Performance Manager Commands › ADVISE PERFORMANCE GRAPH › Change a Rule Literal Value › Qualifiers

Qualifiers

/AVERAGE={DAILY | WEEKLY | MONTHLY | QUARTERLY}

Causes graphs to depict a summarization of a specified time period. The selected data is averaged into the time period selected. See Chapter 4 for more information.

If you also use the /SCHEDULE or /DATES qualifiers, the DAILY and WEEKLY graphs are trimmed to show only the selected hours.

If history data with the periodicity attribute is selected, the /AVERAGE value is automatically set to that periodicity value. This is true regardless of whether the /AVERAGE qualifier is used.

/BEGINNING=date

Specifies the beginning date and time of data selected for graphing.

Where date represents the date and time in standard DCL format.

The date and time format is the standard DCL format, either absolute or relative. If you do not specify the /BEGINNING qualifier, the Performance Manager uses 00:00:00 on the same day for which the ending date and time is specified. If you do not specify an /ENDING qualifier, the Performance Manager uses 00:00:00 of the current day as the default beginning time.

You can also use the keywords TODAY and YESTERDAY. See HP's OpenVMS User's Manual, or access the HELP topic SPECIFY DATE_TIME for complete information on specifying time values.

/BEGINNING is incompatible with the /DATES qualifier.

/COLLECTION_DEFINITION=collection-definition-name

Specifies the name of the Collection Definition, and hence the collected data that you desire to use for the graph. If you omit this qualifier, daily data is obtained from the Collection Definition called “CPD.”

To view the Collection Definitions that you have available, use the DCL command ADVISE COLLECT SHOW ALL.

If you want to use history data instead of daily data, use the /HISTORY qualifier instead of the /COLLECTION_DEFINITION qualifier. /COLLECTION_DEFINITION is incompatible with the /HISTORY qualifier.

/CLASSIFY_BY=USERGROUP=family_name

Specifies the workload family whose workload definitions are to be used for summarizing process activity. This affects the TOP_WORKLOAD graph types as well as custom graphs with WORKLOAD metrics by providing the desired metrics on an individual workload basis. The default is “other” which averages all process activity together. The family_type of USERGROUP is required. No restrictions are made on the family name.

/COMPOSITE

Combines data from all nodes into a single graph. Data from each node is either added or averaged.

The following command produces a graph of the total number of processes in the cluster.

$ ADVISE PERFORMANCE GRAPH/COMPOSITE/TYPE=PROCESSES

When the Performance Manager combines I/O data from more than one node, it is possible to double count I/O operations to a disk device if it is served. Therefore, when you specify /COMPOSITE, the Performance Manager does not count all MSCP-served I/O for individual disks.

When generating a customized graph for a single metric with /COMPOSITE, the Performance Manager graphs the metric by node.

When graphing CPU percentages with the /COMPOSITE qualifier each node's CPU time is scaled according to the VUP rating to produce a cluster average CPU utilization.

For more information, see the chapter Generate Historical Graphs.

/DATES=filespec

Specifies that a file containing a series of date ranges is to be used in place of the /BEGINNING and /ENDING qualifiers. Each line in the dates file should look like the following command:

dd-mmm-yyyy hh:mm:ss.cc,dd-mmm-yyyy hh:mm:ss.cc

The time can be either omitted entirely or truncated. Any truncated parts of the time defaults to 0. The periods of time represented by each line in the file need not be contiguous but they must be in ascending order.

/DATES is incompatible with the /BEGINNING and /ENDING qualifiers.

/ENDING=date

Specifies the ending date and time of the graph. Where date represents the date and time in standard DCL format.

If you do not specify /BEGINNING, /ENDING defaults to the current time. If you do specify /BEGINNING, the /ENDING default are midnight (23:59) of the beginning date.

You can specify either an absolute time or a combination of absolute and delta times. You can also use the keywords TODAY, TOMORROW, and YESTERDAY. See HP's OpenVMS User's Manual, or access the HELP topic SPECIFY DATE_TIME for complete information on specifying time values.

/ENDING is incompatible with the /DATES qualifier.

/FILTER=keyword

The /FILTER qualifier allows you select a subset of the daily or history data for graphing. Process data and disk data can be filtered.

Hotfile data is also filtered. When you specify filtering by process, a hotfile record is selected if accessed by the specified process. When you specify filtering by disk device, a hotfile record is selected if located on the specified device. For hotfile records matching both process and disk device, specify filtering by both process and device.

Process data can be filtered by using any of the filter keywords: USERNAMES, IMAGENAMES, PROCESSNAMES, ACCOUNTNAMES, UICS, PIDS or WORKLOADNAMES. If a process record's identification information matches any of the identification specifications that are specified, then that record is selected.

Likewise, disk data can be filtered by using either of the filter keywords, VOLUMENAMES and DEVICENAMES. If a device record's identification information matches any of the volume names or device names that are specified, then that record is selected.

The following table lists the /FILTER keyword options:

/FORMAT={
ReGIS[=(CHARACTERISTIC={COLOR | LINE | PATTERN}
[,X_POINTS=l ])] |
ANSI=[(HEIGHT=m,WIDTH=n,LINE)] |
TABULAR[=X_POINTS=l] |
CSV[=X_POINTS=l]
POSTSCRIPT=(CHARACTERISTIC={COLOR,LINE,
PATTERN},X_POINTS=l)

Where:

The Performance Manager graphs ReGIS or ANSI graph by default, depending on the device characteristics of the SYS$OUTPUT device. ANSI and ReGIS formats are not available with pie charts. You may override the default with the /FORMAT qualifier. A graph is one of four formats: ANSI, REGIS, TABULAR or PostScript.

Optionally, you may specify whether ReGIS graphs use LINE, PATTERN, or COLOR. COLOR is the default. PATTERN is incompatible with COLOR.

Use the X_POINTS keyword to specify the number of data points to plot across a ReGIS graph. The valid range for X_POINTS is 2 to 480. By default, the Performance Manager chooses a best-fit value for x-points so that the time period represented by each point is even.

As the value of X_POINTS increases, spikes and valleys become more defined and the graph has a higher resolution. A low number of X_POINTS produces a smoother graph because the graphing facility averages any additional data points within the time frame requested. Consider the time frame of a particular graph request when you determine the value of X_POINTS.

For example, over a 12-hour span, the Performance Manager records statistics 360 times (every 2 minutes). If the value of X_POINTS is 24, the graphing facility averages every 15 data records (or 30 minutes) and produces a graph with smooth flow. If the value of X_POINTS is 72, the graphing facility averages every 5 data records (or 10 minutes) and produces a graph with valleys and spikes.

Use the WIDTH keyword to specify the column width of the ANSI graph output. Valid widths range from 40 to 132 columns. If you do not specify the WIDTH qualifier, the Performance Manager uses the terminal width setting. When you specify the /OUTPUT qualifier or generate the graph under batch, the width of the graph is 132 columns.

Use the HEIGHT keyword to specify the graph height of the ANSI graph output. Valid heights are from 20 to 60 lines. If you do not specify HEIGHT, the Performance Manager uses the terminal page length setting. When you use the /OUTPUT qualifier or generate the graph under batch, the height of the graph is 40 lines.

/HISTORY=history_descriptor_name

Allows you to select history data from the Performance Manager database. By default, daily data files are used to supply data for graphing. However, by specifying the name of a history file descriptor, you can select historical data instead.

You must define the history file descriptor in the parameters file and have archived data according to the descriptor's definition. Use the DCL command ADVISE EDIT to invoke the Performance Manager Parameter Edit Utility. From the utility, you can ADD, DELETE, MODIFY, and SHOW history file descriptors. Use the ADVISE ARCHIVE command to create the archived files.

If history data with the periodicity attribute is selected, the /AVERAGE value is automatically set to that periodicity value. This is true regardless of whether the /AVERAGE qualifier is used.

/HISTORY is incompatible with the /COLLECTION_DEFINITION qualifier.

For information on how to produce a graph of history data including a typical time period, see the chapter Generate Historical Graphs.

Note: If model data was not archived, the /CLASSIFY_BY qualifier is restricted to those workload families specified in the history file descriptor.

/NODE_NAMES=(node-name[,...])

Identifies the nodes to graph.

The Performance Manager creates a separate graph for each node unless you specify the /COMPOSITE qualifier. If you omit the /NODE_NAMES qualifier, all the nodes in the schedule file associated with the specified collection definition (CPD by default) are used for the graph(s). If you specify only one node, the parentheses can be omitted. Do not use wildcard characters in the node-name specifications.

/OUTPUT=filespec

Creates an output file that contains the graphs. The default file extension for a ReGIS graph is .REG, the file type for ANSI and TABULAR formatted graphs is .RPT and the file extension for PostScript is .PS.

When you generate multiple graphs with a single command line, you can create a unique output file for each graph. To do this, omit the file name with the /OUTPUT qualifier. The Performance Manager generates a separate file for each graph created and uses the graph type keyword as the unique file name.

For example:

$ ADVISE PERFORMANCE GRAPH/NODE=SYSDEV/END=1/TYPE=(MEM,CPU_U,CPU_MODE) /OUTPUT=.REG
%PSPA-I-CREAGRAPHOUT, PSPA Graph created file MUMMS$DKA300:[CORREY]SYSDEV_CPU_UTILIZATION.REG;1
%PSPA-I-CREAGRAPHOUT, PSPA Graph created file MUMMS$DKA300:[CORREY]SYSDEV_MEMORY_UTILIZATION.REG;1
%PSPA-I-CREAGRAPHOUT, PSPA Graph created file MUMMS$DKA300:[CORREY]SYSDEV_CPU_MODES.REG;1

/RULES[=file-spec], /NORULES

Loads information from the rules file to establish user-defined hardware scaling factors. The file-spec must point to an auxiliary knowledge base which has previously been compiled with the ADVISE PERFORMANCE COMPILE command. The default file type is .KB. If the NORULES qualifier is specified no augmentation of the factory rules occur. See also the Chapter "Customize the Knowledge Base."

/SCHEDULE=({day=(hour-range)[,...]|NOday}[,...])

Specifies that a subset of Performance Manager data is to be used (or not used if keyword negation is specified) to generate graphs. By default, the Performance Manager selects all data between the /BEGINNING time and the /ENDING time, or as specified with the /DATES qualifier.

Where:

If you omit a day keyword, the data for that day is selected. Data selection for individual days of the week can be inhibited by negating the keyword (for example, NOSUNDAY) or for all of the days of the week by specifying the NOEVERYDAY keyword. The values [NO]WEEKDAYS and [NO]WEEKENDS similarly can be used to enable or disable data selection for weekdays and weekends.

You must specify an hour range for any non-negated day keyword. Do not include an hour range if you are specifying a negated day keyword, such as NOMONDAY.

Less inclusive keyword values override more inclusive values. For example, MONDAY=10--12 overrides EVERYDAY=8--17 for Monday, but the Performance Manager selects data from 8:00 a.m. to 5:00 p.m. for all of the other days of the week.

For example:

$ ADVISE PERFORMANCE GRAPH -
_$ /SCHEDULE=(NOEVERYDAY,WEEKDAYS=(8-12,13-17))

Graphs do not depict the time periods deselected by the /SCHEDULE qualifier.

/SELECT[={GREATER_THAN[:percent] | LESS_THAN[:percent]}], /NOSELECT

Use /SELECT in conjunction with the optional threshold values which may be specified on a per graph type basis.

If this qualifier is present, before a graph is produced, a check is made to see if the values to be graphed fall within the threshold values for the indicated percentage of points. If so, then the graph (or pie chart) is produced. If not, no graph is produced. For details on THRESHOLD, see the /TYPE qualifier.

These keywords accept a single value representing the percentage of the points plotted that must meet the threshold criteria before the graph is produced. Each graph point value is determined by the sum (STACKED) of the items depicted (up to 6). If the GREATER_THAN keyword is specified without a value, then 50 percent is assumed. If the LESS_THAN keyword is specified without a value, then 90 percent is assumed.

If the /SELECT qualifier is present without a keyword, then GREATER_THAN:50 is assumed. For example:

$ ADVISE PERFORMANCE GRAPH /BEGINNING=10/ENDING=11/NODE=YQUEM - 
_$ /TYPE=(CPU_U:THRESHOLD:25,CPU_M:THRESHOLD:35,TOP_CPU_I:THRESHOLD:45)-
_$ /SELECT=GREATER/OUTPUT=.REGIS
%PSPA-I-CREAGRAPHOUT, PSPA Graph created file BADDOG:[CORREY.WORK.PSPA]YQUEM_CPU_UTILIZATION.REG;1

This command requests that three graphs be produced. The CPU Utilization graph is produced, if 50 percent or more of the data points exceed 25 percent CPU utilization. The CPU_MODES graph is produced if 50 percent or more of the data points exceed 35 percent CPU utilization. The TOP_CPU_IMAGES graph is produced if 50 percent or more of the data points exceed 45 percent CPU utilization. In this case only one graph is produced.

$ ADVISE PERFORMANCE GRAPH /BEGINNING=10/ENDING=11/NODE=YQUEM -
_$ /TYPE=(CPU_U:THRESHOLD:25,CPU_M:THRESHOLD:35,TOP_CPU_I:THRESHOLD:15)- 
_$ /SELECT=GREATER/OUTPUT=.REGIS
%PSPA-I-CREAGRAPHOUT, PSPA Graph created file BADDOG:[CORREY.WORK.PSPA]YQUEM_CPU_UTILIZATION.REG;3
%PSPA-I-CREAGRAPHOUT, PSPA Graph created file BADDOG:[CORREY.WORK.PSPA]YQUEM_TOP_CPU_IMAGES.REG;1

This command produced two of three graphs because threshold quantity for the last graph was lowered.

$ ADVISE PERFORMANCE GRAPH /BEGINNING=10/ENDING=11/NODE=YQUEM - 
_$ /TYPE=(CPU_U:THRESHOLD:25,CPU_M:THRESHOLD:35,TOP_CPU_I:THRESHOLD:15) -
_$ /SELECT=GREATER:90/OUTPUT=.REGIS 

$

The previous command generated none of the graphs because in all cases 90 percent of the graph points did not exceed the specified thresholds.

/STACK, /NOSTACK

Stacks the values for each category on the graph. Use /NOSTACK to overlay the values on the graph. ReGIS graphs using /NOSTACK may cause some occlusion if you do not specify /FORMAT=ReGIS=CHARACTERISTICS=LINE also. If you are requesting a series of graphs in one command, you can override the /[NO]STACK qualifier by specifying the [NO]STACK keyword following each graph type. See Chapter 4 for an illustration of the use of the /NOSTACK qualifier and for additional information about default behavior.

/TYPE=
([NO]graph_type[=([NO]STACK,Y_AXIS_MAXIMUM=n,
TITLE=string)],...,
ALL_GRAPHS[=([NO]STACK,THRESHOLD=m,
Y_AXIS_MAXIMUM=n)],
CUSTOM=(see below “TYPE=CUSTOM”))

Specifies which of the graphs you want generated.

Use the TITLE keyword to override the Performance Manager supplied title. The text string may be a maximum of 40 characters.

The STACK keyword for a particular graph type overrides the setting established by the /STACK qualifier.

The THRESHOLD keyword specifies a threshold value associated with the graph. The m specifier is a positive decimal value. A horizontal line is placed on the graph at the position on the Y-axis associated with the value. You can use THRESHOLD in conjunction with the /SELECT qualifier to prevent the generation of the graph or pie chart.

The Y_AXIS_MAXIMUM specifies a fixed y-axis to be used for the graph. The default behavior is to setup the y-axis so that the maximum data point appears near the top of the graph. This graph modifier allows you to specify the y-axis so that you can compare data from different graphs without having different scales on the y-axis. The n specifier is a positive decimal value.

You can specify multiple graphs in a single command. For example, you can specify /TYPE=(TOP_IO_DISKS,TOP_HARDFAULTING_IMAGES). Of course, /TYPE=ALL_GRAPHS generate all of the predefined graphs. To suppress a graph type, specify NO graph_type.

CPU_UTILIZATION is the default graph type.

The following list contains all of the available Performance Manager graphs:

The following sections list the graph types and their descriptions. Included are keywords used with the /TYPE qualifier.

• /TYPE=CUSTOM

  You must specify the items for the Performance Manager to graph. The metrics and selection objects are described below.

  ADVISE PERFORMANCE
  GRAPH/TYPE=CUSTOM=({SYSTEM_METRICS=(system_metrics) |
  USER_METRICS=(process_metrics),SELECTION=(usernames) |
  IMAGE_METRICS=(process_metrics),SELECTION=(imagenames) |
  WORKLOAD_METRICS=(process_metrics),SELECTION=(workloadnames)|
  DEVICE_METRICS=(disk_metrics),SELECTION=(devicenames) |
  VOLUME_METRICS=(disk_metrics),SELECTION=(volumenames) |
  CPU_METRICS=(cpu_modes),SELECTION=(Phy-cpu-ids) |
  HSC_METRICS=(hsc_metrics),SELECTION=(HSC-nodenames) |
  SCS_METRICS=(scs_metrics),SELECTION=(SCS-nodenames) |
  RULE_METRICS=(rule_metrics),SELECTION=(Rule-ids) |
  CHANNEL_METRICS=(channel_metrics),SELECTION=(channel-specs) |
  FILE_METRICS=(file_metrics),SELECTION=(file-names) |
  DISK_USER_METRICS=(disk_user_metrics),SELECTION=(username-volumename) |
  DISK_IMAGE_METRICS=(disk_image_metrics),SELECTION=(imagename-volumename)}
  [,[NO]STACK] [,Y_AXIS_MAXIMUM=n] [,THRESHOLD=m] [,TITLE=string])
  

Where:

The CUSTOM graph type allows you to graph a selection of metrics for either the system, or selected users, images, workloads, disk devices, volumes, HSCs, SCS nodes, rule-ids or channels. You may graph up to six selections with a single metric, or up to six metrics with a single selection. The Performance Manager either prompts you in command mode for the data (ADVISE PERFORMANCE) or you can specify the desired metrics and selections in a single DCL command.

For example:

$ ADVISE PERFORMANCE GRAPH/TYPE = CUSTOM = SYSTEM_METRICS = -
_$ (DZROFAULTS,GVALID)

The SELECTION string must be chosen based on the metric class that you use:

• If the metric class is USER_METRICS then the selection strings are interpreted as user names.
• If the metric class is IMAGE_METRICS then the selection strings are interpreted as image names.
• If the metric class is WORKLOAD_METRICS then the selection strings are interpreted as workload names. (Unless you use the /CLASSIFY_BY qualifier to characterize the process data into various workloads, all workload data is grouped into the default workload called “OTHER.”)
• If the metric class is DEVICE_METRICS then the selection strings are interpreted as device names.
• If the metric class is VOLUME_METRICS then the selection strings are interpreted as volume names.
• If the metric class is CPU_METRICS then the selection strings are interpreted as physical processor IDs which are in the form NODENAME_INTEGER, such as, NODE1_3.

  To display a graph which shows active CPUs in an OpenVMS multiprocessing system, enter a command similar to the following:

  $ ADVISE PERFORMANCE GRAPH/END=0:10/NODE=YQUEM -
  _$ /TYPE=TOP_BUSY_PROCESSOR

  Specifying a physical CPU ID allows you to isolate and analyze one CPU of a selected node in an SMP configuration.

• If the metric class is HSC_METRICS then the selection strings are interpreted as HSC node names.
• If the metric class is SCS_METRICS then the selection strings are interpreted as cluster node names.
• If the metric class is FILE_METRICS then the selection strings are interpreted as file names.
• If the metric class is RULE_METRICS then the selection strings are interpreted as rule IDs.

Note: Rule Metrics are available only from history files.

• If the metric class is DISK_USER_METRICS, then the selection strings are interpreted as username-volumename.
• If the metric class is DISK_IMAGE_METRICS, then the selection strings are interpreted as imagename-volumename.
• If the metric class is CHANNEL_METRICS then the selection strings are interpreted as a channel-spec which is the HSC nodename, an underscore, and the HSC's channel number (for example, HSC001_6).

The following tables identify the custom graphing metrics grouped by metric class.

/TYPE=COMPUTE_QUEUE (Number of Processes in COM and COMO)

Plots the number of computable processes categorized by:

• Network processes (NETWORK_COMQ)
• Interactive processes (INT_COMQ)
• Batch processes (BATCH_COMQ)
• Detached processes (DETACHED_COMQ)

/TYPE=CPU_MODES (CPU Modes)

Plots the percentage of CPU time spent in the various processor modes:

• Multiprocessor synchronization (MP_SYNCH)
• User (USER_MODE)
• Supervisor (SUPER)
• Executive (EXEC)
• Kernel (KERNEL)
• Interrupt Stack (INTERRUPT)

/TYPE=CPU_UTILIZATION (CPU Utilization)

Plots 6 metrics for percent CPU utilization:

• Interrupt stack and MP Synch (CPU_MP_INT)
• Detached processes (CPU_DETACHED)
• Interactive processes (CPU_INTERACTIVE)
• Batch processes (CPU_BATCH)
• Network processes (CPU_NETWORK)
• Other (CPU_OTHER)

This is the default graph type.

/TYPE=DECNET (System-Wide DECnet Traffic)

Plots the number of DECnet operations per second in terms of:

• Arriving packets (ARRLOCPK)
• Departing packets (DEPLOCPK)
• Transit packets (ARRTRAPK)

/TYPE=DISKS (DISK I/O)

Plots the disk operations per second categorized by:

• User (DISK_USER)
• Paging (DISK_SWAPPING)
• Swapping (DISK_PAGING)

/TYPE=FAULTS (Page Fault Rate)

Plots the page fault rate per second, and places the rate into these categories:

• Demand zero page faults (DZROFAULTS)
• Free page faults (FREEFAULTS)
• Modified page faults (MFYFAULTS)
• Global page faults (GVALID)
• Hard page faults (PREADIO)
• System page faults (SYSFAULTS)

/TYPE=FILECACHE (File Cache Usage)

Plots the file operation attempt rate to the file system caches categorized by:

• Hits (FILE_CACHE_HIT)
• Misses (FILE_CACHE_MISS)

/TYPE=JOBS (Number of Jobs)

Plots the number of processes categorized by:

• Interactive (INTERACTIVE_PROCESSES)
• Batch (BATCH_PROCESSES)
• Network (NETWORK_PROCESSES)
• Detached (DETACHED_PROCESSES)

/TYPE=LOCKS (Distributed Locking)

Plots the number of distributed lock operations per second categorized by:

• Incoming enqueues and converts (INCOMING_LOCKING)
• Outgoing enqueues and converts (OUTGOING_LOCKING)
• Local enqueues and converts (LOCAL_LOCKING)

/TYPE=MEMORY_UTILIZATION (Memory Utilization)

Plots physical memory usage categorized by:

• Percentage allocated to the free list (FREELIM)
• Percentage in the modified page list (MODIFIED)
• Percentage allocated to user processes (USERWS)
• Percentage allocated to system processes (SYSTEMWS)
• Percentage allocated to OpenVMS (VMSALLOC)

/TYPE=PROCESSES (Number of Processes by State)

Plots the number of processes categorized as:

• Inswapped/active (INPROCACT)
• Inswapped/inactive (INPROCINACT)
• Outswapped/inactive (OUTPROCINACT)
• Outswapped/active (OUTPROCACT)

/TYPE=RESPONSE_TIME (Terminal Response Time)

Plots the terminal response time for interactive processes categorized as:

• Large transactions (LG_RESPONSE)
• Medium transactions (MED_RESPONSE)
• Small transactions (SM_RESPONSE)

/TYPE=TERMINALS (Terminal I/O)

Plots the number of terminal operations per second categorized by the type of terminal used:

• TX (TX_TERMIO)
• TT (TT_TERMIO)
• RT (RT_TERMIO)
• LT (LAT_TERMIO)
• NV (NV_TERMIO)
• TW (TW_TERMIO)

/TYPE=TOP_BDT_W (Top BDT Wait Rate)

Plots five remote nodes with the highest rate of BDT waits (plus “Other”) resulting when a local node issues an I/O, but the connection had to wait for a buffer descriptor. The metric graphed is BDT_W.

/TYPE=TOP_BLKS_R (Top Blk Transfers Requested)

Plots the top five nodes with the highest block transfer requests (plus “Other”) from the remote system to the local system. The metric graphed is BLKS_R.

/TYPE=TOP_BLKS_S (Top Blk Transfers Sent)

Plots the top five nodes with the highest block transfers sent (plus “Other”) from the local system to the remote system. The metric graphed is BLKS_S.

/TYPE=TOP_BUFIO_IMAGES (Top Buffered I/O Images)

Plots the top five (plus “Other Images”) creators of buffered I/O by image names. The metric graphed is BUFIO.

/TYPE=TOP_BUFIO_USERS (Top Buffered I/O Users)

Plots the top five (plus “Other Users”) creators of buffered I/O by user names. The metric graphed is BUFIO.

/TYPE=TOP_BUFIO_WORKLOADS (Top Buffered I/O Workloads)

Plots the top five (plus “Other Users”) creators of buffered I/O by workload names. The metric graphed is BUFIO.

/TYPE=TOP_BUSY_DISKS (Top Busy Disk Device)

Plots the five (plus “Other Disks”) disk devices that experienced the highest busy time percentages. The metric graphed is BUSY.

/TYPE=TOP_BUSY_PROCESSOR (Top Busy Physical Processor)

Plots the five (plus “Other”) processors that experienced the highest busy time percentages. The metric graphed is P_BUSY.

/TYPE=TOP_BUSY_VOLUMES (Top Busy Disk Volume)

Plots the five (plus “Other Volumes”) disk volumes that experienced the highest busy time percentages. The metric graphed is BUSY.

/TYPE=TOP_CHANNEL_IO (Top HSC Channel I/O)

Plots the five (plus “Other”) HSC channels that experienced the largest I/O rate, in I/Os per second. The metric graphed is CHANNEL_IO.

/TYPE=TOP_CHANNEL_QUELEN (Top Queue HSC Channel)

Plots the five (plus “Other”) HSC channels that experienced the largest queue length. The metric graphed is CHANNEL_QUELEN.

Note: The channel names are provided in the format nodename_n, where n represents the channel number (K.SDI) on the HSC node indicated by node name. If the channel cannot be identified, the character u is substituted for n. See logical name PSDC$hscname_hscunitnumber in the Performance Agent Administrator Guide.

/TYPE=TOP_CHANNEL_THRUPUT (Top HSC Channel Thruput)

Plots the five (plus “Other”) HSC channels that experienced the largest throughput rate, in Kilobytes per second. The metric graphed is CHANNEL_THRUPUT.

Note: The channel names are provided in the format nodename_n, where n represents the channel number (K.SDI) on the HSC node indicated by node name. If the channel cannot be identified, the character u is substituted for n. See logical name PSDC$hscname_hscunitnumber in the Performance Agent Administrator Guide.

/TYPE=TOP_CLUSTER_RULE_OCC (Top Cluster Rule Occurrences)

Plots the five (plus “Other”) rule identifiers that fired, as a rate per hour. The metric graphed is CLUSTER_OCCURRENCES and is available only from history data.

/TYPE=TOP_COMPAT_PROCESSOR (Top Compat Mode Processor)

Plots the five (plus “Other”) processors in terms of time spent in compatibility mode, as a percent of CPU time. The metric graphed is P_COMPAT.

/TYPE=TOP_CPU_IMAGES (Top CPU Images)

Plots the top five (plus “Other Images”) consumers of CPU time by image name. The metric graphed is CPUTIME.

/TYPE=TOP_CPU_RULE_OCC (Top CPU Rule Occurrences)

Plots the five (plus “Other”) CPU rule identifiers that fired, as a rate per hour. The metric graphed is CPU_OCCURRENCES and is available only from history data.

/TYPE=TOP_CPU_USERS (Top CPU Users)

Plots the top five (plus “Other Users”) consumers of CPU time by user name. The metric graphed is CPUTIME.

/TYPE=TOP_CPU_WORKLOADS (Top CPU Workloads)

Plots the top five (plus “Other”) workloads as consumers of CPU time. The metric graphed is CPUTIME.

/TYPE=TOP_CR_W (Top Credit Wait Rate)

Plots five nodes with the highest rate of credit waits (plus “Other”) resulting when a connection has to wait for a send credit. The metric graphed is CR_W.

/TYPE=TOP_DGS_D (Top Datagrams Discarded)

Plots five nodes with the most datagrams discarded (plus “Other”)resulting when application datagrams are discarded by the port driver. The metric graphed is DGS_D.

/TYPE=TOP_DGS_R (Top Datagrams Received)

Plots five nodes with the most datagrams received (plus “Other”)resulting when the local system receives datagrams over the connection from the remote system and given to SYSAP. The metric graphed is DGS_R.

/TYPE=TOP_DGS_S (Top Datagrams Sent)

Plots five nodes with the most datagrams sent (plus “Other”) resulting when application datagrams are sent over the connection. The metric graphed is DGS_S.

/TYPE=TOP_DIRIO_IMAGES (Top Direct I/O Images)

Plots the top five (plus “Other Images”) creators of direct I/O by image name. The metric graphed is DIRIO.

/TYPE=TOP_DIRIO_USERS (Top Direct I/O Users)

Plots the top five (plus “Other Users”) creators of direct I/O by user name. The metric graphed is DIRIO.

/TYPE=TOP_DIRIO_WORKLOADS (Top Direct I/O Workloads)

Plots the top five (plus “Other Users”) creators of direct I/O by workload name. The metric graphed is DIRIO.

/TYPE=TOP_DISKIO_IMAGES (Top Image I/O Operations)

Plots the top five (plus “Other Images”) creators of disk I/O by image name. The metric graphed is DSKIO.

/TYPE=TOP_DISKIO_USERS (Top User Disk Operations)

Plots the top five (plus “Other Users”) creators of disk I/O by user name. The metric graphed is DSKIO.

/TYPE=TOP_DISKIO_WORKLOADS (Top Workload Disk Operations)

Plots the top five (plus “Other”) creators of disk I/O by workload name. The metric graphed is DSKIO.

/TYPE=TOP_EXEC_PROCESSOR (Top Exec Mode Processor)

Plots the five (plus “Other”) processors in terms of time spent in executive mode, as a percent of CPU time. The metric graphed is P_EXEC.

/TYPE=TOP_FAULTING_IMAGES (Top Faulting Images)

Plots the top five (plus “Other Images”) creators of page faults by image name. The metric graphed is FAULTS.

/TYPE=TOP_FAULTING_USERS (Top Faulting Users)

Plots the top five (plus “Other Users”) creators of page faults by user name. The metric graphed is FAULTS.

/TYPE=TOP_FAULTING_WORKLOADS (Top Faulting Workloads)

Plots the top five (plus “Other Users”) creators of page faults by workload name. The metric graphed is FAULTS.

/TYPE=TOP_FREEBLK_DISKS (Top Freeblks Disk Device)

Plots the top five (plus “Other”) disk devices in terms of number of free disk pages. The metric graphed is FREEBLKS.

/TYPE=TOP_FREEBLK_VOLUMES (Top Freeblks Disk Volume)

Plots the top five (plus “Other”) disk volumes in terms of number of free disk pages. The metric graphed is FREEBLKS.

/TYPE=TOP_HARDFAULTING_IMAGES (Top Hard Faulting Images)

Plots the top five (plus “Other Images”) creators of hard page faults by image name. The metric graphed is HARDFAULTS.

/TYPE=TOP_HARDFAULTING_USERS (Top Hard Faulting Users)

Plots the top five (plus “Other Users”) creators of hard page faults by user name. The metric graphed is HARDFAULTS.

/TYPE=TOP_HARDFAULTING_WORKLOADS (Top Hard Faulting Workloads)

Plots the top five (plus “Other Users”) creators of hard page faults by workload name. The metric graphed is HARDFAULTS.

/TYPE=TOP_HSC_DISK_IO (Top HSC Disk IO)

Plots the top five (plus “Other”) HSCs in terms of disk I/O operations per second. The metric graphed is HSC_DISK_IO.

/TYPE=TOP_HSC_DISK_THRUPUT (Top HSC Disk Thruput)

Plots the top five (plus “Other”) HSCs in terms of disk throughput in Kilobytes per second. The metric graphed is HSC_DISK_THRUPUT.

/TYPE=TOP_HSC_IO (Top HSC IO)

Plots the top five (plus “Other”) HSCs in terms of I/O operations per second. The metric graphed is HSC_IO.

/TYPE=TOP_HSC_TAPE_IO (Top HSC Tape IO)

Plots the top five (plus “Other”) HSCs in terms of tape I/O operations per second. The metric graphed is HSC_TAPE_IO.

/TYPE=TOP_HSC_TAPE_THRUPUT (Top HSC Tape Thruput)

Plots the top five (plus “Other”) HSCs in terms of tape thruput in Kilobytes per second. The metric graphed is HSC_TAPE_THRUPUT.

/TYPE=TOP_HSC_THRUPUT (Top HSC Thruput)

Plots the top five (plus “Other”) HSCs in terms of total thruput in Kilobytes per second. The metric graphed is HSC_THRUPUT.

/TYPE=TOP_IMAGE_ACTIVATIONS (Top Images Activated)

Plots the top five (plus “Other”) images in terms of image activations per second. The metric graphed is IMAGE_ACTIVATIONS.

/TYPE=TOP_IMAGE_VOLUME_IO (Top I/O Images and the Disk Volumes they access)

Plots the top five (plus “Other”) image and volume name pairs in terms of their I/O rate. The metric graphed is IMAGE_VOLUME_IO.

/TYPE=TOP_INTERRUPT_PROCESSOR (Top Interrupt Stack Processor)

Plots the five (plus “Other”) processors in terms of time spent on the interrupt stack, as a percent of CPU time. The metric graphed is P_INTERRUPT.

/TYPE=TOP_IO_DISKS (Top Operations Disk Device)

Plots the five (plus “Other Disks”) disk devices that incurred the highest I/O rates. The metric graphed is TOTIO.

/TYPE=TOP_IO_FILES (Top IO Operations Files)

Plots the five (plus “Other”) files that incurred the highest I/O rates. The metric graphed is FILE_TOTIO.

/TYPE=TOP_IO_RULE_OCC (Top IO Rule Occurrences)

Plots the five (plus “Other”) IO rule identifiers that fired, as a rate per hour. The metric graphed is IO_OCCURRENCES.

/TYPE=TOP_IO_VOLUMES (Top Operations Disk Volume)

Plots the five (plus “Other Volumes”) disk volumes that incurred the highest I/O rates.

/TYPE=TOP_KB_MAP (Top Kilobyte Mapped Rate)

Plots five nodes (plus “Other”) in terms of the number of kilobytes of data mapped for block transfer. The metric graphed is KB_MAP.

/TYPE=TOP_KB_RC (Top Kilobyte Received Rate)

Plots five nodes (plus “Other”) in terms of the number of kilobytes of data received by the local system from the remote system through request-data commands. The metric graphed is KB_RC.

/TYPE=TOP_KB_S (Top KB Sent Rate)

Plots five nodes (plus “Other”) in terms of the number of kilobytes of data sent from the local system to the remote system through send-data commands. The metric graphed is KB_S.

/TYPE=TOP_KERNEL_PROCESSOR (Top Kernel Mode Processor)

Plots the five (plus “Other”) processors in terms of time spent in kernel mode as a percent of CPU time. The metric graphed is P_KERNEL.

/TYPE=TOP_MEMORY_RULE_OCC (Top Memory Rule Occurrences)

Plots the five (plus “Other”) memory rule identifiers that fired, as a rate per hour. The metric graphed is MEMORY_OCCURRENCES and is available only from history data.

/TYPE=TOP_MGS_R (Top Messages Received)

Plots five nodes (plus “Other”) in terms of number of application datagram messages received over the connection. The metric graphed is MGS_R.

/TYPE=TOP_MGS_S (Top Messages Sent)

Plots five nodes (plus “Other”) in terms of number of application datagram messages sent over the connection. The metric graphed is MGS_S.

/TYPE=TOP_MP_SYNCH_PROCESSOR (Top MP Synch Mode Processor)

Plots the five (plus “Other”) processors in terms of time spent in MP synchronization mode, as a percent of CPU time. The metric graphed is P_MP_SYNCH.

/TYPE=TOP_MSCPIO_FILES (Top MSCP I/O Operations Files)

Plots the five (plus “Other”) files that incurred the highest MSCP I/O rates. The metric graphed is FILE_MSCPIO.

/TYPE=TOP_PAGING_DISKS (Top PG&SWP Operations Disk Device)

Plots the five (plus “Other Disks”) disk devices that incurred the highest I/O paging and swapping rates. The metric graphed is PAGIO.

/TYPE=TOP_PAGING_FILES (Top PG&SWP Operations Files)

Plots the five (plus “Other”) files that incurred the highest I/O paging and swapping rates. The metric graphed is FILE_PAGIO.

/TYPE=TOP_PAGING_VOLUMES (Top PG&SWP Operations Disk Volume)

Plots the five (plus “Other Volumes”) disk volumes that incurred the highest I/O paging and swapping rates. The metric graphed is PAGIO.

/TYPE=TOP_POOL_RULE_OCC (Top Pool Rule Occurrences)

Plots the five (plus “Other”) pool rule identifiers that fired, as a rate per hour. The metric graphed is POOL_OCCURRENCES and is available only from history data.

/TYPE=TOP_PRCT_FREE_DISKS (Top Percent Freeblks Disk Device)

Plots the top five (plus “Other”) disk devices in terms of percentage of free disk blocks. The metric graphed is PRCT_FREE.

/TYPE=TOP_PRCT_USED_DISKS (Top Percent Usedblks Disk Device)

Plots the top five (plus “Other”) disk devices in terms of percentage of used disk blocks. The metric graphed is PRCT_USED.

/TYPE=TOP_PRCT_FREE_VOLUMES (Top Percent Freeblks Disk Volume)

Plots the top five (plus “Other”) disk volumes in terms of percentage of free disk blocks. The metric graphed is PRCT_FREE.

/TYPE=TOP_PRCT_USED_VOLUMES (Top Percent Usedblks Disk Volume)

Plots the top five (plus “Other”) disk volumes in terms of percentage of used disk blocks. The metric graphed is PRCT_USED.

/TYPE=TOP_QUEUE_DISKS (Top Queue Disk Device)

Plots the five (plus “Other Disks”) disk devices that experienced the longest queue lengths. The metric graphed is QUEUE.

/TYPE=TOP_QUEUE_VOLUMES (Top Queue Disk Volume)

Plots the five (plus “Other Volumes”) disk volumes that experienced the longest queue lengths. The metric graphed is QUEUE.

/TYPE=TOP_READ_DISKS (Top Read Operations Disk Device)

Plots the five (plus “Other Disks”) disk devices that incurred the highest read I/O rates. The metric graphed is READIO.

/TYPE=TOP_READ_FILES (Top Read Operations Files)

Plots the five (plus “Other”) files that incurred the highest read I/O rates. The metric graphed is FILE_READIO.

/TYPE=TOP_READ_VOLUMES (Top Read Operations Disk Volume)

Plots the five (plus “Other Volumes”) disk volumes that incurred the highest read I/O rates. The

metric graphed is READIO.

/TYPE=TOP_RESIDENT_IMAGES (Most Resident Images)

Plots the top five (plus “Other Images”) images most resident on the system by image name. The metric graphed is RESIDENCE.

/TYPE=TOP_RESIDENT_USERS (Most Resident Users)

Plots the top five (plus “Other Users”) users most resident on the system by user name. Note that each subprocess adds to the residence for the parent process's user name. The metric graphed is RESIDENCE.

/TYPE=TOP_RESIDENT_WORKLOADS (Most Resident Workloads)

Plots the top five (plus “Other Workloads”) workloads most resident on the system by workload name. The metric graphed is RESIDENCE.

/TYPE=TOP_RESOURCE_RULE_OCC (Top Resource Rule Occurrences)

Plots the five (plus “Other”) resource rule identifiers that fired, as a rate per hour. The metric graphed is RESOURCE_OCCURRENCES and is available only from history data.

/TYPE=TOP_RESPONSE_TIME_DISKS (Top Response Time Disk Device)

Plots the five (plus “Other Disks”) disk devices that incurred the highest response times. The metric graphed is D_RESPONSETIME.

/TYPE=TOP_RESPONSE_TIME_FILES (Top Response Time Files)

Plots the five (plus “Other”) files that incurred the highest response times. The metric graphed is FILE_RESPONSE_TIME.

/TYPE=TOP_RESPONSE_TIME_IMAGES (Top Image Response Time)

Plots the five (plus “Other Images”) images with the highest terminal response time. The metric graphed is RESPONSE_TIME.

/TYPE=TOP_RESPONSE_TIME_USERS (Top User Response Time)

Plots the five (plus “Other Users”) users with the highest terminal response time. The metric graphed is RESPONSE_TIME.

/TYPE=TOP_RESPONSE_TIME_VOLUMES (Top Response Time Disk Volume)

Plots the five (plus “Other Volumes”) disk volumes that have the highest response times. The metric graphed is D_RESPONSETIME.

/TYPE=TOP_RESPONSE_TIME_WORKLOADS (Top Workload Response Time)

Plots the five (plus “Other Workloads”) workloads with the highest terminal response time. The metric graphed is RESPONSE_TIME.

/TYPE=TOP_RULE_OCCURRENCES (Top Rule Occurrences)

Plots the five (plus “Other”) rule identifiers that fired, as a rate per hour. The metric graphed is OCCURRENCES and is available only from history data.

/TYPE=TOP_SPLITIO_DISKS (Top Split Operations Disk Device)

Plots the five (plus “Other”) disk devices that have the highest split I/O operations. The metric graphed is SPLITIO.

/TYPE=TOP_SPLITIO_FILES (Top Split Operations Files)

Plots the five (plus “Other”) files that have the highest split I/O operations. The metric graphed is FILE_SPLITIO.

/TYPE=TOP_SPLITIO_VOLUMES (Top Split Operations Disk Volume)

Plots the five (plus “Other”) disk volumes that have the highest split I/O operations. The metric graphed is SPLITIO.

/TYPE=TOP_SUPER_PROCESSOR (Top Supervisor Mode Processor)

Plots the five (plus “Other”) processors in terms of time spent in Supervisor mode, as a percent of CPU time. The metric graphed is P_SUPER.

/TYPE=TOP_TERMINAL_INPUT_IMAGES (Top Image Terminal Input)

Plots the top five (plus “Other Images”) images with the highest character per second terminal input. The metric graphed is TERM_INPUT.

/TYPE=TOP_TERMINAL_INPUT_USERS (Top User Terminal Input)

Plots the top five (plus “Other Users”) users with the highest character per second terminal input. The metric graphed is TERM_INPUT.

/TYPE=TOP_TERMINAL_INPUT_WORKLOADS (Top Workload Terminal Input)

Plots the top five (plus “Other Workloads”) workloads with the highest character per second terminal input. The metric graphed is TERM_INPUT.

/TYPE=TOP_TERMINAL_THRUPUT_IMAGES (Top Image Terminal Thruput)

Plots the top five (plus “Other Images”) images with the highest character per second terminal thruput. The metric graphed is TERM_THRUPUT.

/TYPE=TOP_TERMINAL_THRUPUT_USERS (Top User Terminal Thruput)

Plots the top five (plus “Other Users”) users with the highest character per second terminal thruput. The metric graphed is TERM_THRUPUT.

/TYPE=TOP_TERMINAL_THRUPUT_WORKLOADS (Top Workload Terminal Thruput)

Plots the top five (plus “Other Workloads”) workloads with the highest character per second terminal thruput. The metric graphed is TERM_THRUPUT.

/TYPE=TOP_THRUPUT_DISKS (Top Throughput Disk Device)

Plots the five (plus “Other Disks”) disk devices that incurred the highest throughput rates. The metric graphed is THRUPUT.

/TYPE=TOP_THRUPUT_FILES (Top Throughput Files)

Plots the five (plus “Other”) files that incurred the highest throughput rates. The metric graphed is FILE_THRUPUT.

/TYPE=TOP_THRUPUT_IMAGES (Top Throughput Images)

Plots the five (plus “Other”) images with the highest throughput rates. The metric graphed is THRUPUT.

/TYPE=TOP_THRUPUT_USERS (Top Throughput Users)

Plots the five (plus “Other”) users with the highest throughput rates. The metric graphed is THRUPUT.

/TYPE=TOP_THRUPUT_VOLUMES (Top Throughput Disk Volume)

Plots the five (plus “Other”) disk volumes that incurred the highest throughput rates. The metric graphed is THRUPUT.

/TYPE=TOP_THRUPUT_WORKLOADS (Top Throughput Workloads)

Plots the five (plus “Other”) workloads with the highest throughput rates. The metric graphed is THRUPUT.

/TYPE=TOP_USER_IMAGE_ACTIVATIONS (Top Image Activations Users)

Plots the top five (plus “Other”) users in terms of image activations per second. The metric graphed is IMAGE_ACTIVATIONS.

/TYPE=TOP_USER_PROCESSOR (Top User Mode Processor)

Plots the five (plus “Other”) processors in terms of time spent in User mode, as a percent of CPU time. The metric graphed is P_USER.

/TYPE=TOP_USER_VOLUME_IO (Top I/O Users and the Disk Volumes they access)

Plots the top five (plus “Other”) user and volume name pairs in terms of their I/O rate. The metric graphed is USER_VOLUME_IO.

/TYPE=TOP_WORKLOAD_IMAGE_ACTIVATIONS (Top Image Activations Workload)

Plots the top five (plus “Other”) workloads in terms of image activations per second. The metric graphed is IMAGE_ACTIVATIONS.

/TYPE=TOP_WRITE_DISKS (Top Write Operations Disk Device)

Plots the five (plus “Other Disks”) disk devices that incurred the highest write I/O rates. The metric graphed is WRITIO.

/TYPE=TOP_WRITE_FILES (Top Write Operations Files)

Plots the five (plus “Other”) files that incurred the highest write I/O rates. The metric graphed is FILE_WRITIO.

/TYPE=TOP_WRITE_VOLUMES (Top Write Operations Disk Volume)

Plots the five (plus “Other Volumes”) disk volumes that incurred the highest write I/O rates. The metric graphed is WRITIO.

/TYPE=TOP_VA_IMAGES (Top VA Space Images)

Plots the top five (plus “Other Images”) images that had the largest combined virtual address space by image name. The metric graphed is VASIZE.

/TYPE=TOP_VA_USERS (Top VA Space Users)

Plots the top five (plus “Other Users”) users that had the largest combined virtual address space by user name. The metric graphed is VASIZE.

/TYPE=TOP_VA_WORKLOADS (Top VA Space Workload)

Plots the top five (plus “Other”) workloads that had the largest combined virtual address space. The metric graphed is VASIZE.

/TYPE=TOP_WSSIZE_IMAGES (Top WS Size Images)

Plots the top five (plus “Other Images”) images that had the largest combined working set sizes by image name. The metric graphed is WSSIZE.

/TYPE=TOP_WSSIZE_USERS (Top WS Size Users)

Plots the top five (plus “Other Users”) users that had the largest combined working set sizes by user name. The metric graphed is WSSIZE.

/TYPE=TOP_WSSIZE_WORKLOADS (Top WS Size Workload)

Plots the top five (plus “Other”) workloads that had the largest combined working set sizes. The metric graphed is WSSIZE.