Introduction to CHESS

CHESS (Clustertech HPC Environment Software Stack) V4.2 Enhanced Version is a cluster man- agement software solution developed by Clustertech. It transforms a stack of servers to a coherent HPC cluster, providing a unified interface for the resources of the cluster to be deployed, managed, monitored, scheduled and reported. It can greatly enhance the efficiency of the cluster and simplify its management. CHESS V4.2 consists of a number of modules: the CUI (Cluster User Interface), the cluster management module, the cluster monitor module, the cluster deployment module, and the cluster report module. CUI is the basic module of Web Portal, while other modules can be selected freely according to the needs of the users. CHESS helps users to deloy the system. It also helps users to install and test cluster software, application environment and applications. It provides a suite of software and services up to cluster applications on top of the hardware equipment. Its main functions are as follows.

CHESS Web Portal

CHESS Web Portal is the user interface of CHESS. By utilizing CUI (Clustertech User Interface, the unified log-on platform for ClusterTech), it unifies the interfaces of the cluster management, cluster monitor, job scheduling and management, and cluster report. It allows all these modules to be logged on in a unified way. It is also responsible for user management and permission management, allowing administrators to modify the access right of each module for users.

Cluster Management

Provide on-site and remote cluster management, including node management, (storage) share management, image management, log management, parallel command and process management. It can be used to remotely power up / down machine in batches, to log into the nodes via SSH and VNC, etc.

Jobs scheduler

The CHESS resource management and job scheduling system manage the software and hardware resources in the system reasonably and efficiently. By adjusting the job scheduling policy via the job scheduling web interface, system administors can optimize resource utilization and reduce response time of the jobs. The system administrator can easily monitor the CPU utilization of each node, and configure the resource manager to optimize the cluster system. This makes complex resource management and job scheduling tasks simpler, unified and efficient. The main function of the module includes job, resources manager configuration, queue configuration, node configuration, resource reservation and application template management.

Cluster Monitor

CHESS provides rich monitoring information. Through the web interface, system administrators can inspect and understand the usage of the cluster, the topology of the cluster, the filesystem of the cluster, the alert system and the performance of nodes.

Cluster Report

The reporting system collects, analyses and summarizes data of the system resources of the cluster as well as jobs of the users. It provides users with information about the utilization of cluster resoures, storage system, as well as the usage of users and queues. With the charging rate configured by the administrator, it can also analyzing the collected data to generate bills according to usage.