In many cases, it is desirable to upgrade a virtual appliance by deploying a virtual appliance at the newer version and importing the old configuration.  For example, the size of the Traffic Manager disk image was increased in version 9.7, and deploying a new virtual appliance lets a customer take advantage of this larger disk.  This article documents the procedure for deploying a new virtual appliance with the old configuration in common scenarios.   These instructions describe how to upgrade and reinstall Traffic Manager appliance instances (either in a cluster or standalone appliances). For instructions on upgrading on other platforms, please refer to Upgrading Traffic Manager.   Upgrading a standalone Virtual Appliance   This process will replace a standalone virtual appliance with another virtual appliance with the same configuration (including migrating network configuration). Note that the Traffic Manager Cloud Getting Started Guide contains instructions for upgrading a standalone EC2 instance from version 9.7 onwards; if upgrading from a version prior to 9.7 and using the Web Application Firewall these instructions must be followed to correctly back up and restore any firewall configuration.   Make a backup of the traffic manager configuration (See section "System > Backups" in the Traffic Manager User Manual), and export it. If you are upgrading from a  version prior to 9.7 and are using the Web Application Firewall, back up the Web Application Firewall configuration - Log on to a command line - Run /opt/zeus/stop-zeus - Copy /opt/zeus/zeusafm/current/var/lib/config.db off the appliance. Shut down the original appliance. Deploy a new appliance with the same network interfaces as the original. If you backed up the application firewall configuration earlier, restore it here onto the new appliance, before you restore the traffic manager configuration: - Copy the config.db file to /opt/zeus/stingrayafm/current/var/lib/config.db    (overwriting the original) - Check that the owner on the config.db file is root, and the mode is 0644. Import and restore the traffic manager configuration via the UI. If you have application firewall errors Use the Diagnose page to automatically fix any configuration errors Reset the Traffic Manager software.   Upgrading a cluster of Virtual Appliances (except Amazon EC2)   This process will replace the appliances in the cluster, one at a time, maintaining the same IP addresses. As the cluster will be reduced by one at points in the upgrade process, you should ensure that this is carried out at a time when the cluster is otherwise healthy, and of the n appliances in the cluster, the load can be handled by (n-1) appliances.   Before beginning the process, ensure that any cluster errors have been resolved. Nominate the appliance which will be the last to be upgraded (call it the final appliance).  When any of the other machines needs to be removed from the cluster, it should be done using the UI on this appliance, and when a hostname and port are required to join the cluster, this appliance's hostname should be used. If you are using the Web Application Firewall first ensure that vWAF on the final appliance in the cluster is upgraded to the most recent version, using the vWAF updater. Choose an appliance to be upgraded, and remove the machine from the cluster: - If it is not the final appliance (nominated in step 2),    this should be done via the UI on the final appliance - If it is the final appliance, the UI on any other machine may be used. Make a backup of the traffic manager configuration (System > Backups) on the appliance being upgraded, and export the backup.  This backup only contains the machine specific info for that appliance (networking config etc). Shut down the appliance, and deploy a new appliance at the new version.  When deploying, it needs to be given the identical hostname to the machine it's replacing. Log on to the admin UI of the new appliance, and import and restore the backup from step 5. If you are using the Web Application Firewall, accessing the Application Firewall tab in the UI will fail and there will be an error on the Diagnose page and an 'Update Configuration' button. Click the Update Configuration button once, then wait for the error to clear.  The configuration is now correct, but the admin server still needs to be restarted to pick up the configuration: # $ZEUSHOME/admin/rc restart Now, upgrade the application firewall on the new appliance to the latest version. Join into the cluster: For all appliances except the final appliance, you must not select any of the auto-detected existing clusters.  Instead manually specify the hostname and port of the final appliance. If you are using Web Application Firewall, there may be an issue where the config on the new machine hasn't synced the vWAF config from the old machine, and clicking the 'Update Application Firewall Cluster Status' button on the Diagnose page doesn't fix the problem. If this happens, firstly get the clusterPwd from the final appliance: # grep clusterPwd /opt/zeus/zxtm/conf/zeusafm.conf clusterPwd = <your cluster pwd> On the new appliance, edit /opt/zeus/zxtm/conf/zeusafm.conf (with e.g. nano or vi), and replace the clusterPwd with the final appliance's clusterPwd. The moment that file is saved, vWAF should get restarted, and the config should get synced to the new machine correctly. When you are upgrading the final appliance, you should select the auto-detected existing cluster entry, which should now list all the other cluster peers. Once a cluster contains multiple versions, configuration changes must not be made until the upgrade has been completed, and 'Cluster conflict' errors are expected until the end of the process. Repeat steps 4-9 until all appliances have been upgraded.   Upgrading a cluster of STM EC2 appliances   Because EC2 licenses are not tied to the IP address, it is recommended that new EC2 instances are deployed into a cluster before removing old instances.  This ensures that the capacity of the cluster is not reduced during the upgrade process.  This process is documented in the "Creating a Traffic Manager Instances on Amazon EC2" chapter in the Traffic Manager Cloud Getting Started Guide.  The clusterPwd may also need to be fixed as above. ... View more