Upgrading Your Cluster to the Latest Product Version

Contents

B.1. Upgrading from SLES 10 to SLEHA 11
B.2. Upgrading from SLEHA 11 to SLEHA 11 SP1

If you have an existing cluster based on SUSE® Linux Enterprise Server 10, you can update your cluster to run with the High Availability Extension on SUSE Linux Enterprise Server 11 or 11 SP1.

For migrating from SUSE Linux Enterprise Server 10 to SUSE Linux Enterprise Server 11 or 11 SP1, all cluster nodes must be offline and the cluster must be migrated as a whole—mixed clusters running on SUSE Linux Enterprise Server 10/SUSE Linux Enterprise Server 11 are not supported.

Upgrading from SLES 10 to SLEHA 11

For convenience, SUSE® Linux Enterprise High Availability Extension includes a hb2openais.sh script with which to convert your data while moving from the Heartbeat to the OpenAIS cluster stack. The script parses the configuration stored in /etc/ha.d/ha.cf and generates a new configuration file for the OpenAIS cluster stack. Furthermore, it adjusts the CIB to match the OpenAIS conventions, converts the OCFS2 file system and replaces EVMS with cLVM. Any EVMS2 containers are converted to cLVM2 volumes. For volume groups referenced in existing resources in the CIB, new LVM resources are created.

To successfully migrate your cluster from SUSE Linux Enterprise Server 10 SP3 to SUSE Linux Enterprise Server 11, you need to execute the following steps:

After the conversion has been successfully completed, you can bring the updated cluster online again.

	Reverting after Update
After the update process to SUSE Linux Enterprise Server 11, reverting back to SUSE Linux Enterprise Server 10 is not supported.

Preparation and Backup

Before updating your cluster to the next product version and converting the data accordingly, you need to prepare your current cluster.

Procedure B.1. Preparing your SUSE Linux Enterprise Server 10 SP3 Cluster

Log in to the cluster.
Review the Heartbeat configuration file /etc/ha.d/ha.cf and check that all communication media support multicasting.
Make sure the following files are equal on all nodes: /etc/ha.d/ha.cf and /var/lib/heartbeat/crm/cib.xml.
Take all nodes offline by executing rcheartbeat stop on each node.
In addition to the general system backup recommended before updating to the latest version, back up the following files, as you need them for running the conversion script after the update to SUSE Linux Enterprise Server 11:
- /var/lib/heartbeat/crm/cib.xml
- /var/lib/heartbeat/hostcache
- /etc/ha.d/ha.cf
- /etc/logd.cf
If you have EVMS2 resources, convert non-LVM EVMS2 volumes to compatibility volumes on SUSE Linux Enterprise Server 10. During the conversion process (see Section B.1.3, “Data Conversion”), these are then turned into LVM2 volume groups. After conversion, make sure to mark each volume group as a member of the High Availability cluster with vgchange -c y.

Update/Installation

After preparing the cluster and backing up the files, you can start updating the cluster nodes to the next product version. Instead of running an update, you can also do a fresh installation of SUSE Linux Enterprise 11 on your cluster nodes.

Procedure B.2. Updating to SUSE Linux Enterprise 11

On all cluster nodes, perform an update from SUSE Linux Enterprise Server 10 SP3 to SUSE Linux Enterprise Server 11. For information on how to update your product, refer to the SUSE Linux Enterprise Server 11 Deployment Guide, chapter Updating SUSE Linux Enterprise.
Conversely, you can also freshly install SUSE Linux Enterprise Server 11 on all cluster nodes.
On all cluster nodes, install SUSE Linux Enterprise High Availability Extension 11 as add-on on top of SUSE Linux Enterprise Server. For detailed information, see Section 3.1, “Installing the High Availability Extension”.

Data Conversion

After having installed SUSE Linux Enterprise Server 11 and the High Availability Extension, you can start with the data conversion. The conversion script shipped with the High Availability Extension has been set up with care, but it cannot handle all set-ups in fully automatic mode. It alerts you of the changes it makes, but needs interaction and decisions from your side. You need to know your cluster in detail—it is up to you to verify that the changes are meaningful. The conversion script is located in /usr/lib/heartbeat (or in /usr/lib64/heartbeat, if you are using a 64-bit system).

	Executing Test Runs
To make yourself familiar with the conversion process, we highly recommend that you test the conversion first (without making any changes). You can use the same test directory to do repeated test runs, but you only need to copy the files once.

Procedure B.3. Testing the Conversion

On one of the nodes, create a test directory and copy the backup files to the test directory:

$ mkdir /tmp/hb2openais-testdir
$ cp /etc/ha.d/ha.cf /tmp/hb2openais-testdir
$ cp /var/lib/heartbeat/hostcache /tmp/hb2openais-testdir
$ cp /etc/logd.cf /tmp/hb2openais-testdir
$ sudo cp /var/lib/heartbeat/crm/cib.xml /tmp/hb2openais-testdir

Start the test run with

$ /usr/lib/heartbeat/hb2openais.sh -T /tmp/hb2openais-testdir -U

or with the following command, if you are using a 64-bit system:

$ /usr/lib64/heartbeat/hb2openais.sh -T /tmp/hb2openais-testdir -U

Read and verify the resulting openais.conf and cib-out.xml files:

$ cd  /tmp/hb2openais-testdir 
$ less openais.conf 
$ crm_verify -V -x cib-out.xml

For detailed information about the conversion stages, refer to /usr/share/doc/packages/pacemaker/README.hb2openais in your installed High Availability Extension.

Procedure B.4. Converting the Data

After doing a test run and checking the output, you can now start with the data conversion. You only need to run the conversion on one node. The main cluster configuration (the CIB) is automatically replicated to the other nodes. All other files that need to be replicated are automatically copied by the conversion script.

Make sure that sshd is running on all nodes with access allowed for root in order for the conversion script to successfully copy the files to the other cluster nodes.
Make sure that all ocfs2 filesystems are unmounted.
The High Availability Extension ships with a default OpenAIS configuration file. If you want to prevent the default configuration from being overwritten during the following steps, make a copy of the /etc/ais/openais.conf configuration file.
Start the conversion script as root. If using sudo, specify the privileged user using the -u option:
```
$ /usr/lib/heartbeat/hb2openais.sh -u root
```
Based on the configuration stored in /etc/ha.d/ha.cf, the script will generate a new configuration file for the OpenAIS cluster stack, /etc/ais/openais.conf. It will also analyze the CIB configuration and let you know if your cluster configuration requires changes, due to the change from Heartbeat to OpenAIS. All file processing is done on the node where conversion runs and replicated to the other nodes.
Follow the instructions on the screen.

After the conversion has been finished successfully, start the new cluster stack as described in Section 3.3, “Bringing the Cluster Online”.

After the upgrade process, reverting back to SUSE Linux Enterprise Server 10 is not supported.

For More Information

For more details about the conversion script and the stages of the conversion, refer to /usr/share/doc/packages/pacemaker/README.hb2openais in your installed High Availability Extension.

Upgrading from SLEHA 11 to SLEHA 11 SP1

To successfully migrate an existing cluster from SUSE Linux Enterprise High Availability Extension 11 to 11 SP1 you can do a “rolling upgrade”, meaning upgrading one node after the other. As the main cluster configuration file has changed from /etc/ais/openais.conf to /etc/corosync/corosync.conf with SUSE Linux Enterprise High Availability Extension 11 SP1, a script takes care of the necessary conversions. They are executed automatically when the openais package is updated.

Procedure B.5. Performing a Rolling Upgrade

Updating Software Packages

	Updating Software Packages
If you want to update any software packages on a node that is part of a running cluster, stop the cluster stack on that node before starting the software update. To stop the cluster stack, log in to the node as `root` and enter rcopenais stop. If OpenAIS/Corosync is running during the software update, this can lead to unpredictable results like fencing of active nodes.

If you want to update any software packages on a node that is part of a running cluster, stop the cluster stack on that node before starting the software update. To stop the cluster stack, log in to the node as root and enter rcopenais stop.

If OpenAIS/Corosync is running during the software update, this can lead to unpredictable results like fencing of active nodes.

Log in as root on the node that you want to upgrade and stop OpenAIS:
```
rcopenais stop
```
Check that your system backup is up-to-date and restorable.
Perform an upgrade from SUSE Linux Enterprise Server 11 to SUSE Linux Enterprise Server 11 SP1 and from SUSE Linux Enterprise High Availability Extension 11 to SUSE Linux Enterprise High Availability Extension 11 SP1. For information on how to update your product, refer to the SUSE Linux Enterprise Server 11 SP1 Deployment Guide, chapter Updating SUSE Linux Enterprise.
Restart OpenAIS/Corosync on the upgraded node to make the node rejoin the cluster:
```
rcopenais start
```
Take the next node offline and repeat the procedure for that node.