Home > 12c, Exadata, Oracle > Linux Kernel Upgrade on Exadata(manual way)

Linux Kernel Upgrade on Exadata(manual way)

Kernel upgrade can be applied node by node on exadata so there will be no service interruption. Kernel upgrades are required when you need new functionality or when you hit bugs on the current kernel version. I had to upgrade kernel of a box. It is a good experience and The following procedure is based on kernel upgrade on Oracle Linux 5.8 with Unbreakable Enterprise Kernel [2.6.32], a compute node of exadata.

PRE-UPGRADE

==> If you have EM12C the targets on the host will be unavailable for upgrade period. Put them in blackout state so that no false alarms generated from them.

==> Run the upgrade step on X-windows like vnc. This will prevent any disconnection issues from ssh clients.

==> Disable all NFS mounts on the system. check the locations /etc/rc.local , /etc/fstab

==> Is there any asm operations going on the system. Wait for them to finish. Make sure no rebalance job is running on the ASM part. check v$asm_operation.

==> Backup the grup startup file /boot/grub/grub.conf. you might need it for rollback.

==> Shutdown the crs and disable crs auto start. Also shutdown any databases or listeners that are not registered with the csr.
[root@host1 ~]# /u01/app/11.2.0.3/grid/bin/crsctl disable crs
CRS-4621: Oracle High Availability Services autostart is disabled.
[root@host1 ~]# /u01/app/11.2.0.3/grid/bin/crsctl stop crs -f

==> Make sure crs is not running
[root@host1 ~]# ps -ef | grep d.bin
root 66664 60395 0 09:55 pts/1 00:00:00 grep d.bin

==> Reboot the system and make sure it is able to restart before any kernel changes 🙂

==> Check the ilom problem page and make sure there is no problem on the server. If there are any like memory problems etc. fix them.

==> Record the current kernel
[root@host1 ~]# uname -r
2.6.32-400.11.1.el5uek

==> Check the server version and make sure the next kernel is designed for the server.
[root@host1 ~]# dmidecode -s system-product-name
SUN FIRE X4170 M3

==> Make sure enough space is available
[root@host1 ~]# df -h

==> Shutdown any database or listeners that hasn’t been registered with the crs. check the crs for the last time.
[root@host1 ~]# ps -ef | grep d.bin
root 66664 60395 0 09:55 pts/1 00:00:00 grep d.bin

UPGRADE
==> upgrade the kernel

[root@host1 ~]# rpm -ivh kernel-uek-firmware-2.6.32-400.34.1.el5uek.noarch.rpm kernel-uek-2.6.32-400.34.1.el5uek.x86_64.rpm ofa-2.6.32-400.34.1.el5uek-1.5.1-4.0.58.1.x86_64.rpm
Preparing… ########################################### [100%]
1:kernel-uek-firmware ########################################### [ 33%]
2:kernel-uek ########################################### [ 67%]
3:ofa-2.6.32-400.34.1.el5########################################### [100%]

==> Reboot the system
[root@host1 ~]# reboot

POST-UPGRADE
==> Check ilom for any errors. Check /var/log/messages for any errors.

==> Check the new kernel version
[root@host1 ~]# uname -r
2.6.32-400.34.1.el5uek

==> Start the crs and enable crs auto start
[root@host1 ~]# /u01/app/11.2.0.3/grid/bin/crsctl enable crs
CRS-4622: Oracle High Availability Services autostart is enabled.
[root@host1 ~]# /u01/app/11.2.0.3/grid/bin/crsctl start crs

==> Check if crs is starting

[root@host1 ~]# ps -ef | grep d.bin
root 11852 1 4 10:22 ? 00:00:00 /u01/app/11.2.0.3/grid/bin/ohasd.bin reboot
oracle 12013 1 0 10:22 ? 00:00:00 /u01/app/11.2.0.3/grid/bin/oraagent.bin
oracle 12025 1 0 10:22 ? 00:00:00 /u01/app/11.2.0.3/grid/bin/mdnsd.bin
oracle 12109 1 1 10:22 ? 00:00:00 /u01/app/11.2.0.3/grid/bin/gpnpd.bin
root 12119 1 0 10:22 ? 00:00:00 /u01/app/11.2.0.3/grid/bin/orarootagent.bin
oracle 12122 1 1 10:22 ? 00:00:00 /u01/app/11.2.0.3/grid/bin/gipcd.bin
root 12137 1 1 10:22 ? 00:00:00 /u01/app/11.2.0.3/grid/bin/osysmond.bin
root 12150 1 0 10:22 ? 00:00:00 /u01/app/11.2.0.3/grid/bin/cssdmonitor
root 12167 1 0 10:22 ? 00:00:00 /u01/app/11.2.0.3/grid/bin/cssdagent
oracle 12169 1 1 10:22 ? 00:00:00 /u01/app/11.2.0.3/grid/bin/diskmon.bin -d -f
oracle 12187 1 2 10:22 ? 00:00:00 /u01/app/11.2.0.3/grid/bin/ocssd.bin
root 12389 10620 0 10:23 pts/0 00:00:00 grep d.bin
[root@host1 ~]#

==> Enable any NFS mount on the system and mount them

==> On EM12c end the blackout period for the targets.

Now you can move on the other server in the cluster.

Advertisements
  1. December 15, 2014 at 5:24 am

    Reblogged this on Dinesh Ram Kali..

  1. No trackbacks yet.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: