AIX - Performance Management - Overview

AIX - Performance Management - Overview

The base priority of a thread is 40. The nice value defaults to 20 for foreground processes and 24 for background processes.The CPUs on a system are shared among all of the threads by giving each thread a time slice of one clock tick (10 ms).
Install the bos.perf.tools (base tools), bos.sysmgt.trace, bos.perf.perfstat and perfagent.tools filesets. Commands no longer supported: bf (bigfoot), bfrpt, lockstat, stem, and syscalls.

topas
Reports selected local system statistics.

schedo
Manages CPU scheduler tunable parameters.
schedo -o 15 : changes the time slice of one clock tick to 15 ms.

bindprocessor
Binds or unbinds the kernel threads of a process to a processor.
bindprocessor 22358 2 : binds the threads in process 22358 to processor 2.

perfpmr
A set of tools and instructions for collecting the data needed to analyze a AIX performance problem.

fdpr
A performance tuning utility for improving execution time and real memory utilization of user-level application programs.

iostat
Reports CPU statistics and input/output statistics for the entire system, adapters, tty devices, disks and CD-ROMs.
iostat -s 2 4 : displays four reports at two second intervals starting with the sum of all activities (-s).
iostat -a : generates an adapter throughput report for all of the disk adapters.
lsattr -E -l sys0 -a iostat : displays the current iostat settings.
chdev -l sys0 -a iostat=false : disable the collection of iostat data.

lvmstat
Reports input/output statistics for LPs, LVs and VGs for hot-spot management.
lvmstat -v rootvg -e : enables statistics collection for all the LVs in VG rootvg.
lvmstat -v rootvg : reports statistics for all the LVs in VG rootvg.
lvmstat -v rootvg -C : clears the counter for VG rootvg.
lvmstat -l hd6 : reports statistics for LV hd6 (paging).
See migratelp to move LPs from one PP to another on a different PV.

AIX - Abbreviations and Acronyms

AIX - Abbreviations and Acronyms

APAR ---> Authorized Program Analysis Report
CDE ---> Common Desktop Environment
CHRP ---> Common Hardware Reference Platform
CuDv ---> Customized Devices object class (ODM)
DHCP ---> Dynamic Host Configuration Protocol
ICMP ---> Internet Control Message Protocol
ISA ---> Instrumentation Systems and Automation Society
ISMP ---> InstallShield Multi-Platform
JFS ---> Journaled File System
JFS2 ---> Enhanced Journaled File System
LFT ---> Low Function Terminal
LTG ---> Logical Track Group
LPP ---> Licensed Program Product
LV ---> Logical Volume
LVCB ---> Logical Volume Control Block
MCA ---> Micro Channel Bus Architecture
ML ---> Maintenance Level
MWC ---> Mirror Write Consistency
NIS ---> Network Information Service
NVRAM ---> NonVolatile Random Access Memory
ODM ---> Object Data Manager Database
PB ---> PetaBytes (1 PB is equal to 1024 TB)
PCI ---> Peripheral Component Interconnect
PCMCIA ---> Personal Computer Memory Card International Association
PdDv ---> Predefined Devices object class (ODM)
PV ---> Physical Volume
PVID ---> Physical Volume IDentifier
PP ---> Physical Partition
PTF ---> Program Temporary Fix
RPC ---> Remote Procedure Call
PReP ---> POWERPC Reference Platform
RPM ---> Red Hat Package Manager
SRC ---> System Resource Controler
SWVPD ---> Software Vital Product Data
TB ---> TeraBytes (1 TB is equal to 1024 GB)
UDI ---> Uniform Device Interface
VG ---> Volume Group
VGDA ---> Volume Group Description Area
VGSA ---> Volume Group Status Area
VPD ---> Vital Product Data

AIX - LPAR / DLPAR

AIX - LPAR / DLPAR

We can configure LPAR and DLPAR in P690 Servers,

LPAR (Logical partitions)

In single server we can install multiple operating systems, like AIX 5.1, AIX 5.2, AIX 5.3, Susi Linux, AIX 4.3, and using LPAR concept

The above Os are installing in single server, before that we have to partition the hardware resources, like processors and memory, Harddisk,

Using HMC (Hardware Management Console) we can partition and operate the LPAR servers

If you have one P690 server with 16 CPU’s and 16GB Ram, in this server we can make 16 partitions,

Now each partitions has 1 CPU and 1GB Ram,

If one particular Os assigned for 500 users that time we have to assign additional hardware resources to that partition, online we can assign the hardware resources using HMC. No need to reboot the server, this concept called DLPAR (dynamic Logical Partition)

Like this we can create the LPAR / DLPAR

HMC - it is a desktop pc (Linux OS) connected with P690 server through Communication (com port) cable or RJ45 Ethernet cat cable

Using HMC we can Create, Delete, and Change the Logical Partitions

This LPAR server has hypervisor Controller, all the information’s stored in this hypervisor, and it is used to maintain the server

Accessing methods:

1. wsm (Web based system management , using this concept we can access the LPAR servers from windows desktops)
2. ssh (Secured shell, using this concept we can access the LPAR servers through telnet)

Cloud Computing an overview.........

Cloud computing is a computing paradigm in which tasks are assigned to a combination of connections, software and services accessed over a network. This network of servers and connections is collectively known as "the cloud." Computing at the scale of the cloud allows users to access supercomputer-level power. Users can access resources as they need them. (For this reason, cloud computing has also been described as "on-demand computing.")

This vast processing power is made possible though distributed, large-scale cluster computing, often in concert with server virtualization software, like Xen, and parallel processing. Cloud computing can be contrasted with the traditional desktop computing model, where the resources of a single desktop computer are used to complete tasks, and an expansion of the client/server model. To paraphrase Sun Microsystems' famous adage, in cloud computing the network becomes the supercomputer.

Cloud computing is often used to sort through enormous amounts of data. In fact, Google has an initial edge in cloud computing precisely because of its need to produce instant, accurate results for millions of incoming search inquries every day, parsing through the terabytes of Internet data cached on its servers. Google's approach has been to design and manufacture hundreds of thousands of its own servers from commodity components, connecting relatively inexpensive processors in parallel to create an immensely powerful, scalable system. Google Apps, Maps and Gmail are all based in the cloud. Other companies have already created Web-based operating systems that collect online applications into Flash-based graphic user interfaces (GUIs), often using a look and feel intentionally quite similar to Windows. Hundreds of organizations are already offering free Web services in the cloud.

In many ways, however, cloud computing is simply a buzzword used to repackage grid computing and utility computing, both of which have existed for decades. Like grid computing, cloud computing requires the use of software that can divide and distribute components of a program to thousands of computers. New advances in processors, virtualization technology, disk storage, broadband Internet access and fast, inexpensive servers have all combined to make cloud computing a compelling paradigm. Cloud computing allows users and companies to pay for and use the services and storage that they need, when they need them and, as wireless broadband connection options grow, where they need them. Customers can be billed based upon server utlilization, processing power used or bandwidth consumed. As a result, cloud computing has the potential to upend the software industry entirely, as applications are purchased, licensed and run over the network instead of a user's desktop. This shift will put data centers and their administrators at the center of the distributed network, as processing power, electricity, bandwidth and storage are all managed remotely.

AIX - Performance Monitoring - I/O Bound

AIX - Performance Monitoring - I/O Bound

Using iostat command we can monitor the I/O bound

In this I/O bound we can monitor the disk usage, If you are using two or three heavy used file system in one PV that time that particular PV usage will be very high

AIX - Performance Monitoring - Network Bound

AIX - Performance Monitoring - Network Bound


Using topas command we can monitor the network bound, in this network monitor we can find KB in and KB out process.

AIX - Performance Monitoring - Memory Bound

AIX - Performance Monitoring - Memory Bound

Using vmstart command we can monitor the memory bound

#vmstart 2 3

After this command execution, it will show the memory page in and page out process. Within 2 seconds, 3 outputs is displayed
If you are checking paging space usage using command lsps

#lsps –s

11% - this is normal
40% - Server running on slow performance
70% - users cannot logon the server,
Some error message will appear (fork exist failed)

AIX - Performance Monitoring - CPU Bound

AIX - Performance Monitoring - CPU Bound

Using Topas command we can check the server performance monitor, the following things we can monitor.

1. processor (CPU Usage)
2. Network (KB in / KB out)
3. Memory (Page in / Page out)
4. I/O (HDD read and write)
5. Top 8 process

AIX - Performance Monitoring

AIX - Performance Monitoring

In performance monitoring the below operations can be performed.

1. Cpu bound
2. Memory bound
3. Network bound
4. I/O bound

AIX - Rewind, Eject, Erase the tape

AIX - Rewind, Eject, Erase the tape

Using tctl command we can rewind, eject and erase the tape drive.

#tctl –f /dev/rmt0 rewind (To rewind the tape)

#tctl –f /dev/rmt0 eject (To eject the tape)

#tctl –f /dev/rmt0 erase (To erase the tape content)