SOLVED cPanel 66 - Reboot Message in WHM

[chrismfz]

Just upgraded to 66 (current).
No matter what I do I am seeing in top right (WHM) the message:
You must reboot the server to apply kernel updates.

of course there is no kernel update.
But rebooted anyway, the message just stays there.

[root@mars ~]# yum clean all
...
[root@mars ~]# yum update
...
No packages marked for update

[root@mars ~]# uptime
03:49:43 up 53 min, 1 user, load average: 0.24, 0.14, 0.09

 

[cPanelMichael]

Hello @chrismfz,

Could you run the following commands and let us know the output?

cat /var/run/system_needs_reboot.cache
uname -r
rpm -q kernel --last
cat /usr/local/cpanel/version
cat /etc/redhat-release
/usr/local/cpanel/bin/envtype
kcarectl --uname
uptrack-uname -a

The last two commands will only work if the system uses KernelCare or KSplice.

Thank you.

 

[chrismfz]

Seems like something "cutting" the version in half and confuses it..

[root@mars ~]# cat /var/run/system_needs_reboot.cache
{"details":{"kernel":{"boot_version":"3.10.0-614.10.2.lve1.4.56.el7.x86_64","running_version":"614.10.2.lve1.4.56.el7"}},"needs_reboot":1}

[root@mars ~]# uname -r
3.10.0-614.10.2.lve1.4.56.el7.x86_64

[root@mars ~]# rpm -q kernel --last
kernel-3.10.0-614.10.2.lve1.4.56.el7.x86_64 Mon 24 Jul 2017 09:15:41 PM EEST
kernel-3.10.0-514.26.2.el7.x86_64 Mon 24 Jul 2017 05:55:00 PM EEST
kernel-3.10.0-514.el7.x86_64 Mon 24 Jul 2017 05:49:23 PM EEST

[root@mars ~]# cat /usr/local/cpanel/version
11.66.0.4
[root@mars ~]# cat /etc/redhat-release
CloudLinux release 7.3 (Yury Malyshev)
[root@mars ~]# /usr/local/cpanel/bin/envtype
standard[root@mars ~]# kcarectl --uname
614.10.2.lve1.4.56.el7
[root@mars ~]#

The same in security advisor:
The system kernel is at version “614.10.2.lve1.4.56.el7”, but is set to boot to version “3.10.0-614.10.2.lve1.4.56.el7.x86_64”.

 

[cPanelMichael]

# kcarectl --uname
614.10.2.lve1.4.56.el7

The output format of the "kcarectl --uname" command on your system doesn't appear to match what I'm seeing on test systems. On your system, it's not including the full kernel version number (i.e. it shows 614.10.2.lve1.4.56 instead of 3.10.0-614.10.2.lve1.4.56).

Could you open a support ticket using the link in my signature so we can take a closer look to determine why the "kcarectl --uname" output is different on this system?

Thank you.

 

[chrismfz]

Just opened it!

Ticket ID: 8743351

 

[cPanelMichael]

Hello,

To update, this relates to KernelCare and how it's providing the kernel version in the "/usr/bin/kcarectl --uname" output. CloudLinux is addressing the issue through case number KCARE-435. You can monitor the KernelCare blog to see when an update is published:

CloudLinux Blog

I'll also update this thread once CloudLinux publishes a resolution.

Thank you.

 

[cPanelMichael]

Hello,

CloudLinux published a KernelCare update to address this issue earlier today. The update ensures the "/usr/bin/kcarectl --uname" command outputs the kernel version in the correct format, and thus fixes the false "You must reboot the server to apply kernel updates" warning message that appears in WHM.

Thank you.

 

[wswd]

I'm running the latest KernelCare but the problem still persists. Any other ideas?

 

[wswd]

The fix is apparently to remove /run/system_needs_reboot.cache

Certainly fixed it.

 

[cPanelMichael]

I'm running the latest KernelCare but the problem still persists. Any other ideas?

Could you let us know the output from the "/usr/bin/kcarectl --uname" command on the affected system? Removing the "/run/system_needs_reboot.cache" file might solve the issue temporarily, but ultimately "/usr/bin/kcarectl --uname" will need to output the version in the correct format to ensure the warning doesn't appear again.

Thank you.