+49 (911) 80 10 30

Give us a call or have us call you back!

Simply leave a message with your callback appointment

Your Name *

Your phone number *

Callback date *

Callback time *

Your request *

In order to process your request, it is necessary to process the data you provide. This means that by submitting the contact form, you consent to your data being processed (see Privacy Policy). You can object to the processing of your data at any time.

* All fields marked with an asterisk are mandatory fields.
Send a message

Your Name *

Your email *

Your phone number

Your request *

In order to process your request, it is necessary to process the data you provide. This means that by submitting the contact form, you consent to your data being processed (see Privacy Policy). You can object to the processing of your data at any time.*

* All fields marked with an asterisk are mandatory fields.
Social Media

Share page

LI250: Linux Performance Analysis & Tuning

Training: Linux/Unix

System planners and administrators with advanced Linux knowledge learn to identify, analyze, and sustainably resolve performance bottlenecks. The course covers kernel and hardware aspects, CPU and memory management, mass storage, file systems, network performance, as well as benchmarking and monitoring with tools such as collectd and RRD.

Appointment selection

Appointments in German language


Unfortunately there are currently no available appointments. Would you like to request an appointment? Then click on 'No matching appointment?'

Possible course languages: German

Request prefered appointment period:

Your Name *

Your email *

Your phone number

Preferred period from *

Preferred period to *

Number of participants *

Your message

* All fields marked with an asterisk are mandatory fields.

Unfortunately there are currently no available appointments.
Would you like to request an appointment? Then click on 'No matching appointment?'

Request prefered appointment period:

Your Name *

Your email *

Your phone number

Preferred period from *

Preferred period to *

Number of participants *

Your message

* All fields marked with an asterisk are mandatory fields.

Agenda:

Analyzing and ultimately improving the performance of existing Linux systems should be an interesting topic for all administrators. This training covers the following topics in detail:

Linux kernel
- Overview of performance-relevant functions
- Current performance-relevant developments
- Overview /proc and /sys filesystem
- Control groups

Hardware aspects
- Architecture of a modern computer system
- Different bus systems
- Different storage systems
- Software and hardware RAID
- IOPS
- Special features of flash memory

Processor
- The process scheduler
- Identifying CPU-bound workloads
- Metrics at system and process level
- Scheduling policies and CPU affinity
- Processes in detail: states and tracing
- Frequency scaling

Main Memory
- Basics of Linux memory management
- Identifying main memory-bound workloads
- Metrics at system and process level
- Swapping behavior and out-of-memory killer
- Shared memory
- Limits for 32-bit and 64-bit Linux
- How much memory does a process need?
- Excursus: Detecting memory leaks
- Excursus: Limiting memory consumption
- Excursus: Compressing main memory

Mass Storage and Filesystems
- How processes generate I/O
- The Linux I/O stack,
- Identifying I/O-bound workloads
- Metrics at system, filesystem and process level
- Recommended filesystem sizes
- I/O scheduling, multipathing
- Special features of SAN systems and flash memory
- Different file systems such as ext2-4 and XFS
- Outlook: BTRFS
- Performance and data integrity
- Performance-relevant mkfs and mount options
- Filesystem alignment
- SSD handling

Network
- Send and receive buffers
- Congestion control
- Analysis of TCP connections
- Netfilter/Conntrack parameters
- Bufferbloat
- Statistics and netflows

Methods of performance analysis
- Some anti-methods
- Systematic methods
- The USE method

Graphing and monitoring
- Basic concepts
- Real-time monitoring
- RRD basics
- Data collection with collectd

Benchmarks
- Correct measurement and typical measurement errors
- Different types of benchmarks
- Test scenarios with common benchmarks
- Determining load limits
- Evaluating results
- Excursus: Flexible I/O Tester

Applications
- NFS and FS cache
- Apache
- MySQL and PostgreSQL
- Rsync
- Outlook: Load balancing via IPVS

Objectives:

The course LI250 Linux Performance Analysis & Tuning provides knowledge to identify and classify performance bottlenecks in operations, conduct performance tests, and evaluate Linux systems based on their hardware and software. Practical tips for optimal sizing of new systems, performance-oriented setup of application and service infrastructures and tuning options at the system level complement the course.

Target audience:

The seminar LI250 Linux Performance Analysis & Tuning is ideal for:

System architects
System administrators

Prerequisites:

To be able to follow the course content and learning pace of the workshop LI250 Linux Performance Analysis & Tuning effectively, you should have the following prerequisites:

Advanced Linux knowledge
I/O and network fundamentals
Basic hardware knowledge
Basic knowledge of a scripting language is advantageous

Description:

The practice-oriented course LI250 Linux Performance Analysis & Tuning targets system planners and system administrators with advanced Linux knowledge, basic knowledge in I/O, networking and hardware as well as ideally experience in a scripting language. The objective of the training is to identify performance bottlenecks in Linux systems, classify them and resolve them through targeted tests and analyses.

The participants receive comprehensive know-how for evaluating hardware and software, learn methods for sizing new systems as well as tuning and optimization options at system level. Focus areas are kernel and hardware aspects, CPU and memory management, mass storage and file systems, network performance as well as methods for performance analysis and benchmarking. Insights into practice-relevant applications and monitoring tools such as collectd and RRD complete the course. After completion, participants are able to sustainably increase the performance of existing systems and optimally design new infrastructure.