Hitachi

JP1 Version 12 JP1/Performance Management - Remote Monitor for Platform Description, User's Guide and Reference


System Summary (PI)

Organization of this page

Function

The System Summary (PI) record stores performance data, taken at a specific interval, about the processors and memory in the entire system of the monitored host.

Note

  • If the connection to the monitored host fails, the information cannot be collected.

  • If the OS of both the PFM - RM host and the monitored host is 64-bit Windows, you can collect a maximum of 32 pieces of CPU information.

  • There is no upper limit to the number of instances you can create. The maximum number of monitoring targets you can create per instance is 50.

  • If the OS of both the monitored hosts is HP-UX, AIX, or Linux, the memory size allocated as the file cache is treated as the size of used physical memory area in the following fields:

    • Free Mem % (FREE_MEM_PERCENT)

    • Free Mem Mbytes (FREE_MEM_MBYTES)

    • Used Mem % (USED_MEM_PERCENT)

    • Used Mem Mbytes (USED_MEM_MBYTES)

  • If the OS of the monitored host is AIX and a user other than the root user is collecting information, 0 is displayed in the following fields if that user does not belong to either the adm group or system group:

    • CPU % (CPU_PERCENT)

    • Idle % (IDLE_PERCENT)

    • System % (SYSTEM_PERCENT)

    • User % (USER_PERCENT)

    • Wait % (WAIT_PERCENT)

    • Free Swap % (FREE_SWAP_PERCENT)

    • Free Swap Mbytes (FREE_SWAP_MBYTES)

    • Used Swap % (USED_SWAP_PERCENT)

    • Used Swap Mbytes (USED_SWAP_MBYTES)

    • Total Swap Mbytes (TOTAL_SWAP_MBYTES)

    • Page Fault Counts/sec (PAGE_FAULT_COUNTS_PER_SEC)

  • If the OS of the monitored host is the Workload Partition (WPAR) environment of AIX V6.1 or later, 0 is displayed in the following fields:

    • CPU % (CPU_PERCENT)

    • Idle % (IDLE_PERCENT)

    • System % (SYSTEM_PERCENT)

    • User % (USER_PERCENT)

    • Wait % (WAIT_PERCENT)

    • Free Mem % (FREE_MEM_PERCENT)

    • Free Mem Mbytes (FREE_MEM_MBYTES)

    • Used Mem % (USED_MEM_PERCENT)

    • Used Mem Mbytes (USED_MEM_MBYTES)

    • Total Mem Mbytes (TOTAL_MEM_MBYTES)

    • Free Swap % (FREE_SWAP_PERCENT)

    • Free Swap Mbytes (FREE_SWAP_MBYTES)

    • Used Swap % (USED_SWAP_PERCENT)

    • Used Swap Mbytes (USED_SWAP_MBYTES)

    • Total Swap Mbytes (TOTAL_SWAP_MBYTES)

    • Page Fault Counts/sec (PAGE_FAULT_COUNTS_PER_SEC)

  • If the OS of the monitored host is the non-global zone environment of Solaris, 0 is displayed in the following fields:

    • Effective Free Mem % (EFFECTIVE_FREE_MEM_PERCENT)

    • Effective Free Mem Mbytes (EFFECTIVE_FREE_MEM_MBYTES)

    • Effective Used Mem % (EFFECTIVE_USED_MEM_PERCENT)

    • Effective Used Mem Mbytes (EFFECTIVE_USED_MEM_MBYTES)

  • Information is not collected if TargetType is set to icmp in the monitoring target settings.

  • If you want to monitor the amount of memory actually available, we recommend that you monitor the Effective Free Mem Mbytes field, not the Free Mem Mbytes field. (Excluding the non-global zone environment of Solaris)

    Similarly, for the Free Mem %, Used Mem Mbytes, and Used Mem % fields, we recommend that you monitor the Effective Free Mem %, Effective Used Mem Mbytes, and Effective Used Mem % fields, which all begin with Effective, respectively.

    For details about the difference between the fields, see 2.2.2(1) Overview of memory monitoring.

  • When a record of the PI record type is displayed in a real-time report, the first time collection data of the field where the delta field is Yes is invalid.

  • When monitoring a guest OS in Hyper-V with dynamic memory enabled, the value of the Total Mem Mbytes (physical memory) field in PI records might not reflect actual memory usage. The same applies to the following fields whose value is derived from the Total Mem Mbytes field of a PI record.

    • Free Mem %

    • Used Mem %

    • Used Mem Mbytes

    • Effective Free Mem %

    • Effective Used Mem %

    • Effective Used Mem Mbytes

    This is because the guest OS is unable to recognize the reduction in available physical memory when the ballooning feature of Hyper-V recovers the excess physical memory allocated to the virtual machine. (When ballooning takes place, there might be a discrepancy between the physical memory allocated to the virtual machine and the physical memory recognized by the guest OS).

    This issue does not occur when the physical memory allocated to a virtual machine increases, because the guest OS recognizes such an increase. Note that because PFM - RM for Platform acquires information from the guest OS, the value of the Total Mem Mbytes field in PI records reflects the amount of physical memory recognized by the guest OS.

Default and changeable values

Item

Default value

Changeable

Collection Interval

300

Y

Collection Offset

0

Y

Log

Yes

Y

LOGIF

(Blank)

Y

Over 10 Sec Collection Time

No

N

Realtime Report Data Collection Mode

Reschedule

Y

Legend:

Y: Changeable

N: Not changeable

ODBC key field

None

Lifetime

None

Record size

Fields

PFM - View name

(PFM - Manager name)

Description

Smry rule

Grpg rule

Format

Delta

Not sprtd on

Record Type (INPUT_RECORD_TYPE)

The record name. This is always PI.

COPY

COPY

char (8)

No

--

Record Time (RECORD_TIME)

Time when the record was created.

COPY

COPY

time_t

No

--

Interval (INTERVAL)

Interval during which the information is collected.

[Units: seconds]

If the data is summarized in historical reports, the last value stored is displayed.

COPY

FIXED

ulong

No

--

VA DeviceID (VADEVICEID)

Device ID of the monitored host.

COPY

COPY

string (256)

No

--

Target Host (TARGET_HOST)

Name of the monitored host.

COPY

FIXED

string (33)

No

--

Polling Time (POLLING_TIME)

Time when performance data was collected on the PFM - RM host.

COPY

FIXED

string (32)

No

--

Target Host Time (TARGET_HOST_TIME)

Time when performance data was collected on the monitored host.

COPY

FIXED

string (32)

No

--

Active CPUs (ACTIVE_CPUS)

Number of processors.

COPY

ADD

ulong

No

--

CPU % (CPU_PERCENT)

Processor usage rate.

[Units: %]

This is the average of all processors.

HILO

AVG

double

No

--

Idle % (IDLE_PERCENT)

Percentage of time the processors are idle.

[Units: %]

This is the average of all processors.

HILO

AVG

double

No

--

System % (SYSTEM_PERCENT)

Percentage of processor usage in kernel mode.

[Units: %]

This is the average of all processors.

HILO

AVG

double

No

--

User % (USER_PERCENT)

Percentage of processor usage in user mode.

[Units: %]

This is the average of all processors.

HILO

AVG

double

No

--

Wait % (WAIT_PERCENT)

Percentage of time the processors are waiting for I/O.

[Units: %]

This is the average of all processors.

HILO

AVG

double

No

Windows

Processor Queue Length (PROCESSOR_QUEUE_LENGTH)

Number of requests in the processor queue that are ready to execute and waiting for processor time. If the length of the queue continuously exceeds 2, the processor is probably busy.

HILO

AVG

double

No

UNIX

Run Queue Avg 5 min (RUN_QUEUE_AVG_5_MIN)

Average number of threads waiting in the execution queue for the past 5 minutes. In HP-UX, Solaris, and AIX, this value includes the number of I/O waiting threads. In LINUX, this value does not include the number of I/O waiting threads.

HILO

AVG

double

No

Windows

Interrupt Counts/sec (INTERRUPT_COUNTS_PER_SEC)

Windows:

Rate at which the processor processed interrupt requests generated by hardware devices (for example, the system clock, mouse, disk drivers, data communication line, and NIC). DPC (delay procedure call) interrupts are not included. If this field increases greatly when there is no system activity, a hardware problem (e.g., a low-speed device) probably exists.

[Units: times/second]

Unix:

Rate at which interrupts are generated.

[Units: times/second]

HILO

AVG

double

No

--

Effective Free Mem % (EFFECTIVE_FREE_MEM_PERCENT)

Percentage of physical memory available to applications.

[Units: %]

HILO

AVG

double

No

--

Effective Free Mem Mbytes (EFFECTIVE_FREE_MEM_MBYTES)

Amount of physical memory available to applications.

[Units: MB]

HILO

AVG

double

No

--

Effective Used Mem % (EFFECTIVE_USED_MEM_PERCENT)

The actual percentage of physical memory used.

[Units: %]

HILO

AVG

double

No

--

Effective Used Mem Mbytes (EFFECTIVE_USED_MEM_MBYTES)

The actual amount of physical memory used.

[Units: MB]

HILO

AVG

double

No

--

Free Mem % (FREE_MEM_PERCENT)

Percentage of physical memory that is unused.

[Units: %]

HILO

AVG

double

No

--

Free Mem Mbytes (FREE_MEM_MBYTES)

Amount of unused physical memory.

[Units: MB]

HILO

AVG

double

No

--

Used Mem % (USED_MEM_PERCENT)

Percentage of physical memory used.

[Units: %]

HILO

AVG

double

No

--

Used Mem Mbytes (USED_MEM_MBYTES)

Amount of used physical memory.

[Units: MB]

HILO

AVG

double

No

--

Total Mem Mbytes (TOTAL_MEM_MBYTES)

Amount of physical memory.

[Units: MB]

COPY

ADD

double

No

--

Free Swap % (FREE_SWAP_PERCENT)

Windows:

Percentage of virtual memory that is unused.

[Units: %]

UNIX:

Percentage of swap area that is unused.

[Units: %]

HILO

AVG

double

No

--

Free Swap Mbytes (FREE_SWAP_MBYTES)

Windows:

Amount of unused virtual memory.

[Units: MB]

Unix:

Amount of unused swap space.

[Units: MB]

HILO

AVG

double

No

--

Used Swap % (USED_SWAP_PERCENT)

Windows:

Percentage of virtual memory used.

[Units: %]

Unix:

Percentage of swap space used.

[Units: %]

HILO

AVG

double

No

--

Used Swap Mbytes (USED_SWAP_MBYTES)

Windows:

Size of area committed for virtual memory.

[Units: MB]

Unix:

Size of swap space.

[Units: MB]

HILO

AVG

double

No

--

Total Swap Mbytes (TOTAL_SWAP_MBYTES)

Windows:

Amount of virtual memory. [Units: MB]

UNIX:

Amount of swap area.

[Units: MB]

COPY

ADD

double

No

--

Page Fault Counts/sec (PAGE_FAULT_COUNTS_PER_SEC)

Frequency of page faults.

[Units: faults/second]

HILO

AVG

double

No

HP-UX, Linux

Page Scan Counts/sec (PAGE_SCAN_COUNTS_PER_SEC)

Frequency of page scans.

[Units: scans/second]

HILO

AVG

double

No

Windows, Linux, Solaris

Page-In Counts/sec (PAGE_IN_COUNTS_PER_SEC)

Rate of page-in operations.

[Units: operations/second]

HILO

AVG

double

No

Linux

Page-Out Counts/sec (PAGE_OUT_COUNTS_PER_SEC)

Rate of page-out operations.

[Units: operations/second]

HILO

AVG

double

No

Linux

Page-In Pages/sec (PAGE_IN_PAGES_PER_SEC)

Rate at which pages are paged in.

[Units: pages/second]

HILO

AVG

double

No

AIX

Page-Out Pages/sec (PAGE_OUT_PAGES_PER_SEC)

Rate at which pages are paged out.

[Units: pages/second]

HILO

AVG

double

No

AIX

Paging Pages/sec (PAGING_PAGES_PER_SEC)

Rate at which pages were being paged in and out when a page fault occurred.

[Units: pages/second]

This is the total of the Page-In Pages/sec and Page-Out Pages/sec fields. If this value continuously exceeds 5, lack of memory might be causing a system bottleneck.

HILO

AVG

double

No

UNIX

Pool Nonpaged KBytes (POOL_NONPAGED_KBYTES)

Amount of physical memory allocated for executing system component tasks for which page-outs could not be performed.

[Units: KB]

If this value continuously increases even though server processing is not becoming busier, a process might have a memory leak.

HILO

AVG

double

No

UNIX

Swap-In Counts/sec (SWAP_IN_COUNTS_PER_SEC)

Frequency of swap-in operations.

[Units: operations/second]

HILO

AVG

double

No

Windows, AIX, Linux

Swap-Out Counts/sec (SWAP_OUT_COUNTS_PER_SEC)

Frequency of swap-out operations.

[Units: operations/second]

HILO

AVG

double

No

Windows, AIX, Linux

Swap-In Pages/sec (SWAP_IN_PAGES_PER_SEC)

Frequency of page loading by swap-in operations.

[Units: pages/second]

HILO

AVG

double

No

Windows, AIX

Swap-Out Pages/sec (SWAP_OUT_PAGES_PER_SEC)

Frequency of page retrieval by swap-out operations.

[Units: pages/second]

HILO

AVG

double

No

Windows, AIX

Ext1 (EXT1)#

Extension field 1

HILO

AVG

double

No

All

Ext2 (EXT2)#

Extension field 2

HILO

AVG

double

No

All

Legend:

--: Supported on all OSs of the monitored hosts.

All: Not supported on any OS of the monitored hosts.

Smry rule: Summary rule

Grpg rule: Grouping rule

Not sprtd on: Not supported on

#

This field is not used for operations.