Hosts page in the AIX PowerVM view

The Hosts page in the AIX PowerVM view provides an overview of all available hosts in the AIX virtualization environment belonging to a selected domain, including details about CPU, pool and memory utilization, memory usage trends, unused resources and more. It also displays data collected from HMC-only terminals and supports LPAR instrumented with Capacity Agent data collection.

Scenario

Teresa is an AIX Technology Specialist in an IT organization. She manages and monitors the AIX infrastructure to ensure that there are sufficient resources (CPU, memory, and storage) to meet the current and future capacity requirements.

Teressa wants to add new partitions to the existing infrastructure. She would like to get the following information:

  • Know the residual capacity of the hosts to be able to determine which host can accommodate them.
  • Check the health of the hosts in the AIX infrastructure to make sure the resources are efficiently utilized by the partitions.

Alan, an administrator, installs the AIX views and provides the required permissions to Teresa to access the view. With AIX views, Teresa can quickly get the information that she wants. She can use the Hosts page in the AIX PowerVM view to get an overview of all the available hosts in the AIX virtualization environment. To know the residual capacity of the hosts, she can review the metrics in the Allocated Capacity page. This page provides the unassigned cores and memory information for a host.

By reviewing and analyzing the metrics related to unused resources and trend of resource utilization on the Used Capacity page, Teressa can evaluate the health of the hosts in the infrastructure. 

To access the page, select Views > Virtualization > AIX Views > PowerVM.

The page shows the following details:

Allocated capacity 

The Allocated Capacity table lists all metrics that are available through HMC-only and fully instrumented collection modes. By default, some columns are hidden. You can show or hide columns using the action menuthat is located next to the Allocated Capacity table title. For more information, see Managing charts and tables in Views.

Each row corresponds to a PowerVM Frame and shows the following information:

Column

Description

Statistical
Type 
Metrics 
HostName of the host.Last value NAME
CPU CoresThe total number of physical CPU cores in the frame.Last value

(1) CPU_NUM_HMC

(2) CPU_NUM

Unassigned CoresNumber of cores that are not assigned to any PowerVM Partition on the host.Last value

CPU_CORES_UNASSIGNED

CPU Configured for Partitions [Cores]Number of cores assigned to all PowerVM partitions on the host.Last value

CPU_NUM - CPU_CORES_UNASSIGNED

CPU Allocated [cores]

The amount of CPU that is allocated to the partitions by the hypervisor, expressed in cores.

95th percentile(1) CPU_ALLOC_CORE_FRACTION_HMC

(2) CPU_ALLOC_CORE_FRACTION

CPU Allocated [%]The percentage of CPU that is allocated to the partitions by the hypervisor.95th percentile

(1) CPU_UTIL_HMC

(2) CPU_ALLOC_CORE_FRACTION_HMC/CPU_NUM_HMC

(3) CPU_ALLOC_CORE_FRACTION/CPU_NUM

CPU Allocated Trend [cores]Daily growth rate of CPU Allocated of the AIX frame, measured in cores/day.TrendTrend value computed based on CPU_UTIL_HMC

Pool - Entitlement Configured

The sum of entitlement configured for the partitions that belong to the frame.

Last value

(1) CPU_ENTC_HMC

(2) CPU_ENTC

Pool - Entitlement Allocated Util [%]

The CPU utilization of the pool normalized by its total entitlement. The total entitlement of the pool is the sum of the entitlement of the SPLPARs in the pool.Computed metric - based on 95th Percentile

 (1) CPU_ALLOC_CORE_FRACTION_HMC/CPU_ENTC_HMC

(2) CPU_ALLOC_CORE_FRACTION/CPU_ENTC

Total Memory [GB]

Total physical memory available for the host.Last value TOTAL_REAL_MEM

Unassigned Memory [GB]

Memory that is not assigned to any PowerVM partition on the host.Last value MEM_UNASSIGNED

Firmware Memory [GB]

Memory consumed to manage PowerVM virtualization by host.Last value MEM_OVERHEAD

Memory Configured for Instrumented Partitions [GB]

Memory configured for all instrumented PowerVM partitions on the host.Last value

SUM (TOTAL_REAL_MEM of partitions with

LPAR_INSTRUMENTED = true)

 

Configured Partitions

Number of partitions that are configured on the PowerVM host, both ACTIVE and IN_ACTIVE.Last valueGM_NUM

Instrumented Partitions

Number of partitions on which Agents are installed for a PowerVM host.Last valueCount of partitions with LPAR_INSTRUMENTED = true
Memory Configured for PartitionsMemory configured for all the partitions in the frameLast valueTOTAL_REAL_MEM - (MEM_UNASSIGNED + MEM_OVERHEAD)
HMC Access

This metric indicates, with a True or False value, if HMC access is available for at least one partition in the PowerVM frame. HMC access is required for collecting certain key configuration metrics. If you cannot view certain metrics, check if HMC access is enabled.

Last value

HMC_CONNECTION

Used capacity

The Used Capacity table lists all PowerVM hosts in the selected domain. By default, some columns are hidden. You can show or hide columns using the action menuthat is located next to the Allocated Capacity table title. For more information, see Managing charts and tables in Views.

Each row corresponds to a PowerVM host and shows the following information.

Note

Metrics marked with (plus) are displayed only in the LPAR instrumented mode. They are not displayed when HMC Only is enabled. You can configure this option from the Settings page.

Column

Description

Statistical
Type 
 Metrics

Host

Name of the host. An alias for the frame, which is available only if HMC access is available for at least one of the partitions in the frame. If HMC access is not available, Host Hrw ID is displayed here.

Last value ALIAS_NAME
Host Hrw IDHardware ID of the host. This ID is unique for each host.Last value

NAME

 

CPU [CORES]

Number of CPU cores in the host.

Last valueCPU_NUM

Unused Cores

Number of unused CPU cores.

Computed metric - based on 95th PercentileCPU_NUM - CPU_UTIL * CPU_NUM 

CPU Util [%]

Percentage of CPU used by the host.

95th PercentileCPU_UTIL 

CPU Trend [%]

The growth rate of the CPU Utilization of the frame.

TrendTrend based on CPU_UTIL
Pool - Entitlement ConfiguredSum of entitlement configured for the partitions in the poolLast valueCPU_ENTC

Pool - Entitlement Utilization [%]

The CPU utilization of the pool normalized by the total entitlement of the pool. The total entitlement of the pool is the sum of the entitlement of the SPLPARs in the pool.

Computed metric - based on 95th Percentile

(CPU_ENTC_CONS / CPU_ENTC) * 100.0

 

Pool - Unused Entitlement

Unused pool entitlement expressed in processor units.

Computed metric - based on 95th PercentileCPU_ENTC - CPU_ENTC_CONS

Memory [GB]

Total physical memory available for the host. This metric is reported only when one of the instrumented partitions has HMC access.

Last valueTOTAL_REAL_MEM 

(plus)Unused Memory [GB]

Unused memory available to the host.

Computed metric - based on 95th PercentileTOTAL_REAL_MEM - MEM_CONSUMED

(plus)Memory Utilization [%]

Percentage of memory used by the host normalized by the total memory of the host.

95th PercentileMEM_UTIL 

(plus)Memory Used Daily Trend [GB]

The growth rate of the memory utilization of the frame.

TrendTrend based on MEM_CONSUMED
 (plus)Network Rate [Mbit/s]

Network received/transmitted rate, measured in megabits per second, aggregated from the partitions in the frame.

95th PercentileNET_BIT_RATE

(plus)I/O Rate [MB/s]

Disk input/output rate, measured in megabytes per second, aggregated from the partitions in the frame.

95th PercentileDISK_TRANSFER_RATE

Total Spec Rating

The total SPEC benchmark performance rating of the host, based on the CPU model and number of CPU cores.

Last valueBYBENCHMARK_VALUE[2006]

Unused Spec Rating

Unused SPEC rating of the host, based on CPU utilization and the total SPEC rating.

Computed metric - based on 95th PercentileComputed using CPU_UTIL and BYBENCHMARK_VALUE[2006]

Configured Partitions

Number of configured partitions in the PowerVM host.

Last valueGM_NUM

Instrumented Partitions

Number of partitions on which agents are installed for a PowerVM host.Last valueCount of partitions with LPAR_INSTRUMENTED = true

Detailed information about a PowerVM host

Click the name of a PowerVM host in the PowerVM Hosts Used and Available Capacity view table. The following host details are displayed on the Host Details page:

  • General details: General host details, such as name, type, description and number of elements within it are displayed at the top of the page. 

  • Tags panel: Displays tags, if any, that are associated with the cluster. For more information, see Filtering data displayed in Views.

  • Related Information: The Related Information panel at the top of the page contains the following links:

  • Summary information: The Summary tab, that is displayed by default, displays details for the selected cluster in a horizontal summary table, as well as charts.

    For more information, see Data Explorer views for PowerVM Hosts.

  • Configuration Details: The Configuration Details tab displays the configuration metrics for the cluster.

Disk Controllers in VIO servers 

NPIV is a Fibre Channel SAN feature from POWER6, VIOS 2.1 FP 20.1. It allows VMs to access SAN devices using N-Port ID Virtualization, where each LPAR gets its own world-wide port name (WWPN).

NPIV is configured on partition controllers of VIO Servers (VIOS), and WWPNs are assigned to client LPARs, but the physical I/O still passes through VIOS. NPIV assists you with configuring LUN mapping for arrays, and zoning to each LPAR.

Through NPIV, BMC Helix Continuous Optimization counts I/O more accurately for LPARs and VIOSs.

A VIOS is equipped to virtualize disks, Logical Volumes, and files, and each of these is presented as a SCSI disk to the client partition. BMC Helix Continuous Optimization reports reports I/O numbers for each of the following entities:

  • Instrumented partition (for example, Client Partition A)
  • Instrumented VIO Server
  • Virtual or physical adapter
For NPIV storage configurations, VIOS presents a VFC Server Adapter, which is a Fibre Channel pass-through.
The Client partition has a VFC Client Adapter and sees the LUN directly. This enables BMC Helix Continuous Optimization to gather and show the FC adapter configuration and I/O metrics for each instrumented partition, including the VIOS.

Click the WWPN link to view the details page that displays detailed charts for NPIV. These charts are useful to monitor I/O traffic of fibre channel adapters in the VIOS. The Disk Controller in VIO Servers table provides the following information. 

Important

The table is displayed only when a VIO Server is configured with Fiber Channel adapters (NPIV).

Column

Description

Statistical
Type 
Metrics 

World Wide Port Number

A World Wide Name (WWN) is a unique identifier used in storage technologies like Fiber Channel or Serial Attached SCSI (SAS). Used on storage area networks, it performs a function equivalent to the MAC address in Ethernet protocol.Last value
PartitionName of the partition on which the disk controller residesLast valueNAME
Interface Name

The name of the interface or the host bus adapter.

Last valueBYHBA_INTERFACE_NAME 
Port Type

The type of the interface, such as, f-port (fabric), n-port (node), nl-port (node loop), and so on.

Last valueBYHBA_TYPE 
Port Speed Supported [GBits/s]The highest supported port speed, measured in Gigabits per secondLast valueBYHBA_BANDWIDTH_RUNNING 
Port Speed Running [GBits/s]The current running port speed, measured in Gigabits per secondLast valueBYHBA_BANDWIDTH_SUPPORTED

Shared Ethernet Adapters in VIO servers

Shared Ethernet Adapters (SEA) enable LPARs to connect to external physical Ethernet switches using broadcast and multicast protocols like ARP, DHCP, BOOTP. You can also count network I/O for the virtual adapters, and Used, Available and Unbalanced I/O information is also available.

The Shared Ethernet Adapters in VIO Servers table displays the following information:

Column

Description

Statistical
Type 
Metrics 

Adapter Name

Name of the SEA adapter. Clicking on the adapter name displays the Viewing the Shared Ethernet Adapter – All Layers details page details page.Last valueBYSEAIF[SUBOBJNAME] 
Device TypeThe SEA device type.Last valueBYSEAIF_DEVICE_TYPE
Byte In Rate [MBytes/sec]The number of Bytes per second received during the interval through the network adapter specified in the subobject.#DRAFTmeet: AverageBYSEAIF_IN_BYTE_RATE 
Byte Out Rate [MBytes/sec]The number of Bytes per second sent during the interval through the network adapter specified in the subobject.AverageBYSEAIF_OUT_BYTE_RATE  
Packet In Rate [Packets/sec]The number of successful packets per second received during the interval through the network adapter specified in the subobject.AverageBYSEAIF_IN_PKT_RATE  
Packet Out Rate [Packets/sec]The number of successful packets per second sent during the interval through the network adapter specified in the subobject.AverageBYSEAIF_OUT_PKT_RATE   
Error In Rate [Errors/sec]Rate of input errors (errors/sec),per second by network adapterAverageBYSEAIF_IN_ERROR_RATE  
Error Out Rate [Errors/sec]Rate of output errors (errors/sec),per second by network adapterAverageBYSEAIF_OUT_ERROR_RATE  

Shared Ethernet Adapter – All Layers details page

SEA Adapter summary and the detailed charts on this page help to analyze the network traffic of the VIOS (all network traffic from the AIX host is routed through the VIOS) and detect potential bottlenecks that can affect the performance of all other LPARs in the frame. SEAs contain a hierarchy of adapter layers.

To view the Shared Ethernet Adapter – All Layers details page and detailed charts, click the name of the adapter in the Adapter Name column. The table displays the following information:

Column

Description

Statistical
Type 
Metrics 

Adapter Name

Name of the SEA adapterLast valueBYSEAIF[SUBOBJNAME] 
PartitionName of the partition the SEA adapter is configured onLast valueNAME 
HostName of the host the SEA belongs toLast valueHost name retrieved using relationship VH_CONTAINS_GM 
Device RoleThe current role of the SEA adapter: Virtual, Real, Primary, Backup or Control Channel Adapter.Last valueBYSEAIF_DEVICE_ROLE
Parent NameName of the adapter's parent SEA adapter.Last valueBYSEAIF_PARENT_NETNM
Device TypeThe SEA device type.Last valueBYSEAIF_DEVICE_TYPE 
Byte In Rate [MBytes/sec]The number of Bytes per second received during the interval through the network adapter specified in the subobject.AverageBYSEAIF_IN_BYTE_RATE  
Byte Out Rate [MBytes/sec]The number of Bytes per second sent during the interval through the network adapter specified in the subobject.AverageBYSEAIF_OUT_BYTE_RATE 
Packet In Rate [Packets/sec]The number of successful packets per second received during the interval through the network adapter specified in the subobject.AverageBYSEAIF_IN_PKT_RATE  
Packet Out Rate [Packets/sec]The number of successful packets per second sent during the interval through the network adapter specified in the subobject.AverageBYSEAIF_OUT_PKT_RATE    
Error In Rate [Errors/sec]Rate of input errors (errors/sec),per second by network adapterAverageBYSEAIF_IN_ERROR_RATE   
Error Out Rate [Errors/sec]Rate of output errors (errors/sec),per second by network adapterAverageBYSEAIF_OUT_ERROR_RATE   

The Shared Ethernet Adapter – All Layers details page also displays the following charts:

  • Byte In/Out Rate (Hourly): A chart that graphically depicts the number of Bytes per second received during the interval through the network adapter specified in the subobject.
  • Packet In Rate (Hourly): A chart that graphically depicts the number of successful packets per second received during the interval through the network adapter specified in the subobject.
  • Error In Rate (Hourly): A chart that graphically depicts the rate of input errors (errors/sec),per second by network adapter.

Was this page helpful? Yes No Submitting... Thank you

Comments