Sun Microsystems
Products & Services
 
Support & Training
 
 

Previous Previous     Contents     Index     Next Next
Chapter 7

Patching Software on the Cluster

When you are patching the software on the cluster, consult the patch README and your hardware documentation for any information about patch dependencies or special installation instructions. You can add a patch to the cluster during initial cluster installation with the nhinstall tool by adding the the patch details to the addon.conf file. For further information, see the addon.conf(4) and nhinstall(1M) man pages.

For information about how to manually add a patch to the Foundation Services and Solaris packages, see the following sections:

Introduction to Patching Software on a Cluster

Some Foundation Services packages are shared packages and some are not shared. A patch for a shared package is installed on the master node only. A patch for a nonshared package is installed on all peer nodes for which the patch is relevant.

Before you install a patch, consider the following:

  • Whether you are patching shared or nonshared packages

  • The type of node on which you are installing the patch

    If you are installing a patch on a diskless node, determine whether you are patching a package installed under /usr or root.

  • Whether the cluster was installed by the nhinstall tool or manually.

    The location of the packages for diskless nodes depends on the installation method. Where the location of a package depends on the installation method, it is stated.

Information about the patches is held in the patch database. The information is node-specific for patches that are applied to nonshared packages. The database describes the file system where the patch contents are installed. To view the patches that are currently installed on a node, use the patchadd -p command. For further information, see the patchadd(1M) man page.

Choosing a Procedure to Patch Software on a Cluster

The following packages can be patched. The path to the patch database for each package and a pointer to how to install the patch are also provided.

  • A Solaris package or nonshared Foundation Services package on the master-eligible nodes

    The patch database for these packages is at /var/sadm/install. To install a patch for a Solaris package or a nonshared Foundation Services package on a master-eligible node, see To Patch a Nonshared Package on Both Master-Eligible Nodes.

  • A Solaris package or nonshared Foundation Services package on a dataless node

    The patch database for these packages is at /var/sadm/install on the dataless node. To install a patch on a dataless node, see To Patch a Nonshared Package on Both Master-Eligible Nodes.

  • A Solaris package for the /usr directory common to diskless nodes

  • A Solaris package or Foundation Services package for the root file system of each diskless node

  • Shared packages

    A shared package can be a Foundation Services package or a user application. The patch database for shared packages is at /SUNWcgha/local/export/services. To install a patch on a shared package, see Patching a Shared Package.

Overwriting Modified Foundation Services Files

The following patches overwrite Foundation Services files:

  • A Solaris patch that changes the IP stack. The IP stack is modified during installation of CGTP.

  • A Solaris patch that changes the modified rc2 and rc3 scripts. The modified scripts are required by the Daemon Monitor.

If you install one of these patches, you must reinstall the Foundation Services.

Patching a Nonshared Package on Both Master-Eligible Nodes

This section describes how to patch a nonshared package on the master-eligible nodes. You must install the patch on both master-eligible nodes.

ProcedureTo Patch a Nonshared Package on Both Master-Eligible Nodes

  1. Log in to a master-eligible node as superuser.

  2. Consult the patch README for the patchid and for any constraints on applying this patch.

    Constraints can include the requirement to be in single-user mode, or the requirement to stop the node or the cluster.

  3. Copy the patch to the /var/spool/patch directory.

    The /var/spool/patch directory is the default directory where the patchadd command looks for a patch.

  4. Take steps to comply with any constraints described in the patch README.

  5. Apply the patch.

    • If you have copied the patch to /var/spool/patch, install the patch as follows:

      # patchadd patchid

    • If you have not copied the patch to /var/spool/patch, specify the location of the patch by using the -M option of the patchadd command:

      # patchadd -M patch-directory patchid

  6. Verify that the patch has been installed:

    # patchadd -p

    A list of all patches on the node is displayed on the screen.

    # patchadd -p | grep patchid

    The patch for which you specified the patchid is displayed, if this patch is installed.

  7. Verify that the node is configured correctly:

    # nhadm check

    For more verification procedures, see Chapter 3, Determining Cluster Validity.

  8. Log in to the second master-eligible node as superuser.

  9. Repeat Step 3 through Step 7.

Previous Previous     Contents     Index     Next Next