Kernel Panic On 6.8 Release With Kernel-2.6.32-642.el6.x86_64
Hi all,
Just tested the new CentOS 6.8 release on a few VMs and a physical box and got a kernel panic.
Is it just me, or anybody else as well?
Just fishing for now, will look into the issue more thoroughly later today.
19 thoughts on - Kernel Panic On 6.8 Release With Kernel-2.6.32-642.el6.x86_64
Include the actual kernel panic you’re seeing, along with the hardware you’re running it on, or this thread is going nowhere.
jh
and today. you’re
Hi and thanks for the reply.
It was a general question, therefore no details (yet).
–HiDGNknbEp97D5bkworSj22h6ovgP2j0w Content-Type: text/plain; charset=windows-1252
Content-Transfer-Encoding: quoted-printable
Did you see it in every case or just one case?
That kernel has been out there since May 16, 2016 (in CentOS CR, 10 days):
https://lists.CentOS.org/pipermail/CentOS-cr-announce/2016-May/002855.html
There are a couple issues with that kernel on an internet search:
https://access.redhat.com/solutions/2314491
https://github.com/ntop/PF_RING/issues/96
https://www.virtualbox.org/ticket/15411
https://bugzilla.redhat.com/show_bug.cgi?id=1333697
But I don’t see any other issues in either the Red Hat or CentOS bugs databases and none of those are oops.
–HiDGNknbEp97D5bkworSj22h6ovgP2j0w
Hi,
Of the four machines tested, two failed with a kernel panic. At the time, I just rebooted to a previous kernel, as a quick fix.
I found some of the issues you list already.
I’ve run the CR-updates on a few test-machines since they became available and have had no problem. Weird…
I’ve tested some more now, and it seems the upgrade went fine on my Virtualbox Win7-host for the test-guests. No kernel-panic, all seems well. The ones that failed were all Hyper-V guests. I need to pick a sacrificial physical computer and see if it’s reproducible.
For now, trouble-shooting and testing on Hyper-V as I write this.
Hello Sorin,
Thursday, May 26, 2016, 1:58:29 PM, you wrote:
Im just updated VM from 6.7 on 6.8 without problems. KVM hypervisor CentOS6.7
Thanks for the feedback.
Right now I’m looking at the specific Hyper-V guest I saw the kernel panic on before. There might be some weirdness going on with NFS, need to test some more though.
The other test-guest on Hyper-V I upgraded to CentOS 6.8 initially had a problem booting up, but a reset seems to have fixed. Works fine now.
–5sv7mI8uA12ExfpmV06kwEdXRP1HH2w2a Content-Type: text/plain; charset=windows-1252
Content-Transfer-Encoding: quoted-printable
If lockd is failling to start:
https://bugs.CentOS.org/view.php?id=10927
–5sv7mI8uA12ExfpmV06kwEdXRP1HH2w2a
Uninstalled NFS, as it wasn’t needed anyway.
Still got the panic.
Sorin Srbu wrote:
I haven’t been following this thread closely, but if it’s 7, have you tried booting it from a rescue (pxe/dvd/usb key), and running journalctl?
mark
Seriously, what panic?
jh
Hi!
It’s CentOS 6.8., but I’ll keep your hint re 7 in the goodie-bag.
Thanks.
The one I posted a screen dump of earlier.
Did some more digging this morning.
All physical machines are quite okay with the 6.8 upgrade.
All virtual machines running on Hyper-V get a kernel panic – if the linux integration services (lis) are installed on the guest.
Uninstalling lis, then applying the kernel-2.6.32-642.el6.x86_64 and rebooting works just fine. No kernel panics.
Supposedly (according to some forum posts I read) the lis shouldn’t care what kernel is actually running, as long as it is compiled with a kernel that is installed and available on the system, in this case the last previous before the 6.8-upgrade.
In any case, problem solved for now.
Thanks for all the feedback!
Marking thread as solved.
I don’t see any posts of a screen dump. The mailing list software (mailman) is set to strip any non-text attachments, so if it was an image, it was probably stripped.
(The thread starts here https://lists.CentOS.org/pipermail/CentOS/2016-May/159655.html )
—
Jonathan Billings
Oh, wasn’t aware of that. My apologies.
Screendump available here:
https://uppsala.box.com/s/iwoh20xsayywbetfdd7uh8vs6hcpmqxt
To recap, the problem was with the linux integration services. Uninstalling them, updating the guest as usual and then rebooting the guest, fixed the panic. Haven’t seen this problem before.
Installer kernel panics on ThinkCenter i7 Haswell desktop and Intel NUC i6260, but works per normal on Thinkpad W530.
I’ve got same issue and found out the hyperv linux integration services kmod-microsoft-hyperv and microsoft-hyper packages are conflicting with with Kernel-2.6.32-642.el6.x86_64 , removing these packages will allow latest kernel to work. Looks like Microsoft LIS haven’t got a compatible version yet for centos 6.8
https://technet.microsoft.com/en-us/library/dn531026.aspx
Same here… Kernel-2.6.32-642.3.X.el6.x86_64 not compatible with kmod-microsoft-hyperv and microsoft-hyper packages
I’m also having a problem with the 6.8 release, and am pretty sure I do something wrong.
My setup: A HP DL380 Gen9, CentOS-6.8-x86_64-LiveCD.iso (being mounted as virtual media, but have the same effect when installing it onto the internal sd card).
Leads to kernel Panic “Atempted to kill init! exitcode=0x00000100”
I’ve already updated all firmware code of the server to the newest version (about 2-3 month old)
As this is – to my opinion – a setup that could not be more standard, I’m really asking myself whats wrong. Using CentOS 7 Live CD works fine, used to have Oracle VM on that machine (which has Oracle Linux in the Dom0, which is a flavour of RedHat) also works fine (actually first ssen the problem when trying to update to Oracle VM 3.4.2, but then drilled down the problem to as much a standard thing as possible).
Any Idea / Hint where I could search ?