Jump to content United States-English
HP.com Home Products and Services Support and Drivers Solutions How to Buy
 Contact HP
HP.com home
HP Tru64 UNIX Version 5.1B-2 and Higher: Patch Kit Installation Instructions > Chapter 4 Patching a Cluster

Using the dupclone Script

 

Table of Contents
Glossary
Index
Content starts here

You can install Version 5.1B-4 and higher on your cluster using an installation method generically referred to as cloning that uses a tool named dupclone . The process consists of two primary steps:

  • Creating an exact duplicate of an existing system on an alternate set of disk drives.

  • Using dupclone to install the patch kit to the alternate disk set. After completion, the system can immediately be rebooted using the alternate disks.

NOTE:

The examples in this section assume a two-node cluster that uses one disk drive each for the cluster-wide root, /usr, and /var file systems; one disk drive for each member boot disk; and 1 drive for the quorum disk.

Benefits and Detriments of Cloning

The benefits of cluster cloning include the following:

  • The process can be performed in multi-user mode as a replacement for the no-roll installation method.

  • It does not require the shutdown of any cluster members while the patch kit is being installed

  • It provides an easy way to reverse the installation by simply rebooting on the original disk set.

Reasons why cluster cloning may not be right for your cluster include the following:

  • It requires a complete cluster reboot and therefore will not be useful for systems required to be online 24/7 and rely on the rolling upgrade procedure.

  • It requires additional hardware. You must have enough spare disk drives that equals or exceeds the existing system disk set for the cluster root, /usr, /var, and member boot file systems. This includes the quorum disk. For example, if you have a three-member cluster and use one disk for the root and /var file systems, one disk for the /usr file system, and one disk for the quorum; you will need six additional disk drives to use cloning technology.

Performing the Cloned Installation

To create an exact duplicate of your system on an alternate set of disk drives, you can find a third-party program that will do it for you, or you can use the following list of tasks as a guide for doing it yourself:

  1. Download and extract the latest Tru64 UNIX version or mount the latest Tru64 UNIX version CD to your Version 5.1B-3 or higher system and run the dupatch command to install the latest tools, including the necessary cloning tools. See Chapter 3 for more information on installing the new tools.

  2. Identify the alternate set of disks you will use for the new software

  3. Partition the alternate drives so those partitions are as large or larger than the current drive set

  4. Duplicate all cluster file systems onto the alternate disk set. This includes cluster_root, cluster_usr, and cluster_var.

  5. Duplicate all member-specific root partitions onto the alternate disk set. (that is, root1_domain, root2_domain, and so forth.)

  6. Create the CNX partitions for all member root drives and the quorum disk (if it exists).

Be careful to ensure that the alternate drive CNX partitions reference the disk device number of the alternate drives, not the drives used by the currently running operating system. For example:

A cluster comprised of two members has the following disk setup.

dsk0a: contains the cluster_root file system
dsk0g: contains the cluster_var file system
dsk1g: contains the cluster_usr file system
dsk2a: contains the member1 root partition
dsk2h: contains the member1 CNX partition
dsk3a: contains the member2 root partition
dsk3h: contains the member2 CNX partition
dsk4h: contains the quorum disk partition

Additionally, the cluster has five unused disks available to it (dsk10, dsk11, dsk12, dsk13, dsk14), which are as large as or larger than, the existing system disks.

The spare drives are used as a duplicate, or “clone,” of the existing operating system and they are mounted on a mount point called /clone” The steps are:

  1. Partition dsk10a to be as large or larger than dsk0a

  2. Partition dsk10g to be as large or larger than dsk0g

  3. Partition dsk11g to be as large or larger than dsk1g

  4. Partition dsk12a to be as large or larger than dsk2a

  5. Partition dsk12h for the special size of a CNX partition

  6. Partition dsk13a to be as large or larger than dsk3a

  7. Partition dsk13h for the special size of a CNX partition

  8. Partition dsk14h for the special size of a quorum disk

  9. Build the CNX partitions for dsk12h, dsk13h, and dsk14h

  10. Create AdvFS domains and file sets for all file systems (that is, alt_cluster_root, alt_cluster_usr, alt_cluster_var, alt_boot1, alt_boot2.)

  11. Use common UNIX commands to exactly duplicate existing cluster and member file systems onto the alternate disks.

  12. Mount all file systems in a hierarchical form on the mount point /clone.

When these steps are complete, the mount command will display output similar to the following:

# mount
cluster_root#root on / type advfs (rw)
root1_domain#root on /cluster/members/member1/boot_partition type advfs (rw)
cluster_usr#usr on /usr type advfs (rw)
cluster_var#var on /var type advfs (rw)
root2_domain#root on /cluster/members/member2/boot_partition type advfs (rw)
alt_cluster_root#root on /clone type advfs (rw)
alt_cluster_usr#usr on /clone/usr type advfs (rw)
alt_cluster_var#var on /clone/var type advfs (rw)
alt_boot1#root on /clone/cluster/members/member1/boot_partition
alt_boot2#root on /clone/cluster/members/member2/boot_partition

The first five file systems in this output represent the current operating system version; the last five represent an exact duplicate of the current operating system mounted on an alternate root path named /clone.”

The final step is to apply the latest patch kit to the alternate root using the dupclone command. This will install the new patch kit onto the alternate set of drives, which can then be booted. If for any reason the installation or the reboot fails, you can simply reboot on the original set of disk drives.

dupclone(8) Reference Page

This section reproduces the dupclone(8) reference page, which is installed on your system with the patch installation tools.

NAME
dupclone – Clone Tru64 UNIX cluster
  
SYNOPSIS
dupclone [–r rootpath] [–k kitpath] [–license]
  
OPTIONS
–r rootpath 

Identifies the root directory of the cloned operating system.

–k kitpath 

Identifies the path to the patch kit being installed.

–license 

Indicates that you accept the license agreement. A printed copy of the license agreement is included in the release notes that came with your patch kit.

DESCRIPTION

The dupclone command installs a Tru64 UNIX patch kit located at kitpath into an alternate root directory defined by rootpath. The specified rootpath must be the root directory of a complete identical copy of the existing Tru64 UNIX operating system, including all members of the cluster.

The dupclone command can only be used on clusters. It is not supported for use on stand-alone systems.

EXAMPLES
 

The following example shows a typical cloning operation:

# dupclone -r /clone -k /usr/T64/patch_kit -license
You have accepted the license agreement.
Cluster Name: jungle
Member-1: apeman.zoo.com 10.0.0.1
Member-2: safari.zoo.com 10.0.0.2
Checking patch kit for transmission errors during download...
Verifying current software on member 1
Verifying current software on member 2
Creating master inventory of current system...
Getting list of patch subsets to install...
.
.
.

FILES
cloneroot/usr/adm/dupatch/log/session.log
  

This file captures dupclone activities.

SEE ALSO

Commands: dupatch(8)

Printable version
Privacy statement Using this site means you accept its terms
2006 Hewlett-Packard Development Company, L.P.