CycleCloud supports a number of schedulers and applications, and ships with HTCondor and Grid Engine. For this demonstration, we will be using a template pre-configured for use with HTCondor. To do this, click Clusters in the menu bar. As there are no existing clusters defined, you will be prompted to add a new cluster.
Select a Configuration
Click the HTCondor logo from the list of cluster templates. Cluster templates are pre-configured infrastructure definitions that encapsulate the composition and basic attributes of compute environments. These templates can be easily modified by system administrators and end users to create standard deployments. CycleCloud also supports numerous pre-configured application environments commonly used in industries such as life sciences, manufacturing, finance, and more.
The cluster templates described in the last section include various parameters that control the compute cluster’s infrastructure, scaling behaviour, and provisioning options. The user interface for modifying cluster parameters is data-driven and can be easily customized to adapt to specific customer requirements. The configuration for the HTCondor cluster type is depicted below. Duplicate the cluster parameter settings and then click “save” to create the configuration. Your AWS Keypair name will match the name of the keypair imported into AWS earlier.
|Cluster Name||The name of the cluster|
|Credentials||Uses credentials created during the “Add Provider Account” step|
|Keypair||The public key name imported into your Cloud Service Provider earlier|
|Keypair Path||The location on the machine running CycleCloud where your private key is stored|
|Manager OS||The operating system used for the scheduler. For this guide, select CentOS 6.|
|Execute OS||The operating system used for the execute nodes. For this guide, use CentOS 6.|
|Cluster-init||The folder in the account bucket associated with your cluster credentials. The process of initialization is covered in the next section of this guide. For this demo, enter “pi” as your folder.|
|CM Type||The machine type to use for your scheduler. For this guide, select m3.large, 2 cores.|
|Execute Type||The machine type to use for the execute nodes. For this guide, select m3.large.|
|Autoscale||Select this checkbox to allow your cluster to dynamically scale up or down with the workload|
|Initial Cores||The number of execute cores to start when the cluster is started. When set to 0, it instructs CycleCloud not to start execute nodes on cluster start.|
|Max Cores||The maximum number of execute cores to autostart|
|Use Spot Instances||When checked, the execute nodes are spot instances. When unchecked, the execute nodes are on demand instances.|
|Networking||Options for Amazon’s Virtual Private Cluster service. For this guide, leave blank.|