When CentOS8 came out I was all excited. It shipped with a newer kernel, version of PHP that was not EOL and a lot of other goodies. My first few installs on “regular” servers went on without a hitch. I then tried to install it on some Generation 8 HP Blade BL460c’s and it worked fine, everything that is except networking. With the help of some people in the CentOS group on Facebook.
Since the blades are sitting in a data center the only way for me to set them up remotely is via the ilo. I put the CentOS8 DVD up a NGNGX host and was easily able to install the operating system. As mentioned above all was fine except there was no networking. The suggestion was made that I run lspci -nnk which gave me this
As you can see the ID of the device is 103c:337b. If we look at the removed adapters list form RedHat we can see that support for our driver was removed. There is no reason to fret elrepo has a driver for it. If we look at the supported device ID’s for elepo we can clearly see 19A2:0710 under be2net.ko. Next I went hunting for elrep8 specifically for be2net and found http://mirror.rackspace.com/elrepo/elrepo/el8/x86_64/RPMS/. But how do we get this rpm onto a server sitting 60 miles away you ask? Very easily with mkisofs. Below are the exact steps that I ran:
On the NGNCX host.
- mkdir /tmp/drivers
- cd /tmp/drivers
- wget http://mirror.rackspace.com/elrepo/elrepo/el8/x86_64/RPMS/kmod-be2net-18.104.22.168-2.el8_0.elrepo.x86_64.rpm
- mkisofs -o drivers.iso /tmp/drivers/
- mv drivers.iso /usr/share/nginx/html/
I then connected to the ilo and mounted the iso that was on my NGNX server. I then did the following
- mkdir /mnt/cd # create a directory for the iso that we created
- mount /dev/sr0 /mnt/cd
- cd /mnt/cd
- rpm -i mkod_be2.rpm
- modprobe be2net
- ip a s # This now showed both network cards.
I did a reboot and sure enough on reboot my network cards were there.