7

Today I was running some yum updates and wanted to verify that everything went through fine by making sure I had a new kernel. To my surprise I noticed that CentOS was still running 2.6.32-431.5.1.el6.x86_64 even though it looked as though 2.6.32-431.23.3.el6 was installed.

Indeed 2.6.32-431.23.3.el6 shows up in /etc/grub.conf but not in the upstart boot options. Any ideas why?

In the update log it says:

---> Package kernel-firmware.noarch 0:2.6.32-431.5.1.el6 will be updated
---> Package kernel-firmware.noarch 0:2.6.32-431.23.3.el6 will be an update

Could this be the reason? What does "will be an update" mean?

My /etc/grub.conf:

# grub.conf generated by anaconda
#
# Note that you do not have to rerun grub after making changes to this file
# NOTICE:  You have a /boot partition.  This means that
#          all kernel and initrd paths are relative to /boot/, eg.
#          root (hd0,0)
#          kernel /vmlinuz-version ro root=/dev/mapper/VolGroup00-root
#          initrd /initrd-[generic-]version.img
#boot=/dev/vda
default=0
timeout=5
splashimage=(hd0,0)/grub/splash.xpm.gz
hiddenmenu
password --encrypted $1$auui(i$sODM4ni/Zts9IlMWu.wWF/
title CentOS (2.6.32-431.23.3.el6.x86_64)
    root (hd0,0)
    kernel /vmlinuz-2.6.32-431.23.3.el6.x86_64 ro root=/dev/mapper/VolGroup00-root rd_NO_LUKS LANG=en_US.UTF-8  KEYBOARDTYPE=pc KEYTABLE=sv-latin1 rd_NO_MD rd_LVM_LV=VolGroup00/swap SYSFONT=latarcyrheb-sun16 crashkernel=auto rd_LVM_LV=VolGroup00/root rd_NO_DM rhgb quiet rhgb quiet audit=1
    initrd /initramfs-2.6.32-431.23.3.el6.x86_64.img
title CentOS (2.6.32-431.5.1.el6.x86_64)
    root (hd0,0)
    kernel /vmlinuz-2.6.32-431.5.1.el6.x86_64 ro root=/dev/mapper/VolGroup00-root rd_NO_LUKS LANG=en_US.UTF-8  KEYBOARDTYPE=pc KEYTABLE=sv-latin1 rd_NO_MD rd_LVM_LV=VolGroup00/swap SYSFONT=latarcyrheb-sun16 crashkernel=auto rd_LVM_LV=VolGroup00/root rd_NO_DM rhgb quiet rhgb quiet audit=1
    initrd /initramfs-2.6.32-431.5.1.el6.x86_64.img
title CentOS (2.6.32-431.el6.x86_64)
    root (hd0,0)
    kernel /vmlinuz-2.6.32-431.el6.x86_64 ro root=/dev/mapper/VolGroup00-root rd_NO_LUKS LANG=en_US.UTF-8  KEYBOARDTYPE=pc KEYTABLE=sv-latin1 rd_NO_MD rd_LVM_LV=VolGroup00/swap SYSFONT=latarcyrheb-sun16 crashkernel=auto rd_LVM_LV=VolGroup00/root rd_NO_DM rhgb quiet rhgb quiet audit=1
    initrd /initramfs-2.6.32-431.el6.x86_64.img
Kristofer
  • 301
  • 1
  • 3
  • 12

3 Answers3

4

Found the problem... Apparently CentOS 6 does not properly sync /etc/grub.conf with /boot/grub/grub.conf so /boot/grub/grub.conf which read during boot. Fixed with:

rm /boot/grub/grub.conf
mv /etc/grub.conf /boot/grub/ 
ln -s /boot/grub/grub.conf /etc/grub.conf
Kristofer
  • 301
  • 1
  • 3
  • 12
  • I was wondering about that. `/boot/grub/grub.conf` is the correct file, and it should be present there, not in `/etc`. – Michael Hampton Aug 25 '14 at 11:33
  • This seems off. There's always a symbolic link from /etc/grub.conf to ../boot/grub/grub.conf - Did something happen with your /boot partition at some point? – ewwhite Aug 25 '14 at 12:26
  • 1
    I don't believe so, we have noticed this happening on all our CentOS 6 boxes. Could be related to our automated installation somehow but I doubt it as we are not messing around with grub in the install script. – Kristofer Aug 25 '14 at 13:46
2

You have to reboot Linux in order for a new kernel to become active.

If you notice, the newly-installed kernel, version 2.6.32-431.23.3.el6.x86_64, is set to be the default in the grub configuration (default=0)

Did you verify that you ran a system reboot following your updates?

  • Can you check the free space in /boot?
  • You can run a yum reinstall of the kernel packages just to ensure everything is installed properly.
  • If you DO reboot, see what shows up in the GRUB menu after the system POST sequence.
ewwhite
  • 197,159
  • 92
  • 443
  • 809
  • Yes, I'm sure it rebooted, several times even. I even looked at the grub menu which just shows the latter two of the entries. – Kristofer Aug 22 '14 at 14:04
  • How much free space is in /boot ? – ewwhite Aug 22 '14 at 14:06
  • This also happened on multiple machines. I ran a cleanup of old kernels (package-cleanup -y --oldkernels --count=2) before and after update since our boot partitions are small. It has 45MB free after both cleaning and both kernels installed. I did another yum reinstall to no avail... – Kristofer Aug 22 '14 at 14:18
1

Make sure you haven't enabled direct kernel boot to boot a kernel from outside the virtual machine.

Direct kernel boot in virt-manager

The direct kernel boot XML looks like:

  <os>
    <kernel>/var/lib/libvirt/images/vmlinuz-3.15.8-200.fc20.armv7hl</kernel>
    <initrd>/var/lib/libvirt/images/initramfs-3.15.8-200.fc20.armv7hl.img</initrd>
    <cmdline>console=ttyAMA0,115200n8 rw root=/dev/mapper/armv9-root rootwait</cmdline>
    <dtb>/var/lib/libvirt/images/vexpress-v2p-ca9.dtb</dtb>
  </os>

Some providers such as Digital Ocean boot all virtual machines via direct kernel boot, and there's no way around this in such cases.

Michael Hampton
  • 244,070
  • 43
  • 506
  • 972