NetWorker V7.3, Disaster Recovery Guide

12

Chapter 12: HP-UX Cluster Disaster Recovery

Performing a Cluster-Wide Recovery

This section describes how to perform a complete cluster recovery, including reinstalling the operating system on all nodes in the cluster.

How to Perform a Cluster-Wide Recovery

To recover an entire cluster, perform the following tasks:

•

"Task 1: Reinstall the Software" on page 176

•

"Task 2: Recover the Cluster Database" on page 176

•

"Task 3: Start the Cluster" on page 177

•

"Task 4: Recover the Data" on page 178

Task 1: Reinstall the Software

To reinstall the software:

1. Reinstall the operating system and cluster software on each node.

2. Reinstall the NetWorker software as required within the HP-UX

MC/ServiceGuard cluster environment. For detailed instructions, refer to the EMC NetWorker Installation Guide, HP-UX Version.

Task 2: Recover the Cluster Database

1. Select any node in the cluster and run the NetWorker server and client software on that node.

2. Re-create the cluster configuration and the NetWorker software configuration and control files. The default location for the cluster configuration file is /etc/cmcluster and the default directory for the

NetWorker configuration and control files is /etc/cmcluster/NetWorker.

• If you know where the latest save sets are located, enter the scanner command to recover the configuration files:

scanner -S saveset_id

device_path

| uasm -i y -rv

where device_path is the path to the tape device.

• If you do not know the save set ID, enter the scanner device_path command to determine the saveset ID, as follows:

scanner

device_path

> /tmp/savefile 2>&1

176

NetWorker V7.3

Disaster Recovery Guide

Chapter 12: HP-UX Cluster Disaster Recovery

3. Run the NetWorker NetWorker.cluster script located in the

/opt/NetWorker/bin/ directory.

Note:

Do not re-create the legato.control and pkg.control files when prompted by the NetWorker.cluster script.

4. If the shared disk has to be replaced, configure the disk and file system of the replacement disk.

5. Create the file system, as follows:

vgchange -c n /dev/vg03 vgchange -a y /dev/vg03 newfs -F hfs /dev/vg03/rlvol1 vgchange -a n /dev/vg03

6. Enter the following command to delete the existing cluster and package configurations:

cmdeleteconf -c cluster1

7. Enter the following command to apply the cluster and package configurations, including the NetWorker package, to all the nodes within the cluster:

cd /etc/cmcluster cmapplycomf -C cluster1.ascii -P networker/pkg.conf \

-P other_pkg/other_pkg

8. Run the cmgetconf command to save the new cluster or package ASCII configuration file whenever the cluster or package configuration is created or modified, as follows:

cmgetconf -c cluster1 cluster1.ascii

cmgetconf -p networker pkg.conf

These ASCII files can be saved in the /etc/cmcluster directory of a node.

Ensure that the NetWorker software backs up /etc/cmcluster so that the configurations can be recovered by using cmapplyconf whenever necessary.

Task 3: Start the Cluster

To start the cluster, enter the following command:

cmruncl

This also starts the NetWorker software.

NetWorker V7.3

Disaster Recovery Guide

177

12

12

Chapter 12: HP-UX Cluster Disaster Recovery

Task 4: Recover the Data

To recover the data:

1. Recover the highly available NetWorker server data as you would a stand-alone NetWorker server. For information about recovering a

NetWorker server, see

"Chapter 3: UNIX Disaster Recovery" on page 21 .

When you shut down the NetWorker software during this step, use

cmhaltpkg

and not nsr_shutdown, as follows:

cmhaltpkg networker exchange -a c /dev/vg03 mount /dev/vg03/lvol1 /vg031 mv /vg031/nsr/res /vg031/nsr/res.old mv /vg031/nsr/res.R /vg031/nsr/res umount /vg031 vgchange -a n /dev/vg03 cmmodpkg -e networker

(restarts NetWorker software)

2. Recover the clients' data from recent backups.

178

NetWorker V7.3

Disaster Recovery Guide