Scyld ClusterWare Updates

You can use Yum update to update ClusterWare once you have upgraded your RHEL5 or CentOS 5 base distribution. See the Section called Updating Red Hat or CentOs Installation for details on updating your base distribution, and the Section called Scyld ClusterWare Installation for how to set up the Yum repo configuration files.

To verify which distribution you are currently running, do the following:

[root@scyld ~]# cat /etc/redhat-release

Updating ClusterWare

  1. Clear the Yum update list with the command:

    [root@scyld ~]# yum clean all
  2. Update ClusterWare as follows:

    [root@scyld ~]# yum groupupdate Scyld-ClusterWare
    This will show you the list of rpms to be updated on your system, and prompt you to accept the list before actually updating the system. If there are any packages you do not want to install, you can add the --exclude=$package parameter to the yum groupupdate command. We do not recommend excluding any ClusterWare packages.

    If the Yum update completes successfully, a complete! message is displayed as the last output to the terminal.

    If Yum fails with a Transaction Check Error that complains that a base distribution rpm is newer than the Scyld ClusterWare rpm that is attempting to replace it, then you must manually install the downlevel Scyld ClusterWare rpm(s). For example, if the complaint is about the kernel rpm, then do:

    [root@scyld ~]# cd /var/cache/yum
    [root@scyld ~]# ls cw-*/packages/kernel-*
    and locate the newest ClusterWare kernel, which should reside in either cw-core/packages or cw-updates/packages. Then install that newest kernel, e.g.:
    [root@scyld ~]# rpm -i --oldpackage cw-updates/packages/kernel-*
    and repeat the groupupdate:
    [root@scyld ~]# yum groupinstall Scyld-ClusterWare
    which should now complete! successfully.

  3. Reboot your system.

  4. To verify that ClusterWare was installed successfully, do the following:

    [root@scyld ~]# uname -r
    The result should match the ClusterWare kernel version noted in the Release Notes.

  5. Restart the compute nodes.