|
| ||||||||||||||||
| ShaoLin HA Cluster 1.0 Installation Manual and Operations Guide | ||
|---|---|---|
| Prev | Chapter 10. Cluster Administration | Next |
Notice! You should only upgrade to kernels that are provided by your Linux distributor to make sure you get support from both your Linux distributor and ShaoLin HA Cluster. ShaoLin HA Cluster is only certified and tested on vendor supported kernel. Custom compiled kernels will not be supported by ShaoLin HA Cluster.
To perform a kernel upgrade and installation, please follow the steps below.
Before install the kernel package, check with ShaoLin Microsystems to see whether the kernel package to be installed is supported by HA Cluster. You can check the support website for detail information. You should look for the package ShaoLin_HA_Cluster_KHBD-<kernelversion>-<khbd version>.<arch>.rpm which matches your to-be-installed kernel version.
If you've obtained the required packages. At the active server, run the command rpm -ivh path-to-your-kernel-package.rpm path-to-KHBD.rpm to install the two packages in the same transaction. Notice, DO NOT use -U in installing kernel rpm packages. Because upgrading kernel rpm packages will remove your old kernel package files, you will not be able to roll back in case an error occurred.
After successful installation of the new kernel and KHBD package, you have to regenerate the cluster boot images for your newly installed kernel. Go to Cluster Configurator, select . In the combo box, select your freshly installed kernel version. Click to generate the new Cluster initrd image file.
You will have to modify the boot loader configuration file to add a selection for your newly generated initrd image (both normal mode and test mode). DO NOT remove your old kernel entries as it provide you the possibility to rollback to the original kernel configuration in case there is an error.
Reboot your secondary standby cluster to use the new kernel. For the first time startup, you should choose to enter the test mode for safe. After starting the Cluster Manager console, check the status of the cluster . If everything is fine, you can reboot your standby server into a normal mode. If you encounter problems with the new kernel image, check step 3 and step 4 to see if you have mis spelled any initrd image file name . If you cannot fix the problem, you should roll back to your original kernel configuration and consult your local support service provider.
If everything works fine, you should use the standby server to take over the system so that your system is now using the new kernel. Again, if problem exists, you should rollback to the original kernel boot image by selecting a proper boot iamge in the Grub or Lilo boot image selection menu.