How To-Set up High Availability and Clustering in Titan

 How To: Set up High Availability and Clustering in Titan

Titan supports high availability (HA) and clustering by letting multiple nodes share configuration, user accounts, and session state. This keeps the service running if a node goes down and spreads the load across nodes. A cluster needs a shared database and shared storage.

Before you begin

1. Create the database. Titan supports SQL Server, PostgreSQL, MariaDB, and MySQL.
2. Make sure the Titan High Availability and Scaling module is enabled in your registration code. Clustering requires it.
3. Create a shared folder that every node can reach by UNC path (for example \\fileserver\TitanCluster). Every node must be able to read, write, and delete files in it. You will use it for two directories:
   • User Data Directory: holds user files so every node serves the same data.
   • Server Database Directory: lets the nodes tell each other about configuration changes (see the note below).

Set up the primary cluster node

1. In Titan Administrator, click Add new Server.
2. Select New Stand Alone or Primary Cluster Server and click Next.
3. Set up the database connection:
   • Select MS SQL Server as the database (or your chosen database type).
   • Select the server instance.
   • Select Username and Password as the authentication type and enter the username and password.
   • Select the database you created in Before you begin, step 1.
4. Click Next, fill in the remaining settings, and click Finish.

Add the UNC account

1. Go to Domain > Server > UNC Accounts tab.
2. Click Add.
3. Enter the UNC path, UNC user, and UNC password.
4. Click Next.

Point the shared directories at the UNC path

1. Go to Domain > Server > Files/Directories > Directories tab.
2. Set User Data Directory to a folder on the UNC share, for example \\fileserver\TitanCluster\UserData.
3. Set Server Database Directory to a folder on the UNC share, for example \\fileserver\TitanCluster\Database.
4. Change any other directories you need, then click Apply.
5. Restart the Titan service so it starts watching the new Server Database Directory.

Important: the Server Database Directory must be on the UNC share.

When you make a change on one node (for example, add a user, change a group, or reconfigure the server), that node writes it to the shared database and drops a small change-notification file into the Server Database Directory. The other nodes watch that folder. When a notification file appears, they reload the changed data from the database right away.

If each node uses its own local Server Database Directory (the default), the other nodes never see these notifications. They keep using cached settings and won't show the change until they're restarted.

All nodes in the cluster must use the same UNC folder for the Server Database Directory.

Note: Use a UNC path for the User Data Directory too, so the primary and secondary nodes serve the same user files.

If you use Titan DMZ

1. Go to Domain > Server > Cluster tab.
2. Enter the cluster node name and click Add DMZ.
3. Click Action on DMZ1.
4. Check Enable DMZ Service.
5. Set the DMZ IP address.
6. Enable the protocols you need and set their ports.
7. Click Save, then click Apply.

Set up a secondary cluster node

1. On the secondary machine, in Titan Administrator, go to Domain > Server and click Add new Server.
2. Select New member server of existing server and click Next.
3. Set up the database connection:
   • Select MS SQL Server as the database (or the same database type as the primary).
   • Select the server instance.
   • Select Username and Password as the authentication type and enter the username and password.
   • Select the same database the primary node uses.
   • Set the server name. It must match the primary server's name.
4. Click Next, set the IP address, and click Finish.
5. Go to Domain > Server > Files/Directories > Directories tab and confirm that Server Database Directory and User Data Directory show the same UNC paths as the primary node. Correct them if they don't, click Apply, and restart the Titan service.
6. Make sure the secondary node can read, write, and delete files in both UNC folders.

Verify the cluster

1. Open Titan Administrator on the primary node and make a small change, such as adding a test user.
2. Open Titan Administrator on the secondary node and refresh. The change should appear within a few seconds, without a restart.
3. If it doesn't appear, check that both nodes use the same UNC path for the Server Database Directory, that both can write to it, and that the Titan service was restarted after the directory was changed.