Cloudlinux kernel does not work on Intel Xeon D CPU (X10SDV-4C-TLN2F motherboard)

[hollowheartx]

Hello,
Were have started new shared hosting company in our area. We wanted to use cloudlinux on new Intel Xeon D 1520/1540 servers (mother board: X10SDV-4C-TLN2F ). Howerver motherboard/CPU does not support cloudlinux. The server does not boot up after changing kernel to cloudlinux. The server also does not boot up on original kernel provided by CentOS (both 6, 7). The server currently using customized kernel provided by our datacenter OVH.

Please advise.

[cloudlinux]

Hi,

Just my 2 cents:
Intel Xeon D are for networking,storage and microservers. They are the succesor of the Intel Atom Avonton.
To give you an idea of the positioning of Intel of the chips:

http://images.anandtech.com/doci/9185/IntelPositioningXeonD.png

So if possible switch to a Xeon E5 for now. It will give you less headache for now. Those are well supported and great processors. (And more powerfull)

[786host.pk]

Thanks for the information. Actually Xeon D servers are half of the price of Xeon E5. The cost matters for startups. I was wondering if cloudlinux team was able release a small fix for Xeon D5 series.

[786host.pk]

I contacted CloudLinux support via help desk and this is the reply I got. I hope it is fixed soon. (I am pasting the response below so that any one who is facing similar problem do know about it).
----------------
Hello.

Yes, we know about this issue.
In the next 2-3 days, this problem will be solved (with OVH server + CL7).

Now we have a temporary solution (but I can not guarantee that it can work on all servers):

"The cloudlinux kernel is naming ethernet devices with different order so the association with mac address is not correct. Example:

cloudlinux | ovh
eth0 -> eth2 (10gbps port)
eth1 -> eth3 (10gbps port)
eth2 -> eth0 (1gbps port)
eth3 -> eth1 (1gbps port)

Yes, eth2 es the main ethernet device with conectivity. Only ovh knows.

Using "alias eth2 ixgbe" for /etc/modules-load.d/ixgbe.conf allows to asign what interface should using the driver"

[info]

I just installed cloudlinux in a server just like that with Centos 6 + Cl6. You have to download the ixgbe source drivers and recompile them with the cloudlinux kernel.

After cloudlinux install before rebooting.

yum install kernel-devel

And follow the instructions in:

http://ask.xmodulo.com/compile-ixgbe-driver-centos-rhel-fedora.html

When you reboot the boot will fail but you can access via IPMI and install the kernel module.

Regards

[bogdan.sh]

Is server not booting, or it boots but no network? What does console says when connected over KVM?

Yes we are preparing ixgbe drivers to be released soon, however if you want them faster you may want to install them from source, here is one that looks to be related to your motherboard: https://downloadcenter.intel.com/download/14687

[786host.pk]

@Bogdan: Actually I didnt look into it via IPMI. The server never came online after rebooting. The OVH monitoring did not detect fault for some time means that most probably the network configuration issue but I am not sure.

I do not want to take risk of compiling things manually for cloudlinux. I am waiting for cloudlinux to release a fix so that we can start proper configuration. The support told that it would not take more than few days (2-3).

Thank you in advance.

[jmanuel]

We are having the same problem, can run CloudLinux on the new servers provided by OVH.

Any idea where we can be notified of the ixgbe drivers?

[bogdan.sh]

Looks like you have to install ixgbe driver, we have it in repos since Sep 22, so please run

yum install kmod-ixgbe

[patrickfetue]

Hi

I am having the same issue,

I install cloudlinux on top of my cpanel and since 2 days I cannot access my server. customer are complaining

I can just access the website in rescue mode, and I dont know anything in linux

How can I get my server back online either remove cloudlinux and fix it

Please provide me the step by step procedure

thank you

[bogdan.sh]

Its hard to say what is the core of the issue, the fastest way to help is to create support ticket and provide access to rescue console.

Also, boot process screenshot could help, where it actually fails.