This section provides information that you may need to follow in case you are upgrading one of these specific versions (see Checking for new versions section if you want to check which version is installed in your company):
Kiuwan On-Premises version | Specific upgrade process needed | Apache load balancer optional changes |
---|---|---|
2.8.1910.1 | YES | Added health check for Kiuwan front instances Added new error pages Added kiuwanDomain variable to ease configuration Removed unneeded modules |
If you have currently installed Kiuwan On-Premises version 2.8.1910.1, you will need to follow this guide in order to adapt your Kiuwan On-Premises infrastructure to the latest version.
Due to changes in how Redis Cluster is created through the installation process, in order to keep your installation up to date with the latest infrastructure you will need to follow a different approach to perform a full upgrade from this version.
You will need both the kiuwan-cluster version you used to install 2.8.1910.1 and the latest available kiuwan-cluster (that can be downloaded here https://static.kiuwan.com/download/onpremise/kiuwan-cluster.tar.gz). We will refer to this locations:
Note that if you are using AWS elasticache service to run Redis Cluster services you can ignore this section and follow the standard upgrade process.
Make sure there are no running analyses before continuing to the next step. You can check this in the System Administration Console:
If you are using an automation tool to launch Kiuwan analyses, make sure you pause the needed jobs so no more analyses enter your Kiuwan On-Premises queues.
Once there are no more active analyses, you should wait for the pending analyses to be completely freed before stopping the Kiuwan server. This should take up to 30 minutes maximum.
Change the current directory to the old kiuwan-cluster docker scripts and stop all the Kiuwan On-Premises services:
cd [INSTALLER_DIR_OLD]/docker sudo ./stop-kiuwan.sh |
This will stop all Kiuwan frontal, analyzer and schedulers containers currently running.
In order to update the current Kiuwan On-Premises clients and your volumes global configuration file, you should run the update script located in the new kiuwan-cluster:
cd [INSTALLER_DIR]/docker sudo ./update.sh |
Pruning containers is needed to safely upgrade from this version. To do so, run the uninstall.sh script from the new kiuwan-cluster installation directory.
Note that this step will NOT remove any data from your current installation:
cd [INSTALLER_DIR] sudo ./uninstall.sh |
Redis Cluster current data must be cleaned up in order to start a new fresh cluster. Run the following command:
sudo rm -rf [VOLUMES_DIR]/data-local/Redis/data |
Redis Cluster default configuration has been changed in the following versions of Kiuwan On-Premises. If you are running the provided service you will need to update the Redis Cluster nodes locations.
Edit the global configuration file:
sudo vim [VOLUMES_DIR]/config-shared/globalConfig/globalConfig.properties |
You must change properties "redis.cache.nodes" and "redis.store.nodes" to these values:
redis.cache.nodes=172.17.0.1:6379,172.17.0.1:6380,172.17.0.1:6381,172.17.0.1:6382,172.17.0.1:6383,172.17.0.1:6384 redis.store.nodes=172.17.0.1:6379,172.17.0.1:6380,172.17.0.1:6381,172.17.0.1:6382,172.17.0.1:6383,172.17.0.1:6384 |
If you need to update any other configuration property, you can modify now any other configuration option. Changes will be taken into account when Kiuwan services are started in the next steps.
If you made manual changes in the Wildfly instances configuration to support TLS based email servers, you must set the property "kiuwan.mail.secure.layer" to "tls".
|
Running the installation script will recreate the new Redis Cluster and start all needed services:
cd [INSTALL_DIR] sudo ./install.sh |
You can now continue with the standard upgrade process. See Step 8: load balancer manual configuration (optional).