In this 2-hour session, Jeremy Eder, D. John Shakshober, and Larry Woodman explore how to configure and tune Red Hat Enterprise Linux for optimal performance while running a variety of common applications.
You’ll learn how to:
- Analyze the performance of heavily loaded systems.
- Tune the systems to maximize performance on bare-metal x86 systems.
- Tune both Linux containers and clouds virtualized with Kernel-based Virtual Machines (KVM).
In the first hour, we look at the Linux virtual memory system and how to tune for non-uniform memory (NUMA). We share tools like "numastat," "tuned," and techniques that are used to identify and resolve performance issues in a number of combined systems and applications, such as database servers, Internet servers, financial applications, bare metal, and Linux containers.
We go on to discuss performance as it relates to disk and network IO. We show you examples that illustrate the latest performance analysis tools and techniques like using "perf", "tuna", and "performance copilot" to identify any bottlenecks that are impacting system and application performance.
- 2015年 6月 25日