IBM Spectrum Conductor with Spark Basic Configuration and Administration for Linux

The course is designed for the tuition of IBM Spectrum Conductor with Spark v2.2.
The course is designed to give system administrators the knowledge required to implement and maintain a working Conductor with Spark environment. They will gain solid understanding of the product’s architecture and main features as well as workload resource management, cluster setup, configuration, and administration.
The workshops provide valuable hands-on experience with the installation of Conductor with Spark, cluster configuration and administration. The system administrators will also learn helpful hints and tips and develop fundamental troubleshooting skills.

Profil uczestnika

System administrators with basic knowledge of Linux, Spectrum LSF, and Spectrum Symphony, especially EGO.

Agenda

• Introduction to Conductor with Spark
• Architecture
• Installation and configuration
• Cluster management
• Spark workload
• Scheduling and sharing
• Troubleshooting and debugging
• Monitoring
• Additional features
• GPU scheduling
• Conductor with Spark on Docker

Wymagane przygotowanie uczestników

•Basic knowledge of Linux systems administration.
•Basic knowledge about HPC solutions.
•Previous experience in an EGO-based solution like IBM Spectrum Symphony or IBM Spectrum LSF is highly recommended.

Zagadnienia

•Describe the product features and value proposition
•List the product basic concepts
•Describe the high-level architecture or Conductor with Spark
•Explain EGO resource manager
•List the product prerequisites
•Define the required settings and variables
•Perform a step-by-step installation
•Provide tools and tips for troubleshoot a failed or incomplete installation
•Check how to find setup guides for different installation options
•Perform cluster management operations
•Manage and monitor Spark applications
•Describe REST API integration
•Describe how scheduling works in Conductor with Spark
•Schedule batch applications by using the GUI
•Schedule batch applications by using the CLI
•Retrieve useful information (logs)
•Debug and troubleshoot Conductor with Spark
•Explain the performance achieved by Conductor with Spark under benchmark execution
•List additional features of Conductor with Spark and identify the suitable use cases
•Use GPUs in Conductor with Spark environments
•Explain the GPU workload scheduling mechanisms and GPU shuffle service
•Describe how the product integrates with Docker
•Use Docker with Conductor with Spark