The course teaches how to configure and optimize Red Hat Enterprise Linux systems for peak performance, covering versions 5, 6, and 7 beta. Students will learn to analyze system performance, identify bottlenecks, and apply tuning techniques for various applications. The course covers hardware features affecting performance, performance analysis tools, system tuning parameters, and resolving bottlenecks. The teaching method includes lectures, demonstrations, and practical examples. This course is intended for IT professionals, system administrators, and individuals interested in enhancing the performance of Red Hat Enterprise Linux systems.
Performance Analysis and Tuning of Red Hat Enterprise Linux
Overview
Syllabus
Intro
Red Hat Enterprise Linux Performance Evolution
Industry Standard Benchmark Results 4/2014
RHEL Benchmark platform of choice 2014
RHEL Performance Coverage
Red Hat Performance RHEL7 RC1 vs R6.5
Tuned: Storage Performance Boost
Tuned: Network Throughput Boost
Tuned: Network Latency Performance Boost
Tuned: Profile Inheritance
RHEL "tuned" package
RHEL NUMA Scheduler
Automatic NUMA Balancing - balanced
NUMA Nodes and Zones
Tools to display CPU and Memory (NUMA)
RHEL7 Database Performance Improvements Autonuma-Balance
Autonuma-balance and numad aligns process memory and CPU threads within nodes
numastat shows guest memory alignment
Resource Management using cgroups Ability to manage large system resources effectively Control Group (Cgroups) for CPU/Memory/Network Disk
C-group Dynamic resource control
Cgroup default mount points RHELG
Cgroup how-to
Correct bindings
cpu.shares default
Cgroup node migration
Cgroup automation with Libvirt and Docker
RHEL Scheduler Tunables
Finer Grained Scheduler Tuning
Load Balancing
fork() behavior
RHEL Hugepages/ VM Tuning Standard HugePages 2MB
2MB standard Hugepages
1GB Hugepages
Taught by
Red Hat
