1000e from ELrepo the 2 onboard 82574L NICs now work.
You may want to look through this ELRepo web site:
http://elrepo.org/tiki/Driver+Versions
As for the e1000e driver, the version offered by ELRepo is newer than what is in the EL6 kernel. Often times, ELRepo drivers are built from the manufacturer's source files, which would not happen with the upstream kernels because their source is from kernel.org).
But if you'd like more info or questions regarding ELRepo's drivers, you will need to ask on the ELRepo mailing list.
Akemi
Thanks Akemi for your response.
I know you are part of the ELrepo team and saw you contributing to a bugzilla ticket, asking for updating the e1000e kernel module with recent Intel sources. I understand the point the Red Hat team is making, when they say, they will only update when the Intel version found their way into Linus's vanilla kernel tree.
Though I am planning to open an upstream ticket because the 2 Intel 82574L NICs are not usable with plain RHEL 6, but they are using the ELrepo module. Unfortunately I was not able to trace and dump much useful information to fill in the ticket besides to note that simply no network traffic is going through the NICs with the RHEL 6 kernel module.
Sidenote: The e1000e module coming with CentOS 5.5 does drive the Intel 82574L chips correctly on another box. The boards are btw. a Supermicro X8SIL.
Glad there is the ELrepo! :)
Regards
Alexander
On Wed, Feb 16, 2011 at 10:21 AM, Alexander Dalloz ad+lists@uni-x.org wrote:
Am 16.02.2011 18:09, schrieb Akemi Yagi:
But if you'd like more info or questions regarding ELRepo's drivers, you will need to ask on the ELRepo mailing list.
Akemi
I know you are part of the ELrepo team and saw you contributing to a bugzilla ticket, asking for updating the e1000e kernel module with recent Intel sources. I understand the point the Red Hat team is making, when they say, they will only update when the Intel version found their way into Linus's vanilla kernel tree.
For people wondering which bugzilla report you are referencing:
https://bugzilla.redhat.com/show_bug.cgi?id=632650
Though I am planning to open an upstream ticket because the 2 Intel 82574L NICs are not usable with plain RHEL 6, but they are using the ELrepo module. Unfortunately I was not able to trace and dump much useful information to fill in the ticket besides to note that simply no network traffic is going through the NICs with the RHEL 6 kernel module.
By all means, try filing a support request. Unless the RHEL-6 kernel gets the updated driver, CentOS-6 would not. I suppose (hope) you can send in a request without having to provide all technical details. Just mention the "fact" that the driver in the RHEL-6 kernel does not work while a newer version from ELRepo works and see how they handle the case.
Glad there is the ELrepo! :)
Always good to hear a success story like this. [But once again, it's best mentioned on the ELRepo ML]
Akemi