About Us

RSInfoMinds, a web based IT Training and Consultancy firm. It is established with high dreams in training people in IT Infrastructure Field. We provide Online and Class Room training in various fields of IT Infrastructure Management.

Join Us: http://www.facebook.com/RSInfoMinds
Mail Us: rsinfominds@gmail.com
Twitter: @RSInfoMinds

We are specialized in the below courses:

Redhat Linux Admin Redhat Linux Cluster
Redhat Virutualization IBM AIX Admin
IBM AIX Virtualization IBM AIX Cluster
HP Unix Admin HP Unix Cluster
HP Unix Virtualization Shell Scripting
Veritas Volume Manager Veritas Cluster
Oracle Core DBA VMWare


We provide training in such a way, So that you get in depth knowledge on the Courses you look for.

And we ensure you are very confident from each and every Techincal aspect that the IT Industry needs and expects from you.

We also conduct Workshops on the latest technology and the real time faculties sharing their work experiences to make you the best.

Thursday, 8 December 2011

CPU Virtualization - Power 5 Based





On a Power 5 system it is possible to create a logical parition with dedicated processor or shared processor. The concept of shared processore refers to virtualization of CPU's processing unit.

Power5 introduced the concept of "Global Shared Processor Pool". It contains group of processor that will be shared among the logical partition that are created using the shared resources (CPU).

Processing unit of a CPU is always measured in 1/10 th of the CPU Power.  So, with the single CPU 1/10=0.1 Processing unit can be assigned to a logical partition. So a single CPU can support 10 LPAR's (logically).

The granulartiy of increasing the processing unit is 0.01 ( As per IBM Standards).

The above diagram shows a P5 machine with 6 Physical Processors.
The machine has 4 partitions namely A,B,C and D.

Partitions A and B are created with the dedicated processors. 3 Processor per partition. So, A and B are allocated with 6 Physical Processor.

Remaining processors available : 16-(3+3)=10 Processors.

Processing unit of a single processor is expressed as 1/10, therefore 10 Processors hold : 10*10 : 100 Processing units.

The 10 Processors are combined in a "Global Shared Processor Pool" with the processing capacity of 100 units.

Logical Partitions use these processing units as a Virtual Processor. i.e., Virtual processors provide required amount of processing units to the logical partitions.

Now, Partition C has 52 Processing units with 7 virtual processor and Partition D has 23 Processing units with 5 Virtual Processor.

So, Total processing units occupied by C and D : 52+23=75 Processing units.

Now, available processing unit in the Global Shared Processor Pool = 100-75 = 25 Processing units.

These 25 Processing units are not allocated.

No.of Processing unit allocated to a partition is called : Entitlement.

Capped Partition: Partition cannot get more processing unit than that of allocated to it.

Uncapped Partition: Partition can acquire more processing unit than that of allocated to it.

Similarily, an "uncapped partition" can also consume less processing units than that of allocated to it. This should give a clear understanding between "Physical consumption" and "Entitlement consumption".

Physical Consumption: Refer to percentage of processor power/unit consumed by a partition.

Entitlement Consumption: Refer to percentage of processor power/unit currently consumed compared to the no.of processing units allocated to that partition.

So, its evident that an "uncapped partition" can have an "Entitlement Consumption" of more than 100%.




Explore # topas -C

In this report, you see the global memory consumption metrics:

mem : Represents the sum of monitored memory for all shared partitions in the pool.
muse : Represents the sum of memory consumed for all shared partitions in the pool.
Mem : Represents the amount of memory allocated to the partition.
InU : Represents the amount of memory consumed by the partition.
mon : Represents the total amount of memory allocated to the partitions.

Restoring an AIX System with mksysb

In the event of a disaster, you can boot from tape, CD/DVD, or the network and do a full restore. The processes are very similar, so we'll cover them as one procedure:
  1. If the system is still running, you can set it up to boot from tape or CD before you shut it down:
    1. To boot from CD/DVD:
    2. Client # bootlist m normal cd0
                              
    3. To boot from tape:
    4. Client # bootlist m normal rmt0
                              
    5. To boot from the network:
    6. Client # bootlist m normal en0
                              
  2. Power down the system.
  3. Attach a system terminal (or monitor and keyboard).
  4. Prepare the media:
    1. If this is a tape restore, attach a tape drive and insert the tape in the drive.
    2. If this is a CD restore, insert the first CD/DVD in the drive.
  5. Power up the system.
  6. Display the SMS menu. If you weren't able to set the boot list to the proper device before powering down, you need to tell the system to boot from the proper device on power up. On HMC-attached systems, you can select the boot mode before you start the partitions; it then defaults to the SMS menu. On other AIX systems, you need to interrupt the boot process to display the SMS menu. On some older systems, you press F1 while the POST icons are appearing on the monitor. On most later model systems, you press F5 while the POST icons are appearing on the monitor. Consult the hardware guide that shipped with your AIX system for the interrupt sequence for your model of hardware.
  7. Once in the SMS menu, select Maintenance Mode, and set one of the following restore/boot options:
    1. Choose Select Boot Device, and specify CD, DVD, or tape restore.
    2. For a NIM install, select "Installation from Network." Selecting NIM install may require you to supply additional network parameters depending on how your NIM server is set up.
  8. The backup is checked for compatibility. If the backup is compatible with the hardware it is being restored onto, it pauses for five seconds and then continues unprompted and restores the rootvg based on the settings in image.data and/or bosinst.data. If you want to interrupt the install, press the 0 key three times (000) during the five-second pause. This forces the restore into prompted mode. In the event that the backup does not match the hardware of the system, it automatically begins a prompted restore. The prompted restore looks very much like the initial AIX installer, but at the end you have a restore of your system.
  9. If you are not performing a prompted restore, the restore continues uninterrupted until it completes. If you are performing a prompted restore, it allows you to change the physical disk(s) that the restore is performed to, shrink filesystems, etc.
  10. When the restore finishes, the system changes its boot device to the install target devices for the rootvg and reboot.