Topics: HMC, Logical Partitioning

Useful HMC commands

Here are some very useful commands for the Hardware Management Console (HMC): Show vital product data, such as the serial number:

# lshmc -v
Show the release of the HMC:
# lshmc -V
Show network information of the HMC:
# lshmc -n
Reboot the HMC:
# hmcshutdown -r -t now
Show the connected managed systems:
# lssysconn -r all
Change the password of user hscpe:
# chhmcusr -u hscpe -t passwd -v abc1234
List the users of the HMC:
# lshmcusr
These are intersting log files of the HMC:
# ls -al /var/hsc/log/hmclogger.log
# ls -al /var/hsc/log/cimserver.log
Monitor the disk space:
# monhmc -r disk
This can be used to view the file systems of the HMC. Try using "proc", "mem" and "swap as well. By default this command will loop forever and update the screen every 4 seconds. You can run it only once, with the following command:
# monhmc -r disk -n 0
Zero out log files:
# chhmcfs -o f -d 0
This will delete any temporary files. Extremely useful if the HMC calls home to IBM about high usage of one of its file systems. Open a virtual console from the HMC:
# vtmenu
Exit by typing "~." (tilde dot) or "~~." (tilde tilde dot). Force the closure of a virtual terminal session:
# rmvterm -m SYSTEM-9117-570-SN10XXXXX -p name
Change the state of a partition:
# chsysstate -m SYSTEM-9131-52A-SN10XXXXX -r lpar -o on -n name
-f default_profile
# chsysstate -m SYSTEM-9131-52A-SN10XXXXX -r lpar -o shutdown 
-n name --immed
To start all partitions of one managed server:
# chsysstate -m Prd2-Server-8233-E8B-SN0XXXXXX -r lpar -o on --all
List partition profiles for a managed system:
# lssyscfg -r prof -m SYSTEM-9117-570-SN10XXXXX
List partition information:
# lspartition

Topics: AIX, Installation, Logical Partitioning, Virtualization

Workload Partitioning (WPAR) in AIX 6.1

The most popular innovation of IBM AIX Version 6.1 is clearly workload partitioning (WPARs). Once you get past the marketing hype, you'll need to determine the value that WPARs can provide in your environment. What can WPARs do that Logical Partitions (LPARs) could not? How and when should you use WPARs? Equally as important, when should you not use Workload Partitioning. Finally, how do you create, configure, and administer workload partitions?

For a very good introduction to WPARs, please refer to the following article: https://www.ibm.com/developerworks/aix/library/au-wpar61aix/ or download the PDF version here.

This article describes the differences between system and application WPARs, the various commands available, such as mkwpar, lswpar, startwpar and clogin. It also describes how to create and manage file systems and users, and it discusses the WPAR manager. It ends with an excellent list of references for further reading.

Topics: Logical Partitioning, Virtual I/O Server, Virtualization

Introduction to VIO

Prior to the introduction of POWER5 systems, it was only possible to create as many separate logical partitions (LPARs) on an IBM system as there were physical processors. Given that the largest IBM eServer pSeries POWER4 server, the p690, had 32 processors, 32 partitions were the most anyone could create. A customer could order a system with enough physical disks and network adapter cards, so that each LPAR would have enough disks to contain operating systems and enough network cards to allow users to communicate with each partition.

The Advanced POWER Virtualization feature of POWER5 platforms, makes it possible to allocate fractions of a physical CPU to a POWER5 LPAR. Using virtual CPU's and virtual I/O, a user can create many more LPARs on a p5 system than there are CPU's or I/O slots. The Advanced POWER Virtualization feature accounts for this by allowing users to create shared network adapters and virtual SCSI disks. Customers can use these virtual resources to provide disk space and network adapters for each LPAR they create on their POWER5 system.

There are three components of the Advanced POWER Virtualization feature: Micro-Partitioning, shared Ethernet adapters, and virtual SCSI. In addition, AIX 5L Version 5.3 allows users to define virtual Ethernet adapters permitting inter-LPAR communication.


An element of the IBM POWER Virtualization feature called Micro-Partitioning can divide a single processor into many different processors. In POWER4 systems, each physical processor is dedicated to an LPAR. This concept of dedicated processors is still present in POWER5 systems, but so is the concept of shared processors. A POWER5 system administrator can use the Hardware Management Console (HMC) to place processors in a shared processor pool. Using the HMC, the administrator can assign fractions of a CPU to individual partitions. If one LPAR is defined to use processors in the shared processor pool, when those CPUs are idle, the POWER Hypervisor makes them available to other partitions. This ensures that these processing resources are not wasted. Also, the ability to assign fractions of a CPU to a partition means it is possible to partition POWER5 servers into many different partitions. Allocation of physical processor and memory resources on POWER5 systems is managed by a system firmware component called the POWER Hypervisor.

Virtual Networking

Virtual networking on POWER5 hardware consists of two main capabilities. One capability is provided by a software IEEE 802.1q (VLAN) switch that is implemented in the Hypervisor on POWER5 hardware. Users can use the HMC to add Virtual Ethernet adapters to their partition definitions. Once these are added and the partitions booted, the new adapters can be configured just like real physical adapters, and the partitions can communicate with each other without having to connect cables between the LPARs. Users can separate traffic from different VLANs by assigning different VLAN IDs to each virtual Ethernet adapter. Each AIX 5.3 partition can support up to 256 Virtual Ethernet adapters.

In addition, a part of the Advanced POWER virtualization virtual networking feature allows users to share physical adapters between logical partitions. These shared adapters, called Shared Ethernet Adapters (SEAs), are managed by a Virtual I/O Server partition which maps physical adapters under its control to virtual adapters. It is possible to map many physical Ethernet adapters to a single virtual Ethernet adapter, thereby eliminating a single physical adapter as a point of failure in the architecture.

There are a few things users of virtual networking need to consider before implementing it. First, virtual networking ultimately uses more CPU cycles on the POWER5 machine than when physical adapters are assigned to a partition. Users should consider assigning a physical adapter directly to a partition when heavy network traffic is predicted over a certain adapter. Secondly, users may want to take advantage of larger MTU sizes that virtual Ethernet allows, if they know that their applications will benefit from the reduced fragmentation and better performance that larger MTU sizes offer. The MTU size limit for SEA is smaller than Virtual Ethernet adapters, so users will have to carefully choose an MTU size so that packets are sent to external networks with minimum fragmentation.

Virtual SCSI

The Advanced POWER Virtualization feature called virtual SCSI allows access to physical disk devices which are assigned to the Virtual I/O Server (VIOS). The system administrator uses VIOS logical volume manager commands to assign disks to volume groups. The administrator creates logical volumes in the Virtual I/O Server volume groups. Either these logical volumes or the physical disks themselves may ultimately appear as physical disks (hdisks) to the Virtual I/O Server's client partitions, once they are associated with virtual SCSI host adapters. While the Virtual I/O Server software is packaged as an additional software bundle that a user purchases separately from the AIX 5.3 distribution, the virtual I/O client software is a part of the AIX 5.3 base installation media, so an administrator does not need to install any additional filesets on a Virtual SCSI client partition.

Topics: AIX, Hardware, Logical Partitioning

Release adapter after DLPAR

An adapter that has previously been added to a LPAR and now needs to be removed, usually doesn't want to be removed from the LPAR, because it is in use by the LPAR. Here's how you find and remove the involved devices on the LPAR:

First, run:

# lsslot -c pci
This will find the adapter involved.

Then, find the parent device of a slot, by running:
# lsdev -Cl [adapter] -F parent
(Fill in the correct adapter, e.g. fcs0).

Now, remove the parent device and all its children:
# rmdev -Rl [parentdevice] -d
For example:
# rmdev -Rl pci8 -d
Now you should be able to remove the adapter via the HMC from the LPAR.

If you need to replace the adapter because it is broken and needs to be replaced, then you need to power down the PCI slot in which the adapter is placed:

After issuing the "rmdev" command, run diag and go into "Task Selection", "Hot Plug Task", "PCI Hot Plug Manager", "Replace/Remove a PCI Hot Plug Adapter". Select the adapter and choose "remove".

After the adapter has been replaced (usually by an IBM technician), run cfgmgr again to make the adapter known to the LPAR.

Topics: HMC, Logical Partitioning, Virtualization

Stopping and starting a LPAR from the HMC prompt

It is possible to stop and start an LPAR from the HMC prompt:

# lssycfg -r lpar
This command will list all partitions known to this HMC.
# chsysstate -o osshutdown -r lpar -n [partition name]
This command will send a shutdown OS command to the lpar.
# chsysstate -o on -r lpar -n [partition name]
This command will activate the partition.
# lsrefcode -r lpar -F lpar_name,refcode
This command will show the LED code.

Topics: AIX, HMC, Logical Partitioning

Lpar tips

The uname -Ls command will show you the partition number and the partition (lpar) name. When setting the resource allocation for a partition profile, set the minimum to the absolute bare minimum, and set the maximum as high as possible.

For memory there are special considerations:

  • If you set the maximum too low and you wish to exceed above the maximum amount of memory defined in the active profile, you can't simply adjust the profile and put extra memory in via DLPAR, because the LPAR has been initialized with a certain page table size, based on the maximum amount of memory setting. Therefore, a reboot will be required when you wish to use more memory than defined in the active profile. If you do try it however, you'll receive the following error:
    HMCERRV3DLPAR018: There is no memory available for dynamic logical partioning on this partition.
  • If you set the maximum too high, the partition will be initialize with a large page table size, which uses too much memory for overhead, which you might never use.

Number of results found for topic Logical Partitioning: 6.
Displaying results: 1 - 6.