To increase the availability of Orchestrator services, you can configure a cluster of Orchestrator server instances.
Before you begin
Configure the database that you plan to use as a shared database to accept multiple connections, so that it can accept connections from the different Orchestrator instances.
To prevent possible transactional deadlocks when the database is Microsoft SQL Server database, you must set the ALLOW_SNAPSHOT_ISOLATION and READ_COMMITTED_SNAPSHOT database options on.
Install and configure at least two identical Orchestrator server instances.
If you export the configuration of one Orchestrator server instance and import it to another Orchestrator server or if you clone the machine on which the Orchestrator server is running, you must type the credentials for the new Orchestrator server that you want to use to establish the connection to your vCenter Server instance. You can do this on the vCenter Server tab of the Orchestrator configuration interface.
Verify that the Orchestrator instances use the same database.
Synchronize the clocks of the machines on which the Orchestrator server instances are installed.
About this task
An Orchestrator cluster consists of at least two Orchestrator server instances that share one database.
To work properly in the cluster, all Orchestrator server instances must be configured identically with each other and must have the same plug-ins installed. After you set up the Orchestrator cluster, do not change the configurations of its nodes.
- Log in to the Orchestrator configuration interface of the first Orchestrator server as vmware.
- Click Server Availability.
- Select the Cluster mode check box.
If you have configured the Orchestrator server nodes properly, Orchestrator detects the other nodes when you select the check box.
- (Optional) : Provide values for the Cluster mode settings and click Apply changes.
Number of active nodes
The maximum number of active Orchestrator server instances in the cluster.
Active nodes are the Orchestrator server instances that run workflows and respond to client requests. If an active Orchestrator node stops responding, it is replaced by one of the inactive Orchestrator server instances.
The default number of active Orchestrator nodes in a cluster is one.
Heartbeat interval (milliseconds)
The time interval, in milliseconds, between two network heartbeats that an Orchestrator node sends to show that it is running.
The default value is 5000 milliseconds.
Number of failover heartbeats
The number of heartbeats that can be missed before an Orchestrator node is considered failed.
The default value is 12 heartbeats.
The default failover timeout is 1 minute and is equal to the value of the default heartbeat interval multiplied by the number of the default failover heartbeats.
- Log in to the Orchestrator configuration interface of the second Orchestrator server as vmware.
- Repeat 3 and 4.
You have set up an Orchestrator cluster.
What to do next
You can add more Orchestrator cluster nodes.
When you configure Orchestrator to work in cluster mode, you must first start one of the Orchestrator servers and wait until it starts and initializes the database. A cluster node is considered running when on the Server Availability tab, the node appears under Started cluster nodes with a Running status. If you start more than one Orchestrator servers at the same time, concurrency issues occur as all of the started Orchestrator servers try to initialize the database.