CentOS

Hi guys,
after update this morning my VM's don't boot with kernel config-3.10.0-693.21.1.el7.x86_64.

I don't have any nVidea drivers installed. It's a server machine (apache only) with minimal install.

I can boot with previews version: config-3.10.0-693.17.1.el7.x86_64

This machine, when boots, keeps with CPU and Memory at 100%.

How can i solve it?

If this is a xen VM then it's a previously reported problem (one without a solution other than use the older kernel).

Yes,
it's a XenServer 7.1 in a PowerEdge R430.

The strange thing is that with one XenServer it starts, and with other host it won't boot.

There are two identical VM's (cloned) in two different hosts - maybe because one Xen is patched (the one that works) and the other is not updated.

Well, i will wait then for a new kernel.

TrevorH wrote:If this is a xen VM then it's a previously reported problem (one without a solution other than use the older kernel).

Hi Trevor, do you know if will be a new kernel to solve this issue?

Because it will require a new license to update my XenServer host. So it would be easier update VM's kernel.

I've been checking. Are you sure you were running kernel-3.10.0-693.17.1.el7.x86_64 and not the "plus" kernel, kernel-plus-3.10.0-693.17.1.el7.centos.plus.x86_64? As far as I can gather, there is an upstream bug that was introduced with the meltdown/spectre fixes that causes all non-plus kernels to crash on xen PV guests. There is a fix for it in the plus kernel series maintained by CentOS but not in the mainstream one yet.

TrevorH wrote:I've been checking. Are you sure you were running kernel-3.10.0-693.17.1.el7.x86_64 and not the "plus" kernel, kernel-plus-3.10.0-693.17.1.el7.centos.plus.x86_64? As far as I can gather, there is an upstream bug that was introduced with the meltdown/spectre fixes that causes all non-plus kernels to crash on xen PV guests. There is a fix for it in the plus kernel series maintained by CentOS but not in the mainstream one yet.

Hi Trevor, I'm sure. I'm not running "plus" kernel - I saw this meltdown/spectre issue, but I'm running an HVM VM (it's not PV).



[root@]# rpm -qa | grep kernel

kernel-3.10.0-514.21.1.el7.x86_64
kernel-3.10.0-514.26.2.el7.x86_64

kernel-3.10.0-693.21.1.el7.x86_64
kernel-3.10.0-693.17.1.el7.x86_64
kernel-3.10.0-693.11.6.el7.x86_64

kernel-headers-3.10.0-693.21.1.el7.x86_64
kernel-tools-3.10.0-693.21.1.el7.x86_64
kernel-tools-libs-3.10.0-693.21.1.el7.x86_64

Just to note:

Freezes: kernel-3.10.0-693.21.1.el7.x86_64


3.10.0-693.17.1.el7.x86_64 runs alright.

Do you have a Redhat subscription? If so you can download a RHEL iso image and try that and see if it gets the same error when that is updated from their repos. If you don't have one then you can sign up for a developer subscription for nothing and get access that way.

For more information on the free Red Hat Developer subscription please see http://developers.redhat.com/blog/2016/ ... available/

Also, can you remove rhgb quiet from the kernel command line and boot like that so you see all the kernel messages. If it does it again, take a screenshot and post it to an image hosting site and link to it (or here if it's not too big).

TrevorH wrote:Also, can you remove rhgb quiet from the kernel command line and boot like that so you see all the kernel messages. If it does it again, take a screenshot and post it to an image hosting site and link to it (or here if it's not too big).

Hi Travor, i've made a test:

I have migrated this VM to the XenServer host that is patched and it worked with this kernel (21).

Then I have migrated it back to XenServer host that it's not patched and the problem come back. I will try to update this host later, because now i cannot stop this since it's a production server.

Heres the boot without quiet and graphical:

https://ibb.co/bsqUkH

It freezes after this smptboot line.