Základní info
The Oracle Solaris 10 Performance Management course introduces students to performance tuning principles, monitoring utilities and tuning tools for the Oracle Solaris 10 Operating System (Oracle Solaris 10 OS). The course includes a review of Solaris subsystems and the utilities provided to monitor system efficiency including kstat, sar, vmstat, iostat, netstat, mpstat, nfsstat, ps, prstat, pmap, the proc tools, truss, dtrace, the DTrace Toolkit, cpustat, cputrack, swap, lockstat, and mdb. This revision also includes a section on resource management using zones and pools, as well as content related to Solaris Studio tools. The course format is divided into three major sections: tools, OS theory, and strategy and actions.The Oracle Solaris 10 Performance Management course includes practices to reinforce skills development. The Practices will be of two types: 1. Learning and practice using the monitoring and configuration tools. 2. Problem solving practices where the student will have to determine the cause of a performance problem or figure out a way to measure the load on a subsystem. Learn To: Identify ways to measure or load subsystemsConfigure resource management with zones and poolsUse system monitoring tools including DTraceDetermine the cause of system performance issuesUse system configuration tools
Audience:
Prerequisites:
- Manage pseudo and distributed file systems
- Describe fundamentals of TCP operations
- Employ advanced systems administration skills in a networked Oracle Solaris 10 OS server environment
- Create and manage logical volumes
- Configure a network adapter and use ndd to modify its settings
Objectives:
- Use the Oracle Solaris 10 OS and third-party tools to analyze performance
- View and set tunable parameters for the global or a non-global zone
- Monitor and report on process and thread activity
- Modify CPU scheduling and virtual memory operations
- Describe system caches and system buses
- Tune I/O, and ZFS and UFS File Systems, and network subsystems
- Describe and employ proper testing procedures to measure throughput, utilization, saturation and latency
- Describe performance management fundamentals
Topics:
Introducing Performance Management
- Describe the principles of performance tuning
- Describe the performance tuning process
- Define the terms used to describe performance aspects
- Describe the drill-down analysis strategy
- Describe the performance analysis approach
- Describe the solaris observability tools
- List kstat-based and procfs-based utilities
- List DTrace-based utilities
The kstat Monitoring Tools
- Describe and use sar
- Describe and use vmstat
- Describe and use iostat
- Describe and use mpstat
- Describe and use netstat
- Describe and use nfsstat
- Describe and use kstat
The procfs Monitoring Tools
- Describe and use ps
- Describe and use prstat
- Describe and use truss
- Describe and use apptrace
- Describe and use pmap
- Describe and use pargs
- Describe and use pwait and preap
- Describe and use ptree, psig, pstop, and prun
DTrace
- List the advantages of DTrace
- Describe the DTrace architecture
- List the four parts of a Pprobe description
- Define a probe clause
- Write and execute DTrace commands
- Write some simple DTrace scripts
- Download and use scripts from the DTrace Toolkit
Other Significant Tools
- Describe and interpret the swap command
- Describe and use cpustat
- Describe and use mdb -k to examine the kernel data structures
- Describe and use dtrace to examine kernel environment variables
Processes and Threads
- Define a process and describe the process address space
- Describe the lifetime of a process
- Describe threads
- Describe and use lockstat(1M)
- Describe and display parameters for process management
- Compare the performance of single-threaded and multithreaded processes
Memory
- Describe physical and virtual memory
- Describe the x86 and SPARC MMUs
- Describe the TLB and cache
- Describe MPSS and how to use them
- Describe the cyclic page cache
- Describe paging and swapping
- Monitor memory utilization
- List tunable parameters for paging and swapping
System Buses and Caches
- Define a cache and its characteristics
- Describe the overall memory hierarchy
- Describe the concept of caching and the cost of a cache miss
- Identify cache problems associated with multiple CPUs
- Identify cache problems associated with cache design
- Describe a bus and diagnose problems associated with buses
- Describe the prtdiag utility
- Describe and use the Solaris Studio cputrack utility
CPU Scheduling
- Control and monitor CPUs
- Create and monitor processor sets
- List and describe the different scheduling classes
- Change the scheduling behavior of a process using nice or priocntl
File Systems and Disk I/O
- Explain the basic concepts of UFS
- Describe driver and bus features that enhance performance
- View device properties and configuration
- Describe I/O performance planning
- Explain the basic concepts of ZFS and identify the layers of its architecture
- Create a ZFS storage pool and describe performance-related considerations
- Create a basic ZFS file system, snapshot, and clone
- Describe ZFS tunable parameters
Monitoring Network Performance
- Describe terms used for network analysis
- Describe network utilization
- Understand effects of misconfigured components
- Describe and use ping and traceroute
- Describe and use snoop
- Describe TCP and UDP
- Describe ICMP
- Describe the performance of network hardware
Applying the Performance Analysis Approach
- Describe steps for maintaining system performance
- Describe common utilities for measuring system performance and setting tunable parameters
- Explain the use of resource management and virtualization software to isolate and distribute workloads
- Describe specific types of bottlenecks and methods for reducing them
- List the basic steps for the performance analysis approach
- Use the performance analysis checklist to identify problems
Performance Testing
- Plan to conduct performance testing
- Understand common pitfalls
