Creating a new Cluster
Before creating a new cluster or joining a node to an existing cluster, keep in mind the following restrictions:
- It's strongly suggested to configure all Application Suite nodes to send email notifications
- Ensure that all Application Suite nodes clocks are up-to-date and synchronized
- All nodes in the same cluster must have the same Application Suite version. If they have not, upgrade them to the latest Application Suite version before creating the cluster
- It's highly recommended that all nodes use the same SQL server version
- At the beginning, all nodes in the cluster will get the configuration and data from the first node. A node that joins an existing cluster will completely loose both its configuration and data. So, choose carefully the first node that will be used to create a cluster.
- Joining a node to a cluster requires to stop all applications on that node. Once added, a server restart of all other nodes is required, too.
- TCP port 52700 will be allocated to the communication between the nodes
- A set of TCP ports above 52700 is used for data synchronization between the nodes (the number of ports depends on the number of nodes)
Creating a new cluster
Before creating an Application Suite cluster you have to choose the node that will be used as the first source of data (i.e. the node that contains the data that will be shared in the cluster). Notice that this is a critical decision, since all the data in the other nodes will be lost. Also, before continuing, ensure that the node is already configured to use the desired database server (you can check this through the Support web page).
Once you've chosen the node, log in as administrator in that node's Application Suite web interface and click on Admin -> High Availability.
You will be asked to choose whether you want to create a new cluster or to add this node to an existing cluster.
On the first node choose New Cluster and click Next.
In the next page you will have to choose one of the IP addresses of the machine. This will be the address used by the other nodes of the cluster to communicate with this node.
You will also have to specify if every node will use its own database instance (Replicated Database) and the cluster will need to synchronise data between all of them, or if the cluster will use a single database instance (Shared Database).
If you choose to use multiple database instances, it's highly recommended that all of them have the same SQL Server version.
Once you've made your choices, click Next.
If you confirm your choices, the cluster creation procedure will begin and it may take a long time. While the cluster is being prepared, all the applications will keep running as usual.
As soon as the operation completes, if the cluster has been created successfully, the local node will send a success notification. If the creation fails, it will send a failure notification.
From now on, every time you'll open this web page, you'll se the list of all the nodes in the cluster.
Of course, when the cluster is prepared it contains just the local node. Other nodes will appear when they'll be added.
Adding a node to an existing cluster
Warning: when you add a node to a cluster you will lose all the data that is on that node.
Before you try to add a node to a cluster, ensure that:
- The new node has the same Application Suite version of the other nodes in the cluster, otherwise you won't be able to complete the procedure
- The node is already configured to use the desired database server (you can check this through Support web page)
- If the cluster is using the Replicated Database strategy, there's enough free disk space on the machine to contain at least twice the size of the database of a node that already belongs to the cluster.
You should notice also that during this procedure all the applications of the Suite will be stopped and unavailable on this node.
To add a node to a cluster, log in as administrator in that node's Application Suite and click on Admin -> High Availability.
Choose Add to existing Cluster and click Next.
On the next page you will have to choose one of the IP addresses of the machine. This will be the address used by the other nodes of the cluster to communicate with this node.
Then you must enter the IP address of an existing node (chosen from the displayed nodes list) of the cluster you want to join. That node must remain up and running until the end of this procedure.
When you click Next and confirm your choices, the node addition procedure will start. It may take a long time, depending on the amount of data contained at that moment in the cluster (this procedure will automatically create a backup of source node's database, download and restore it to the node to be added). As a coarse idea, the first node that joins an existing cluster could take 16 minutes per GB of database size. Nodes that join later, could take 3 minutes per GB.
When the procedure is complete, the node will send a success notification, but if the addition fails it will send a failure notification.
From now on, this page will show the list of all the nodes in the cluster (including the node that you just added).
Now, in order to propagate the topology change to all the other nodes, you should restart them one by one.
Pls. keep in mind that above cluster's NODE ID corresponds to Imagicle server's host name. You can later change host name, if required; just beware that NODE ID will remain unchanged (with initial host name).
There are some actions an administrator won't be able to do on a cluster:
- On Application Suite versions lower than Summer 2015, you can't remove a node from a cluster.
- Once a node is part of a cluster, you can't change its IP Address.
- You can't change cluster's NODE IDs. To change NODE ID, please contact Imagicle Support.
- You can't restore a IAS backup to a cluster node: data integrity is granted by the cluster itself.
For any problem regarding one of the activities above, you'll need to contact Imagicle technical support.
This article was:
Thank you for your feedback!
|Monitoring a Cluster||Product Integration API|