Cluster Management
Using the Dispatcher Phoenix Web’s “Manage Cluster” app, admins can set up server clusters, a group of servers working together as a single system to ensure high availability of services.
Important! Only administrators have access to Manage Cluster capabilities.
Each Dispatcher Phoenix server in a cluster can have a different role:
-
Node - A primary Dispatcher Phoenix server that runs all “Started” workflows.
-
Primary - A fully licensed Dispatcher Phoenix server. (Please note that Dispatcher Phoenix servers licensed as “Demo” cannot be a primary server).
-
Failover - A Dispatcher Phoenix server that is licensed as a Failover. When the primary server goes down, the Failover server processes all “Started” workflows. Failover servers ONLY run workflows when the primary server goes down. When setting up a cluster for Failover, all workflows on the primary should be running as a service. When a Failover server is acting as the primary, running workflows can only be stopped or modified by users who log into the Failover server with the same credentials as the primary server.
Note: To ensure that failover is automatic for MFPs, the administrator should have the MFPs registered with a hostname that is either controlled by a proxy with failover (e.g., HAProxy, nginx, etc.) or governed by a DNS server with failover policies.
-
Offload - A Dispatcher Phoenix server that is licensed to operate as an offload server. You can add Workflow Processing (Advanced and Basic) Offload servers to the cluster.
Configuration Requirements
-
If setting up your cluster with multiple primaries, the packages of Dispatcher Phoenix on the primary servers must match. For example, your primary cluster can include 2 servers with licenses of Dispatcher Phoenix Office installed on both of them; however, your cluster cannot include 1 server with Dispatcher Phoenix Office and 1 server with Dispatcher Phoenix Finance.
-
All servers in the cluster must have the same driver(s) installed. You can use the Dispatcher Phoenix client application (under Tools > Options > Database Setup) to control the driver/version. For example, if SQL Server Native Client 11.0 is configured in the Dispatcher Phoenix client, SQL Server Native Client 11.0 must be installed on the other servers in the cluster.
Quick Start
To set up your cluster, you must do the following:
-
Configure each system in the cluster to broadcast to the cluster, using the Workflow Services Manager.
-
Access the Cluster Management app within Dispatcher Phoenix Web.
-
Add a server to the cluster.
Broadcast Set Up
All members of the cluster must be able to communicate with one another. By default, Dispatcher Phoenix servers are set up as “private.” You must use the Workflow Services Manager to enable your servers to be public so that they are accessible for cluster set up.
Do the following:
-
Launch the Workflow Services Manager (Start Program > Workflow Services Manager). See the following illustration for an example:
-
In the Workflow Services Manager window, do the following:
-
Select the Workflow Service tab.
-
Select the Stop Service button to stop Workflow Services first.
-
Choose which servers should be made public by doing one of the following:
-
Choose the User Specified IP/Hostname option from the Broadcast to Cluster drop-down list; then enter the specific server’s IP address in the empty field that is provided.
-
Choose the Local IP address that Dispatcher Phoenix has automatically detected and is listed in the Broadcast to Cluster drop-down list.
-
Choose Cluster (automatic) from the Broadcast to Cluster drop-down list to allow Dispatcher Phoenix to automatically detect the computer’s host name (i.e., this option is applicable if your computer is part of a domain, etc.)
-
-
-
Select the Save Settings button when you are done.
-
Select the Start Service button.
-
Select the Close button to close the Workflow Services Manager.
Access the Manage Cluster App
From the primary system’s Dispatcher Phoenix Web, select the Manage Cluster button or the Tools > Manage Cluster option from the slide out menu. The Manage Clusters page will display with the host/IP address of the Dispatcher Phoenix server running locally, along with a friendly name and its status (online or offline).
Add A Server
To add a new server within the DP Web Manage Cluster app, do the following:
-
Select the Add Server (+) button.
-
On the Add Server pop-up window, enter information about the server in the provided field; then select the Save button. See the following illustration:
You must add a server in the same way that it was broadcasted to the cluster. For example:
- If the server was broadcast with an IP address, the server must be added using that IP address.
- If the server was broadcast with hostname, the server must be added using that hostname.
- If the server was broadcast automatically, the server must be added using that computer’s fully qualified domain name.
The new server will appear in the Servers list with its licensed role listed (Failover, Offload, etc.). See the following illustration:
Edit A Server
To edit an existing server’s friendly name, do the following:
-
Select the server’s Edit Server button.
-
On the Edit Server pop-up window, type in a new friendly name; then select the Save button.
The modified server name will appear in the Servers list. See the following illustration:
Delete A Server
To remove a server, select the Delete server icon; then select the Delete button on the following confirmation message:
Applying Updates in Clustered Environments
When applying updates to a primary and failover server in a cluster, follow these guidelines:
- Apply updates to the primary server. This will turn the server off. The following will occur:
-
After updates complete and the server powers up again, it will act as a Failover.
-
The server that was licensed as Failover will now act as the primary.
- Apply updates to the failover server. This will turn the server off. The following will occur:
-
The server that was originally licensed as a primary will now act as the primary again.
-
After updates complete and the server powers up again, it will act as a Failover again.