Sun Microsystems, Inc.  Oracle System Handbook - ISO 7.0 May 2018 Internal/Partner Edition
   Home | Current Systems | Former STK Products | EOL Systems | Components | General Info | Search | Feedback

Asset ID: 1-72-1537133.1
Update Date:2017-07-05
Keywords:
Solution Type  Problem Resolution Sure

Solution  1537133.1 :   ODA : Removing Old Kernel to Reclaim Disk Space on mount point: /boot "/boot is (XX%)"  

Related Items 
    

  • Oracle Database Appliance
    •  
  • Oracle Database Appliance Software
    •  
  • Oracle Database Appliance X5-2
    •  
Related Categories 
    

  • PLA-Support>Eng Systems>Exadata/ODA/SSC>Oracle Database Appliance>DB: ODA_EST
    •  
  • _Old GCS Categories>ST>Server>Engineered Systems>Oracle Database Appliance>Patching
    •  




ODA updates can lead to boot percent available warnings in version 2.1.0.x up to 2.3.0.x - Two enhancements  will help avoid this problem. In 2.4 we avoid this warning by cleaning up unneeded files from previous patch attempts. In 2.6 we will also increase the available boot area from 150M to 500M

Applies to:

Oracle Database Appliance - Version All Versions to All Versions [Release All Releases]
Oracle Database Appliance X5-2 - Version All Versions to All Versions [Release All Releases]
Oracle Database Appliance Software - Version 2.1.0.1 to 12.1.2.7 [Release 2.1 to 12.1]
Information in this document applies to any platform.
/boot , ODA , /boot cleanup, reclaiming space, ODA boot

Symptoms

There are different times when you may find /boot needs more usable space

  • Installing a patch / upgrade on ODA and you find the message 

       Disk Space Usage on mount point: "/boot is (xx%)" full

  • Running pre-patch checks for newer ODA versions may report /boot as having insufficient space

  • Running df -k on both nodes may show /boot has under 50m free

    df -k

 

Changes

 Upgrade of the OS on Oracle Database Appliance (ODA) on due to patching

Cause

 A relatively small allocated boot space allocated during the deployment of 100m by OS binaries from a few ODA versions
 Regular patching decreases usable space in /boot due to the collection of several no longer used kernel versions

Solution

 As there is no cleanup of the previously installed or attempted installed versions the boot area
can throw warnings that the boot area is getting FULL

    e.g. Disk Space Usage on mount point: "/boot is (84%)"

The recommended resolution is to go to the most current available patchset where

  • An enhancement in 2.4 cleans up some (but not all) old unneeded kernel versions
  • An enhancement in 2.6 increases the /boot size to 500M *

 * bug detected for 1 newer version where /boot was resized to /100m
This can be avoided using manual steps and has been corrected for newer versions.

ACTION

You can manually cleanup the previous binaries which may include versions from failed patching attempts

In some cases you can easily identify older kernel version due to updates of the same kernel
These are identified with img.dup_orig 

 

Example 2 from older ODA versions

#root@ /boot
#ls -al
...
-rw------- 1 root root 3920573 Dec 15 20:09 initrd-2.6.32-300.11.1.el5uek.img
-rw------- 1 root root 3927510 Aug 28 2012  initrd-2.6.32-300.11.1.el5uek.img.dup_orig    < remove the duplicates
-rw------- 1 root root 9432195 Dec 15 20:10 initrd-2.6.32-300.11.1.el5uekkdump.img
...
-rw------- 1 root root 3926910 Dec 15 20:08 initrd-2.6.32-300.32.4.el5uek.img.dup_orig    < remove the duplicates
-rw------- 1 root root 9434725 Dec 15 21:01 initrd-2.6.32-300.32.4.el5uekkdump.img   


#root@ /boot]  rm initrd-2.6.32-300.11.1.el5uek.img.dup_orig
#root@ /boot]  rm initrd-2.6.32-300.32.4.el5uek.img.dup_orig

..etc

 
Example 1

1. Check current kernel being used

 uname -a

2. List kernels that have been installed

 rpm -qa | grep kernel

3. List the files present in the boot directory

 ls -al /boot

4. Remove old kernel - take care not to delete the current kernel that was identified in the first step

 rpm -e <old_kernel>

5. Remove older kernels until you have about 50MB space in the boot partition. Under normal circumstances, removing just one old kernel may be good enough.
For example,

 [root@ boot]# uname -a
 Linux c1n1 2.6.39-400.276.1.el6uek.x86_64 #1 SMP Wed Jan 27 15:36:25 PST 2016 x86_64 x86_64 x86_64 GNU/Linux

 [root@ boot]# rpm -qa | grep kernel

 kernel-uek-firmware-2.6.39-400.276.1.el6uek.noarch
 kernel-uek-2.6.39-400.264.13.el6uek.x86_64
 kernel-uek-firmware-2.6.39-400.264.13.el6uek.noarch
 dracut-kernel-004-388.0.1.el6.noarch
 kernel-transition-2.6.32-0.0.0.1.el6.x86_64
 kernel-uek-2.6.39-400.276.1.el6uek.x86_64
 kernel-headers-2.6.32-573.18.1.el6.x86_64

 [root@ boot]# ls
 System.map-2.6.39-400.264.13.el6uek.x86_64 efi initrd-2.6.39-400.276.1.el6uek.img vmlinuz-2.6.39-400.264.13.el6uek.x86_64
 System.map-2.6.39-400.276.1.el6uek.x86_64 grub lost+found vmlinuz-2.6.39-400.276.1.el6uek.x86_64
 config-2.6.39-400.264.13.el6uek.x86_64 initramfs-2.6.39-400.264.13.el6uek.x86_64.img symvers-2.6.39-400.264.13.el6uek.x86_64.gz
 config-2.6.39-400.276.1.el6uek.x86_64 initramfs-2.6.39-400.276.1.el6uek.x86_64.img symvers-2.6.39-400.276.1.el6uek.x86_64.gz

 [root@ boot]# rpm -e kernel-uek-2.6.39-400.264.13.el6uek.x86_6

[root@c1n1 boot]# ls
 System.map-2.6.39-400.276.1.el6uek.x86_64 efi initramfs-2.6.39-400.276.1.el6uek.x86_64.img lost+found vmlinuz-2.6.39-400.276.1.el6uek.x86_64
 config-2.6.39-400.276.1.el6uek.x86_64 grub initrd-2.6.39-400.276.1.el6uek.img symvers-2.6.39-400.276.1.el6uek.x86_64.gz

 Also - kdump related files can be moved when no rpms are available to be deleted

initrd-2.6.39-400.264.13.el6uek.x86_64kdump.img
initrd-2.6.39-400.282.1.el5uekkdump.img

  

 

 


Attachments 
This solution has no attachment
  Copyright © 2018 Oracle, Inc.  All rights reserved.
 Feedback