[CentOS] Trouble with kernel-3.10.0-1160.80.1.el7.x86_64

Wed Nov 16 19:22:46 UTC 2022
Bill Gee <bgee at campercaver.net>

On the other hand ---  I just upgraded one of my CentOS7 systems to 
3.10.0-1160.80.1 today.  With that kernel it will not run any KVM guest. 
  The system hangs - HARD - within one or two seconds of issuing the 
"virsh start ..." command.  It hangs so hard that a power-off restart 
using the back panel switch is required.  It does not log any error I 
can find, nor display any message about a kernel panic.

Rebooting with 3.10.0-1160.76.1 works fine.

The system is an older SuperMicro mainboard with a Q9400 processor.

===============
Bill Gee

On 11/16/22 12:38, Chris Schanzle via CentOS wrote:
> On 11/15/22 7:50 PM, Petko Alov wrote:
>>>
>>> On 2022-11-08 15:49, Orion Poplawski wrote:
>>>
>>>     On 11/8/22 13:12, Simon Matter wrote:
>>>
>>>             Is anyone else experiencing trouble with
>>>             kernel-3.10.0-1160.80.1.el7.x86_64?
>>>
>>>             I'm seeing a kernel panics in the kvm module on one of 
>>> our VM hosts with
>>>
>>>             it.
>>>
>>>             I did notice a new libvirt update as well, but it seems 
>>> to work fine with
>>>
>>>             the
>>>             older kernel (.76.1).
>>>
>>>         Where did you get the .80.1 kernel from? I'm a bit confused 
>>> because I can
>>>
>>>         only see .76.1 on my systems.
>>>
>>>         Simon
>>>
>>>     I'm actually running Scientific Linux, which seems to be a little 
>>> ahead here.
>>> Probably not related, but vmlinuz-4.18.0-372.32.1.el8_6.x86_64 
>>> (AlmaLinux 8.6) had a kernel panic on a Intel Xeon E5504 processor, 
>>> but works fine on Xeon E56XX processors (5620 specifically in our 
>>> tests). * Believe from kernel version, the original email is for 
>>> Centos 7, possible the same change that went into the EL8 kernel went 
>>> also to EL7 Triggered right after I used virsh start to start the vm. 
>>> (Caused Black Screen, and reboot. Found this in the crashed kernel 
>>> logs on /var/crash) Works fine on 
>>> vmlinuz-4.18.0-372.26.1.el8_6.x86_64 and 
>>> vmlinuz-4.18.0-372.16.1.el8_6.x86_64 
>>
>> In fact, related - on our systems attempt to start qemu-kvm VM under 
>> kernel-3.10.0-1160.80.1.el7.x86_64 freezes any of 5 workstations with  
>> dual E5507 (all worked OK under kernel-3.10.0-1160.76.1.el7.x86_64 and 
>> any previous version). The workstations with E5-2609, E5-2650 or 
>> E5-2630 are not affected - all of them run qemu-kvm VM under 
>> kernel-3.10.0-1160.80.1.el7.x86_64 without problems.
>>
> 
> Thankfully, no problems with 2x Xeon E5-2667 v2 and CentOS 7.9 kernel 
> 3.10.0-1160.80.1.el7.x86_64.
> 
> _______________________________________________
> CentOS mailing list
> CentOS at centos.org
> https://lists.centos.org/mailman/listinfo/centos