Configuring a Cluster to Run the Windows Agent
The Agent for Windows can run in a Windows cluster environment. A cluster
environment is defined as multiple machines working together as one system. The cluster
environment provides a level of redundancy so that if one of the machines in the cluster
fails, another machine is available to replace the failed component.
The following instructions describe how to configure a two node cluster environment to
run the Windows agent offered by Scheduler.
Prerequisites
Before installing the Tidal Agent for Windows on the nodes of a cluster, you must first
complete and/or verify the following on each node:
Verify that the systems on each node are identical
Verify that the agent machines in each node meet the hardware and software
requirements specified in the Installing the Agent for Windows chapter of the
Installation and Configuration Guide for Scheduler.
Verify that the user installing the Windows agent has the specified user rights
including access to the registry on each machine.
Verify that the cluster group has the following resouce types:
Network name
IP address
Physical disk
Configuring the Agents for a Cluster
During configuration, you should complete a step on a machine and then go around to the
other machines in the cluster and do the same step. When that step has been performed on
each machine in the cluster, return to the first machine and do the next step and then
again do that step on the other machines in the cluster, and return to the first machine and
do the next step, etc.
An agent instance must exist on every node in the cluster before it can be configured to
run as a cluster. This means that if you add a third agent instance to a machine, before you
configure that instance, go to all of the other machines in the cluster and add a third
instance.
To configure the agents:
1
2
Install the Agent for Windows on the first cluster node.
Verify that the cluster works correctly.
Check that the cluster software is installed and configured correctly by forcing a
failure on a server. Be sure that a failover to another server occurs as intended and
that control can be returned to the server that failed. Your Windows Cluster
Administrator should help you with this.
Be sure to install the agent to a non
‐
clustered physical disk on the local machine
using the default directory path during installation.
3
Stop the agent if it is running.
4
An agent instance must exist on every node in the cluster before it can be configured
to run as a cluster. If you are adding an agent instance, add the agent instance to a
machine. See
“Installing the Agent for Windows”
. Go to each of the other nodes in
the cluster and add that same instance.
Once each of the nodes on the cluster have the same agent instance on it, you can edit
the agent instance to configure it for the cluster.
5
From the Start menu button, select
Programs>TIDAL Software>Agent>Instance
Manager
to display the Agent Instance Manager.
Figure 25
Agent Instance Manager
6
Select the first agent instance and click the Edit
button to display the Agent Instance
Manager’s configuration screen.
Figure 26
Adding an Agent Instance
Caution
You must install the agent on the same disk drive letter on each cluster node.
For example, if you install the agent on the C drive of one node, the agent must
be installed on the C drive of the other nodes also.
56
If this agent instance is on a node that is configured for a cluster with an existing
agent service, the
Run on a cluster group
option is available. If the node is not part
of a cluster than this option is unavailable. You cannot proceed any further without
verifying with your Windows Cluster Administrator that the node is correctly
configured as a member of the cluster.
7
8
9
10
11
Save
button.
12
13
When each agent instance on each node is configured properly, start each clustered
agent instance from its active node using Scheduler’s Service Control Manager.
Starting the agent instance in the Service Control Manager automatically starts the
agent resource in the Windows Cluster Administrator.
Go to each node and repeat this procedure for each agent instance.
In the Cluster Nodes field, select which node this agent instance is on. When the
fields are completed, click the
In the Work Directory field, enter the pathname to the work directory that was
created for the cluster group.
In the Physical Disk field, select the disk that the agent instance resides on. All the
disks that were created on all of the cluster groups are listed. Be sure to select a disk
that exists on the cluster group you selected.
In the Cluster Group field, select which cluster group that this agent instance
belongs to.
Select the Run on a cluster group option to expand the screen to display the cluster
configuration fields.
