Measure Reference Manual Abstract This manual describes the command and programmatic interfaces of the Measure performance monitor. This manual is for system analysts and performance-management specialists who monitor performance on NonStop systems. HP Part Number: 523324-020 Published: November 2013 Edition: J06.03 and subsequent J-series RVUs, H06.03 and subsequent H-series RVUs, G06.03 and subsequent G-series RVUs, and D40.00 and subsequent D-series RVUs.
© Copyright 2013 Hewlett-Packard Development Company, L.P. Legal Notice Confidential computer software. Valid license from HP required for possession, use or copying. Consistent with FAR 12.211 and 12.212, Commercial Computer Software, Computer Software Documentation, and Technical Data for Commercial Items are licensed to the U.S. Government under vendor’s standard commercial license. The information contained herein is subject to change without notice.
Contents About This Document...................................................................................17 Supported Release Version Updates (RVUs)................................................................................17 Intended Audience..................................................................................................................17 New and Changed Information................................................................................................
ADD MEASUREMENT.............................................................................................................45 Syntax .............................................................................................................................45 Related Commands............................................................................................................46 Usage Notes for All PVUs and RVUs.....................................................................................
Examples..........................................................................................................................60 HISTORY...............................................................................................................................61 Syntax .............................................................................................................................61 Related Commands...........................................................................................
Related Commands............................................................................................................93 Usage Notes for All PVUs and RVUs.....................................................................................93 Usage Notes for H-Series and J-Series PVUs and RVUs............................................................93 Usage Note for G-Series PVUs.............................................................................................94 Examples..................
Example.........................................................................................................................108 SET REPORT.........................................................................................................................109 Syntax............................................................................................................................109 Related Commands...............................................................................................
Examples........................................................................................................................131 SWAPVOL...........................................................................................................................132 Syntax ...........................................................................................................................132 Usage Note......................................................................................................
Usage Note for G-Series CLUSTER Entities...........................................................................158 CONTROLLER......................................................................................................................158 Entity Specification Syntax for CONTROLLER Entities.............................................................158 DDL Record for CONTROLLER Entities (Legacy Style).............................................................
DISKFILE..............................................................................................................................217 Entity Specification Syntax for DISKFILE Entities....................................................................217 DDL Record for DISKFILE Entities (ZMS Style).......................................................................218 DDL Record for DISKFILE Entities (Legacy Style)....................................................................
OSSCPU..............................................................................................................................270 Entity Specification Syntax for OSSCPU Entities....................................................................270 DDL Record for OSSCPU Entities (ZMS Style).......................................................................271 DDL Record for OSSCPU Entities (Legacy Style)....................................................................
Common ID Field DDL Definitions for SQLPROC Entities........................................................339 Common Counter Field DDL Definitions for SQLPROC Entities................................................340 Usage Note for New Format SQLPROC Entities....................................................................341 SQLSTMT............................................................................................................................
Reading in the Declaration Files.........................................................................................376 Allocating Space for the Measure Control Block...................................................................376 Specifying Entity Descriptors..............................................................................................376 Creating the Configuration Table.......................................................................................
Usage Notes...................................................................................................................435 MEAS_RETRIEVE_TIMERCELLS_..............................................................................................436 MEAS_SQL_MAP_INIT_........................................................................................................437 Usage Notes...................................................................................................................
A Error Messages......................................................................................495 B Error Codes...........................................................................................518 C Subsystem Files......................................................................................548 C and C++ Language Usage Notes........................................................................................549 Process Identification Numbers.......................................
Figures 1 Operational Overview of the Measure Performance Monitor..................................................29 Tables 1 2 3 4 5 6 C-1 MEASCOM Commands ...................................................................................................35 Measure Entity Types .....................................................................................................136 ZMS Style Format Error Field Values ................................................................................
About This Document This manual describes the command and programmatic interfaces of the Measure performance monitor. Supported Release Version Updates (RVUs) This manual supports J06.03 and all subsequent J-series RVUs, H06.03 and all subsequent H-series RVUs, G06.03 and all subsequent G-series RVUs, and D40.00 and all subsequent D-series RVUs, until otherwise indicated by its replacement publication.
New and Changed Information for H06.26/J06.15 RVUs (523324-018) • Added a usage note to Usage Notes for H-Series and J-Series PVUs and RVUs (page 121) for START MEASSUBSYS to indicate that you can configure the Measure subsystem to emit EMS warning event 4063 if the per-CPU CID usage exceeds a configured threshold. • Revised Max Value Counters (page 140) to improve clarity. • Corrected Units Displayed values for Response Time Counters (page 141) and Syslink Counters (page 143).
• • ◦ AF-UNIX-SOCKET-OPENS-LIMIT ◦ DIR-OPENS ◦ DIR-OPENS-LIMIT ◦ DIR-STREAMS ◦ DIR-STREAMS-LIMIT ◦ DISK-CACHE-BUF-BYTES ◦ DISK-CACHE-BUF-BYTES-LIMIT ◦ DISK-OPENS ◦ DISK-OPENS-LIMIT ◦ NFS-OPENS ◦ NFS-OPENS-LIMIT ◦ OSS-OPENS ◦ OSS-OPENS-LIMIT ◦ PIPE-FIFO-BUF-BYTES ◦ PIPE-FIFO-BUF-BYTES-LIMIT ◦ PIPE-FIFO-OPENS ◦ PIPE-FIFO-OPENS-LIMIT ◦ PXS-BUF-BYTES ◦ PXS-BUF-BYTES-LIMIT ◦ PXS64-BUF-BYTES ◦ PXS64-BUF-BYTES-LIMIT ◦ TTY-OPENS ◦ TTY-OPENS-LIMIT Add usage note to Usa
Definitions for PROCESS Entities (page 299), and added their usage considerations to Usage Notes for H-Series and J-Series PROCESS Entities (page 310): 20 ◦ CURRENT-PRIORITY-END ◦ CURRENT-PRIORITY-START ◦ INITIAL-PRIORITY-END ◦ INITIAL-PRIORITY-START • Revised description of MEASMONCONTROL (page 411) to add PROCESS_CREATE_ error information.
◦ MEASFT 3262 (ERR_PARSE_SPLIT_TIME_WC): SPLIT, BY TIME * command not allowed. ◦ MEASFT 3301 (ERR_CANNOT_SPLIT_BY_TIME): Only H01 and later data files can be split by TIME, file version VNN. ◦ MEASFT 4016 (ERR_NOMEMORY_TIME128): Not enough memory for TIME128 array, bytes remaining = #nn. New and Changed Information for H06.24/J06.13 RVUs (523324-016) • Revised information about active counter measurements in LISTACTIVE entity-type (page 86), CPU (page 161), and PROCESS (page 291).
• Revised DISC (page 189) entity syntax and Usage Notes to indicate support for partition parameter and multiple names in the sac parameter for CLIM-attached devices. • Revised DEVICE^CLIM^DESC (page 457) to include new partition field and new PLPT flag: MEAS_MULTIPLE_CLIMS. • Revised text for error 3122 in Appendix A (page 495) to reference partition parameter. • Added the following new error messages to Appendix A (page 495): 3124, 3125, and 3126.
Document Organization This document is organized as follows: Chapter 1: Introduction to Measure Provides an overview of the Measure performance monitor Chapter 2: MEASCOM Commands Describes the Measure command interface Chapter 3: Entities and Counters Describes how to identify each system resource to be measured while using the command interface Chapter 4: Measure Callable Procedures Describes the Measure programmatic interface Chapter 5: Entity Descriptors Describes the entity descriptors used by the prog
Notation Conventions General Syntax Notation This list summarizes the notation conventions for syntax presentation in this manual. UPPERCASE LETTERS Uppercase letters indicate keywords and reserved words. Type these items exactly as shown. Items not enclosed in brackets are required. For example: MAXATTACH Italic Letters Italic letters, regardless of font, indicate variable items that you supply. Items not enclosed in brackets are required.
{ } Braces A group of items enclosed in braces is a list from which you are required to choose one item. The items in the list can be arranged either vertically, with aligned braces on each side of the list, or horizontally, enclosed in a pair of braces and separated by vertical lines. For example: LISTOPENS PROCESS { $appl-mgr-name } { $process-name } ALLOWSU { ON | OFF } | Vertical Line A vertical line separates alternatives in a horizontal list that is enclosed in brackets or braces.
!i and !o In procedure calls, the !i notation follows an input parameter (one that passes data to the called procedure); the !o notation follows an output parameter (one that returns data to the calling program). For example: CALL CHECKRESIZESEGMENT ( segment-id , error ) ; !i !o !i,o In procedure calls, the !i,o notation follows an input/output parameter (one that both passes data to the called procedure and returns data to the calling program).
A group of items enclosed in brackets is a list of all possible items that can be displayed, of which one or none might actually be displayed. The items in the list can be arranged either vertically, with aligned brackets on each side of the list, or horizontally, enclosed in a pair of brackets and separated by vertical lines.
ZCOM-TKN-OBJNAME token-type ZSPI-TYP-STRING. !r !o The !o notation following a token or field name indicates that the token or field is optional. For example: ZSPI-TKN-MANAGER token-type ZSPI-TYP-FNAME32. !o Related Information Refer to the Operator Messages Manual for a list of Measure operator messages.
1 Introduction to Measure The Measure performance monitor is a tool that collects performance statistics about system resources. It lets you gather data from system components, network components, and your own business applications. You can use the data to balance and tune your system, detect bottlenecks, balance workloads, and do capacity planning. Multiple users can run Measure sessions at the same time, and each user can configure and take measurements independently.
For a description of the command interface, see Chapter 2: MEASCOM Commands (page 35) and Chapter 3: Entities and Counters (page 136). • The programmatic interface is a set of callable procedures. By calling these procedures from your applications, you can access all Measure functions. For example, you can write an application that configures a measurement, starts the measurement, stops it, and writes the data to a file for analysis.
The first line starts MEASCOM, the Measure command interface. The plus sign (+) is the MEASCOM command prompt. The next two lines configure the measurement, instructing Measure to take measurements on CPUs 5 and 6. In these commands, CPU is the entity type, and 5 and 6 are entity specifications. Step 2.
Other output options exist as well. If you specified a measurement interval, you can examine the counter values taken at each interval, in addition to the final values. You can display uninterpreted values—that is, values that are not converted to percentages—and you can choose between a full report and an abbreviated report. For most entity types, you can also view counter values while the measurement is in progress. When a measurement is complete, you can display one or more counter values as a plot.
Example 1 Using INFO Command to Determine When an Active Measurement Will Write Interval Data $DATA PERF 31> MEASCOM MEASURE Performance Monitor - T9086J04 - (01FEB11) - \PROD1 (C) Copyright 1986-2010 Hewlett-Packard Development Company, L.P. 1+ STATUS MEASSUBSYS Number of Active (or Configured) Measurements = 3 $DATA.PERF.MDFGUEST $SYSTEM.MEASURE.DOFF3 $GUEST.USER1.
Example 2 Using LISTALL Command to Determine When an Active Measurement Will Write Interval Data $DATA PERF 32> TIME April 13, 2011 14:23:45 $DATAS PERF 33> meascom MEASURE Performance Monitor - T9086J04 - (01FEB11) - \PROD (C) Copyright 1986-2010 Hewlett-Packard Development Company, L.P. 1+ ADD $GUEST.USER1.
2 MEASCOM Commands This chapter explains how to use the Measure command interface, MEASCOM, and describes each MEASCOM command. For more information, see General Information About Using MEASCOM (page 37). Summary of MEASCOM Commands Table 1 MEASCOM Commands Command Function Page START MEASSUBSYS Start the Measure subsystem. 120 STATUS MEASSUBSYS Display status of the Measure subsystem. 127 STOP MEASSUBSYS Stop the Measure subsystem. 130 ADD entity-type Select an entity to measure.
Table 1 MEASCOM Commands (continued) Command Function Page LISTENAME Translate a Guardian file name and creation version serial 97 number (CRVSN) to its corresponding external name (ANSI SQL name or OSS pathname). LISTGNAME Translate an OSS file pathname to its corresponding Guardian file name and creation version serial number (CRVSN). LISTPNAME Translate a Guardian file name and creation version serial 99 number (CRVSN) to its OSS file pathname equivalent.
Table 1 MEASCOM Commands (continued) Command Function Page VOLUME Select a default volume. 133 WARNINGS Selectively display or suppress warning messages. 133 ! (Exclamatory Symbol) Reexecute a command. 134 General Information About Using MEASCOM Syntax Conventions for MEASCOM Commands Multiple Commands on a Line A MEASCOM command line can contain more than one command. Use a semicolon (;) to separate multiple commands.
You can use an asterisk as a wildcard to replace an element of the file name. For example: • $SR8.QUAL.* refers to all files in the subvolume QUAL on the volume $SR8. • $*.QUAL.STATUS refers to files named STATUS in subvolume QUAL on any volume. • $SR8.*.* refers to all subvolumes and all files on volume $SR8. • $*.*.* refers to all files in all subvolumes on all volumes. You cannot use an asterisk to represent only part of a file-name element.
• You cannot use abbreviations in a command (OBEY) file or in an input file used during noninteractive command entry. Because abbreviations might change from RVU to RVU, this restriction prevents command files from becoming obsolete. • You cannot abbreviate the topic in a HELP command. For example, you must enter HELP ADD TERMINAL, not HELP ADD TER.
35> MEASCOM /IN $MEAS.CMDS.MCONFIG / MCONFIG is a file containing these commands: ADD CPU * ADD DISC * START MEASUREMENT $DATA.PERF.MAY04 For the full syntax of the MEASCOM command, see MEASCOM (page 40). Creating a Custom Startup File You can create a file of MEASCOM commands, similar to a command (OBEY) file, to execute each time MEASCOM is invoked. The file must be an edit file, named MEASCSTM, that contains only MEASCOM commands. MEASCOM searches for the MEASCSTM file on the current subvolume.
Related Command Command Function Page EXIT Ends a MEASCOM session 57 Usage Note If multiple calls are needed while attempting to retrieve data from MEASFH, MEASCOM checks for BREAK before each call. Example To start the Measure command interface, check the status of the Measure subsystem, and stop the command interface: 11> MEASCOM MEASURE Performance Monitor - T9086H06 - (01AUG11) - \SYSTEM (C) Copyright 1986-2010 Hewlett-Packard Development Company, L.P.
INFO entity-type Displays the entities in the configuration 61 START MEASUREMENT Starts a measurement after defining the configuration 123 Usage Note for All PVUs and RVUs For a description of how to include and exclude subsets of entities in a measurement, refer to the DELETE entity-type command under Examples (page 52). Usage Notes for H-Series and J-Series PVUs and RVUs • In H06.23/J06.
• To add all processes running a specific object file to a measurement and collect procedure (code-range) execution data: + ADD PROCESSH $SYSTEM.SYS00.TCPIP ($SYSTEM.SYS00.TCPIP) • To add measurement of specific code ranges and provide recognizable names via an EDIT (code 101) file, either ranges or range offsets can be specified. This example shows use of offsets to define code ranges: OBJECT $a.b.
counter-type is the type of counter to create.
2. • Add the counter to the configuration by using the ADD COUNTER command. Identify each process that modifies the counter by specifying an entity-spec exactly as specified in the ADD USERDEF command. The record for user-defined counters can be up to 1000 words long. Each process can maintain up to 85 named user-defined counters, and each counter can be an array of up to 128 counter values, as long as the 1000-word limit is not exceeded.
data-file is a disk file (local or remote) containing measurement data. You can specify a currently active data file. MEASFH measfh specifies a MEASFH process other than the default for your system. measfh is an object file name. Specify an alternate MEASFH process when you: • Need to process a measurement data file created by an earlier product version of the Measure subsystem. MEASFH and measurement data files are release dependent. You must use D40 MEASFH with a D40 data file, and so on.
G-series version of MEASFH on the system where you ran MEASFT, using the MEASFH parameter to specify the object file. For information about the MEASFT utility, see Appendix D: Measure Data File Tool (MEASFT) (page 550). Usage Notes for G-Series PVUs and RVUs • Prior to Measure G05, a limit of 127.5 MB existed on the size of the measurement data file that could be opened for analysis. With Measure G05 and later PVUs, the Measure data file size is 890 MB.
record-number is the ordinal number of a single record to be plotted (1 through 200). For the PROCESSH entity type, the syntax is: ADD [ PLOT ] counter [ ( code-space ) ] PLOT is optional if you specified PLOT as the default command object by using the ASSUME command. counter is one of: procedure-name Adds an individual procedure to the plot. CODE-RANGE Adds all code samples. Measurements for accelerated, unaccelerated, and native mode object-code execution are combined in one plot item.
In H-series RVUs, no code-space specification applies. Any value you specify will be accepted but ignored. For the USERDEF entity type, the syntax is: ADD [ PLOT ] counter [ ( array-index ) ] PLOT is optional if you specified PLOT as the default command object by using the ASSUME command. counter is the name of a user-defined counter. array-index is the index number of an item in an array (0 through 127).
• This example lists the OSSCPU records by ascending order of the C0-LS-SENDS counter, where C0 is the CPU number of the other CPU involved in the socket operation. + LIST OSSCPU *, BY C0-LS-SENDS ( ASCENDING ) • This example lists the OSSCPU record for CPU 1 and then shows a plot of the values for the C6-LS-AWAKES counter, where C6 is the CPU number of the other CPU involved in the socket operation.
COMMENTS The COMMENTS command specifies which comment messages are displayed. Comments are identified by the word COMMENT and a number in the range 2000 through 2999. Syntax COMMENTS { DISPLAY { ALL | comm-num [, comm-num ] ... } } { SUPPRESS { ALL | comm-num [, comm-num ] ... } } DISPLAY specifies which comment messages are to be displayed. ALL causes all comments to be displayed except those specified by one or more succeeding COMMENTS SUPPRESS commands. COMMENTS DISPLAY ALL is the default.
The entity-type keyword is optional if you specified it as the default command object by using the ASSUME command. entity-spec identifies the entity to be measured. Entity specifications are specific to each entity type. For the syntax, see the description of the specified entity type in Chapter 3: Entities and Counters (page 136).
+ + + + + + ADD ADD ADD ADD ADD ADD FILE FILE FILE FILE FILE FILE $DISC1.*.* $DISC2.*.* $DISC3.*.* $DISC4.*.* $DISC5.*.* $SYSTEM.SYS00.* NOTE: It is not possible to add a subset of an item that has been deleted. You CANNOT use a wildcard to specify all files, then delete the $SYSTEM volume and add back the SYS00 subvolume of the $SYSTEM volume (that is, the following example will NOT work: + ADD FILE $* + DELETE FILE $SYSTEM.*.* + ADD FILE $SYSTEM.SYS00.
DELETE MEASUREMENT The DELETE MEASUREMENT command deletes one or more measurement data files from the set of those accessible by MEASCOM. Syntax DELETE [ MEASUREMENT ] data-file [ , data-file ] ... MEASUREMENT is optional if you specified MEASUREMENT as the default command object by using the ASSUME command. data-file is the name of the measurement data file to be deleted. To indicate all accessible data files, use an asterisk (*).
char is the character previously designated for plotting the counter. Use an asterisk (*) to indicate all counters named counter. You must specify char if the current plot definition contains multiple counters that have the same name.
env-param is one of: ASSUME Displays the default command object COMMENTS Displays the current settings for the COMMENTS command LOG Displays the name of the default log file OSSPATH Displays the current default OSSPATH OUT Displays the name of the default output file PAGESIZE Displays the current PAGESIZE setting SETPROMPT Displays the current prompt text SQLCATALOG Displays the current default catalog value used to resolve partially qualified ANSI SQL names SQLSCHEMA Displays the current d
For example, if an ANSI SQL table name is specified as: ‘TABLE table_56’ MEASCOM will prepend the default catalog and schema, if set, and the fully qualified name will be (with the values above): ‘TABLE CATALOG_12.SCHEMA_34.table_56’ • In Measure G09 and later PVUs, to display the current default path for OSSPATH and the current setting for PAGESIZE, respectively: + ENV System \BUYER Volume $DATA.
Example This example shows the EXIT display for an abnormal voluntary termination: + EXIT CPU time 0:00:02.430 2: Process terminated with fatal errors or diagnostics FC The FC command operates like the TACL FC command. FC retrieves a command from the history buffer, displays it so you can modify it, then executes the modified command. For a complete description of FC, see the TACL Reference Manual.
HELP The HELP command displays quick-reference information about commands, objects, entities, and error messages. Syntax HELP [ / OUT filename / ] [ [ [ [ [ [ [ [ [ [ [ command-name command-name object command-name ENTITY command-name entity object entity counter entity COUNTERS error-number ABBREVIATIONS ALL ] ] ] ] ] ] ] ] ] ] ] OUT filename directs command output to filename. If filename does not exist, MEASCOM creates an edit file by that name and writes command output to it.
ALL lists all Measure keywords (commands, objects, and attributes) by category and provides a guide to more specific help topics. Issuing HELP with no parameters is the same as issuing HELP ALL. Usage Notes • You cannot abbreviate the help topic. For example, you must enter HELP START MEASUREMENT, not HELP START MEASU. • A double prompt (++) signals that more information is available for a topic. Press Return (or any other key) to see the additional information.
HISTORY The HISTORY command lists one or more of the most recently entered commands. Syntax HISTORY [ number ] number specifies the number (0 through 999) of history-buffer lines to display. If you do not specify a number, the last 10 lines in the history buffer appear. If the buffer contains fewer than the specified number of lines, HISTORY lists the existing lines.
DEVICE DISC LINE NETLINE PROCESS PROCESSH SYSTEM TERMINAL The entity-type keyword is optional if you specified it as the default command object by using the ASSUME command. Related Commands Command Function Page ADD entity-type Adds an entity to the configuration 41 DELETE entity-type Deletes an entity from the configuration 51 Usage Notes for All PVUs and RVUs • If the INFO display shows an entity as both added and deleted, the delete operation takes precedence.
+ ADD PROCESSH $SYSTEM.SYS00.TCPIP (ULR $SYSTEM.SYS00.ZCRTLSRL) + ADD PROCESSH $SYSTEM.SYS00.TCPIP (SCR $SYSTEM.SYS00.TSYSCLR) + ADD PROCESSH $SYSTEM.SYS00.TCPIP (SLR $SYSTEM.SYS00.TSYSCLR) + INFO PROCESSH Add Processh $SYSTEM.SYS00.TCPIP ($SYSTEM.SYS00.TCPIP) Add Processh $SYSTEM.SYS00.TCPIP ($SYSTEM.SYS00.ZCRESRL) Add Processh $SYSTEM.SYS00.TCPIP ($SYSTEM.SYS00.ZINETSRL) Add Processh $SYSTEM.SYS00.TCPIP ($SYSTEM.SYS00.ZLANCSRL) Add Processh $SYSTEM.SYS00.TCPIP ($SYSTEM.SYS00.
+ INFO COUNTER * Add counter TRANSACTIONS, process $PERF.QUOTAS.BILLING, accum Add counter WAITING, process $PERF.QUOTAS.BILLING, busy Add counter TRANSACTIONS, process $PERF.QUOTAS.INCOMING, accum INFO MEASUREMENT The INFO MEASUREMENT command displays the following information for one or more of the measurement data files currently accessible by MEASCOM: • File name, displayed as an ADD MEASUREMENT command. • Start time, stop time, and the collection interval of the measurement.
Related Command Command Function Page ADD MEASUREMENT Makes a data file accessible by MEASCOM 45 Usage Notes for All RVUs • In the INFO MEASUREMENT display, the name of the current data file (the file MEASCOM uses when generating reports) is followed by the comment “Current Data File.” To make a different file the current data file, use the ADD MEASUREMENT command. • The output of the INFO MEASUREMENT command is a list of the commands used to define the configuration.
if you requested one either explicitly, in the START MEASUREMENT command, or by making journal segment construction the default, as described in the usage notes for START MEASSUBSYS (page 120).) OSS journal segment states can be: ◦ OSS journal segment attached ◦ OSS journal segment under construction ◦ OSS journal segment error: nnnn Journal segment construction occurs after the measurement is stopped. When construction is complete, current openers of the file are not notified.
-- Delete Process system-processes -- Add CPU * -- Add Process "/bin" • This example shows an INFO MEASUREMENT report with an OSS file pathname translation error (indicated by "*OSSPath*"): + INFO MEASUREMENT * Add measurement $SPOOL.MEASDATA.MDATA -- Current Data File -Data collected from system \DEV, MEASURE release version G09.
Usage Note The output of INFO PLOT is a list of the commands used to add the counters to the plot definition. You can use /OUT filename/ to create a command (OBEY) file that stores a given plot definition. Example To display the current plot definition: + INFO PLOT * Add measurement $SPOOL.QUOTAS.
entity-spec identifies the entity to be measured. Entity specifications are specific to each entity type. For the syntax, see the description of the specified entity type in Chapter 3: Entities and Counters (page 136). list-option is one of: BY item-name [ ( ASCENDING ) | ( DESCENDING ) ] sorts the report in ascending or descending order. item-name is one of: { counter { identification-item } } where counter is a counter name. By default, counter items are sorted in descending order.
UNQUALIFIED displays procedure names in traditional form with no qualifiers. QUALIFIED displays procedure names with the Guardian or OSS object file name of the associated code. This feature can be useful in differentiating between like-named procedures in different object files. Guardian file names have their associated CRVSN appended to the end of the name. DOTS { ON | OFF } specifies whether report displays include connecting dots between labels and numeric values. Valid only if STYLE is ZMS.
{ [d]d mmm[ yyyy]} { mmm [d]d[ yyyy] } where dd is a day of the month, a number in the range 1 to 31. mmm is the first three letters of the month; for example, JAN, MAR, OCT. yyyy is the year. Valid years are 1984 through 2047. If you omit date, the start date of the measurement is used. start-time-of-day is the time the report window starts, in the format: hh:mm[:ss] where hh is the hour (0 through 23). mm is minutes (0 through 59). ss is seconds (0 through 59).
operation is one of: > (greater than) = (equal to) < (less than) <> (not equal to) value is a number in the range 0 through 2147482.999. In Measure G09 and later PVUs the range is 0 through 999999999999. If item-name is a time counter, value must be expressed in seconds. For example, 26 milliseconds must be entered as .026. LOADID loadid specifies the name to be placed in the loadid field of the records generated by this command. The LOADID option applies only to structured reports.
and the measurement interval is 30 minutes, the actual FROM time can be as early as 7:45. Similarly, if the specified TO time is 10:00, the actual TO time can be as late as 10:15. OFF applies the FROM and TO limits exactly as specified. TOTALS { INCLUDE | ONLY | SUPPRESS } determines whether TOTALS are displayed. The default is SUPPRESS. INCLUDE indicates that both the per-process and aggregated totals are displayed.
Related Commands Command Function Page LISTACTIVE Displays an active measurement 86 SET REPORT Specifies the report format 109 Usage Notes for All PVUs and RVUs • The FILE and PROCESS entities display the names of open files. A file can include resources whose names do not conform to file name syntax, such as a process. To display reports for these entities, specify the CPU, PIN, and file number of the resource rather than its file name. For an example, see Examples (page 75).
• In Measure H03/J01 and later PVUs, the LIST command allows SERVERNET, DEVICE, and DISC entity specifications using CLIMs. For the syntax and examples, see the SERVERNET (page 323), DEVICE (page 179) and DISC (page 189) entities. • In Measure H01 and later PVUs, the LIST command allows DISCOPEN, DISKFILE, FILE, and SQLSTMT entity specifications using ANSI SQL names. For the syntax, see the DISCOPEN (page 208), DISKFILE (page 217), FILE (page 227), or SQLSTMT (page 341) entities.
Ending-Free-Mem 4092 Ending-UCME ++ Cpu 6 VLX Initial Lock Pgs 2048 Mem Pages 8192 Memory MB 16 PCBs 100 Page Size 2048 Local System \BUYER From 22 Oct 1994, 12:20:47 For 118 Secs Cpu-Busy-Time Ending-Free-Mem ++ Cpu 7 VLX Memory MB 16 Local System \BUYER Cpu-Busy-Time Ending-Free-Mem • 18.42% Swaps 4092 Ending-UCME 0.27 0 Initial Lock Pgs 2048 Mem Pages 8192 PCBs 100 Page Size 2048 From 22 Oct 1994, 12:20:47 For 118 Secs 5% Swaps 4092 Ending-UCME 0.
Cpu-Busy-Time Ending-Free-Mem 36.29% Swaps 4092 Ending-UCME 0.42 0 + LIST CPU 1, FROM 12:30, TO 12:45 Cpu 1 VLX Initial Lock Pgs 2048 Mem Pages 8192 Memory MB 16 PCBs 256 Page Size 2048 Local System \BUYER From 5 Nov 1994, 12:30:24 For 15 Minutes Cpu-Busy-Time Ending-Free-Mem • 15.05% Swaps 4092 Ending-UCME 0.05 0 If you are examining a measurement data file from a remote system, you must use complete specifications when referencing files within the data file.
Local System \RAMBLER From 10 Sep 2004, 13:22:08 For 22.2 Seconds Total samples = 69 # Code-Map UC.0 Samples 1 # 1.45 % of total Code file $SYSTEM.SYS46.FUP:134451329318623 Percent Percent Code-Range Name Accel TNS of Code-Map of Total ------------------------------ ----------- ----------- ----------- ---------FILE^HAS^FORMAT1 1 # 100 % 1.45 Code-Map UC.1 Samples 2 # 2.90 % of total Code file $SYSTEM.SYS46.
tser__transfer() WAIT 3 WAIT() tip_avtrelease_flags 2 tip_avtrelease_flags() tip_avtsetup_data 2 tip_avtsetup_data() LOCKMEMORY 2 LOCKMEMORY() searchInitIndex__45NSKOrderedArr 1 NSKOrderedArrayOf::searchInitIndex(unsigned , unsigned int) const # 4.62 % 4.35 % # 3.08 % 2.90 % # 3.08 % 2.90 % # 3.08 % 2.90 % # 1.54 % 1.
MID has two subfields: PATHID is an internal format representation of an OSS file or SQL/MX file. For other files, the field contains zeros. CRVSN is a creation version serial number that identifies a unique instance of an OSS or SQL/MX file. For other files, the field contains zeros. FILE-SYSTEM-NAME The EXPAND system name of a file described by file-name if the file is located on a node other than that measured in the data file. Otherwise, this field contains spaces.
03 volume 03 subvol 03 filename 02 pathname-len 02 pathname * key is MID duplicates not end type character type character type character type binary 16 type character allowed 8. 8. 8. unsigned. 1024. DDL Record Description Fields For descriptions of the header fields used by all entities, see Common Entity Header Fields (page 144). MID A Measure identifier for an OSS file or SQL/MX file.
LIST PLOT The LIST PLOT command displays the currently designated plot, as specified by one or more ADD PLOT commands. In Measure G11 and later PVUs, ANSI SQL names appear in plots.
where hh is the hour (0 through 23). mm is minutes (0 through 59). ss is seconds (0 through 59). If you omit start-time-of-day, the start time of the measurement is used. INTERVAL interval specifies the time interval for the time line in the plot. interval is a time interval in one of these formats: n SECOND[S] n MINUTE[S] n HOUR[S] where n is an integer in the range 1 through 9999 and a multiple of the collection interval specified when the data was collected.
OFF for a two-axis plot, displays time on the horizontal axis. For a bar graph, displays the bars vertically (base on the horizontal axis). WIDE-ITEM { ON | OFF } sets the density of the plot. The default is OFF. ON for a two-axis plot, fills in the area between the time axis and the lowest counter value. For a bar graph, displays bars two characters wide to six characters wide depending on the number of counter values on the graph.
D: REQUESTS E: REQUESTS • Disc $SYSTEM-P (0, X, C1002531, 101) Disc $DISC4-P (1, X, C1002531, 107) To display a plot in the default format, assuming that the measurement data was collected using a collection interval of 4 seconds: + LIST CPU * . . (reports for CPUs appear here) . + ADD PLOT CPU-BUSY-TIME + LIST PLOT 14:28:19 14:28:23 14:28:27 14:28:31 14:28:35 14:28:39 14:28:43 14:28:47 14:28:51 14:28:55 14:28:59 14:29:03 14:29:07 0::::::+:::20.0::::::+:::40.0::::::+:::60.0 ...
0::::::+:::20.0::::::+:::40.0::::::+:::60.0:::::+:::80.0:::+::::100 A **************************+++++++++++++++++++| B **********************************++| C *******************************************************++++++++++++++++++++++++++++++++++| D +++++++++++++++| E + F +++++++++++++++++++++++++++++++++++++| 0::::::+:::20.0::::::+:::40.0::::::+:::60.0:::::+:::80.0:::+::::100 Min. value = 2.
A double prompt (++) at the bottom of the screen signals that another report follows the one currently on the screen. Press Return (or any other key) to view the next report. Press Ctrl-Y or type BREAK at the double prompt to ignore subsequent reports: • If you entered multiple commands on the previous command line, BREAK interrupts the current command and executes the next command. • If you did not enter multiple commands, BREAK interrupts the current command and returns the MEASCOM prompt.
OFF displays do not use dots to connect labels to formatted numbers. FOR duration for each field in the record, displays the difference between the immediate value and the value after duration. If FOR is omitted, the field values reflect activity since the measurement began. duration is a time interval in one of these formats: n SECOND[S] n MINUTE[S] n HOUR[S] where n is an integer in the range 1 through 9999. FORMAT { BRIEF | NORMAL | STRUCTURED } sets the format of the report. The default is NORMAL.
LEGACY displays and structured records use the legacy style. ZMS displays and structured records use the ZMS style. ZERO-REPORTS { INCLUDE | SUPPRESS } determines whether records containing all zero values are displayed. The default is SUPPRESS. INCLUDE displays entity reports even if all counter values are zero. SUPPRESS does not display entity reports if all counter values are zero. NOTE: Snapshot counters return a measurement value taken at a specific moment.
Devices connected by an FCSA or CLIM are dual fabric adapters, that is, they are connected to both the X and Y fabrics. For these devices, only "*" or "X" should be specified in the LIST command, and only "X" should be specified in the LISTACTIVE command. group (G-series and later RVUs) is the group number of the device to measure. (The group corresponds to the physical enclosure.) module (G-series and later RVUs) is the module number of the device to be measured.
group (G-series and later RVUs) is the group number of the disk to be measured. (The group number corresponds to the physical enclosure.) module (G-series and later RVUs) is the module number of the disk to be measured. slot (G-series and later RVUs) is the slot number of the disk to be measured. scsi-id (G-series and later RVUs) is the SCSI port identifier of the disk to be measured. Used for ServerNet/DA and FCSA disks, but not for CLIM disks.
LINE and NETLINE LISTACTIVE [ LINE | NETLINE ] [ WAN-spec [ , listactive-spec ] WAN-spec identifies the line to measure, specified as one of: { cpu ( line-name ) } { cpu ( trackid , clip , line ) } cpu is the number of the processors, in the range 0 through 15. The default is all CPUs. line-name is the name of the line, in the form $name7, where name7 is up to seven characters. trackid is the number of the track, in the form %Hhexnum6, where hexnum6 is up to six characters. The default is all track IDs.
cpu ( line-name.subdevice ) cpu is the number of the processor, in the range 0 through 15. line-name is the name of the line, specified as: { $name7 | $num | logical-Define-name }. subdevice is the name of the subdevice, specified as: { #name7 }. Related Commands Command Function Page LIST Reads counter values from a data file 68 SET REPORT Specifies a default report format 109 Usage Notes for All PVUs and RVUs • You can use the LISTACTIVE command to read the active records of DISKFILE entities.
partition number was specified for the measurement. See also the LISTACTIVE Entity Specification Special Cases subsection for DISC (page 90). • In Measure H04/J02 and later PVUs, the LISTACTIVE command creates a format 1 or format 2 structured file, depending upon the measurement data file size, if the report output file is a structured file. • In Measure H03/J01 and later PVUs, the LISTACTIVE command allows SERVERNET, DEVICE, and DISC entity specifications using CLIMs.
LISTALL entity-type The LISTALL entity-type command reads data from the current data file (the file most recently specified in an ADD MEASUREMENT command) and displays a report for each entity included in the entity-spec parameter. The LISTALL command is similar in function and syntax to the LIST entity-type command. However, LISTALL differs in that it lists each interval record within the specified measurement window. A double prompt (++) signals that another report follows the one on the screen.
TOTALS { INCLUDE | ONLY | SUPPRESS } ZERO-REPORTS { INCLUDE | SUPPRESS } ZERO-VALUES { INCLUDE | SUPPRESS } For more detailed syntax and descriptions of these list options, see the LIST entity-type (page 68) command.
• In Measure H01 and later PVUs, the LISTALL command allows DISCOPEN, DISKFILE, FILE, and SQLSTMT entity specifications using ANSI SQL names. For syntax, see the DISCOPEN (page 208), DISKFILE (page 217), FILE (page 227), and SQLSTMT (page 341) entities. • In H-series RVUs, IF clauses cannot refer to fields that appear only in legacy style records and reports, even in the legacy style report mode.
LISTGNAME The LISTGNAME command translates an OSS file pathname (in Measure G09 and later PVUs) or an ANSI SQL name (in Measure H01 and later PVUs) to its corresponding Guardian file name and creation version serial number (CRVSN). In the case of file name reuse, the CRVSN identifies a specific instance of a Guardian file name. If the OSS file pathname specified identifies a directory, the contents of the directory are listed.
Examples • This information appears when “pathname” specifies an OSS file: + LISTGNAME "/Path/to/the/file" $DATA01.ZYQ00001.Y00003D2:340359 • This information appears when “pathname” specifies an OSS directory: + LISTGNAME "/Path/to/the" Directory: Directory: $DATA01.ZYQ00001.Y00003D2:340359 $DATA01.ZYQ00001.Y00003D2:340223 $DATA01.ZYQ00001.
3023 WARNING. CRVSN was not specified, translation may be incorrect. 3114 OSS journal segment is required in this context and is not available for access. Examples This example shows the output from a LISTPNAME command where the CRVSN is not specified. In this case, a warning indicates a potentially incorrect OSS file pathname translation. + LISTPNAME $DATA01.ZYQ00001.Y00003D2 MEAS 3023 WARNING. Crvsn was not specified, OSS file pathname translation may not be correct. $DATA01.ZYQ00001.
OBEY The OBEY command reads MEASCOM commands from a specified file. Syntax OBEY filename [ , NOECHO ] filename is the name of a command file that contains MEASCOM commands. NOECHO prevents the command file from echoing its commands to an output file. Usage Notes • The terms command file and OBEY file are often used interchangeably. • You can nest up to four command files.
OUT The OUT command directs command output to a specified file. You can use the OUT command to create a command (OBEY) file based on MEASCOM command output. For example, the output of the INFO MEASUREMENT command can be used, with little editing, to rerun a measurement. Syntax OUT filename filename is the name of the listing file. If filename already exists, command output is appended to the file. If filename does not exist, MEASCOM creates an edit file by that name and sends command output to it.
Example To prevent data from scrolling out of a standard 24-line by 80 character display area: + PAGESIZE 24 RESET PLOT The RESET PLOT command resets one or more plot options to their default values. NOTE: The RESET PLOT command accepts attributes, but no values. To set a plot to a value other than the default, see SET PLOT (page 106). Syntax RESET [ PLOT ] { * | plot-attribute [ , plot-attribute ] ... } PLOT is optional if you specified PLOT as the default command object by using the ASSUME command.
RESET REPORT The RESET REPORT command sets one or more report options to default values. NOTE: The RESET REPORT command accepts attributes, but no values. To set a report to a value other than the default, see SET REPORT (page 109). Syntax RESET [ REPORT ] { * | report-attribute [ , report-attribute ] ... } REPORT is optional if you specified REPORT as the command object by using the ASSUME command.
Related Command Command Function Page SET REPORT Sets the report format options 109 RUN The RUN command lets you run another process without exiting from Measure. The MEASCOM RUN command is similar to the TACL RUN command. Syntax RUN [ / OUT out-file / ] run-file [ / options / ] [ "parameters" ] out-file is an output file for any error messages generated by the RUN command. run-file is the program file to be run. options are run options.
+ RUN FUP INFO MYSUBVOL.* SET PLOT The SET PLOT command sets one or more plot attributes that govern subsequent plots, until changed by a new SET PLOT or RESET PLOT command. Syntax SET [ PLOT ] plot-attribute value [ , plot-attribute value ] ... PLOT is optional if you specified it as the default command object by using the ASSUME command. plot-attribute value is one of: FOR duration is the duration of the plot window. FOR and TO are mutually exclusive.
number is a number in the range 0 through 999999999999.999. The default is 0. SCALE-TO number sets the upper boundary of the data-value axis. number is a number in the range 0 through 999999999999.999. The default is 100. TIME-BASE { ON | OFF } displays a bar graph or a two-axis plot of counter values. If a time interval was specified when the data was collected, the default is TIME-BASE ON. Otherwise, the default is TIME-BASE OFF. ON displays a plot of counter values over time.
ON for a two-axis plot, fills in the area between the time axis and the lowest counter value. For a bar graph, displays bars two to six characters wide, depending on the number of counter values on the graph. OFF for a two-axis plot, displays a single plot character for each counter value at each time interval. For a bar graph, displays bars one character wide.
5.45 5.08 4.72 4.36 3.99 3.63 3.27 2.90 2.54 2.18 1.82 1.45 1.09 .726 .363 0 ----| | | | - -----| | -| | | -| | | -| | | -| | | -| | |--------------------------A B Min Value = 1.80 A: READ-BUSY-TIME B: READ-BUSY-TIME Max Value = 2.92 Disk $BOOKS1 (6, 1, 17, 1) Disk $BOOKS1 (6, 1, 17, 2) SET REPORT The SET REPORT command governs the format of subsequent reports until changed by a new SET REPORT command, a RESET REPORT command, or the addition of a new data file.
CR-NAME-QUAL { UNQUALIFIED | QUALIFIED } specifies whether procedure (code-range) names are displayed with object file name qualifiers, if available, in the "Code-Range Name" column on the line immediately following the traditional line of code-range output. The default is UNQUALIFIED. UNQUALIFIED displays procedure names in traditional form with no qualifiers. QUALIFIED displays procedure names with the Guardian or OSS object file name of the associated code.
For structured records, if a preexisting structured file exists, the first record of the file is used to determine the template version, and the new records added to the structured file will use that template version. Note that it is actually possible for the data file to be a newer version than the installed system and for Measure to still determine the record format and write relevant counters in the correct places according to the “future” template version.
loadid is an alphanumeric string, 1 through 8 characters long. The string can contain letters, numbers, carets (^), hyphens (-), and underscores (_). The first character must be a letter. RATE { ON | OFF } determines how counter values are displayed. The RATE attribute has no effect on structured files. The default is ON. ON displays interpreted counter values (counts per second and percent busy). OFF displays uninterpreted counter values.
ONLY For all entities except PROCESSH and USERDEF, displays only the final TOTALS report. If only one entity report is generated, the TOTALS report is not displayed. For the PROCESSH entity, only aggregated PROCESSH data is displayed. NOTE: PROCESSH data can be collected on either a per-code-file basis or a per-procedure basis, depending on whether a code-file-spec was supplied when the PROCESSH measurement was configured (via the ADD PROCESSH command).
Examples • To display all CPU reports in the brief format, total the counter values across the reports, and produce a final report containing the totaled counter values: + SET REPORT FORMAT BRIEF, TOTALS INCLUDE + LIST CPU * Cpu 6 VLX Initial Lock Pgs 2048 Mem Pages 8192 Memory MB 16 PCBs 100 Page Size 2048 Local System \SAIID From 7 Aug 1994, 14:53:01 For 75 Seconds Cpu-Busy-Time Ending-Free-Mem ++ Cpu 7 VLX Memory MB 16 Local System \SAIID • 17.86% Swaps Ending-UCME 0.
Swaps 1,999 # Page-Requests 3,999 # MM-Page-Scans 3,999 # Page-Scans 14,995 # Unsp-Pages-Qtime 499,999 sec Mem-Qtime Unsp-Pages-Start 14 # Unsp-Pages-End 14 # Starting-SCL 29,999 # Ending-SCL 7,999 # Starting-UCL 25,999 # Ending-UCL 1,999 # Starting-Free-CIDs 31,999 # Ending-Free-CIDs 31,999 # Ending-UCME ------------ Device I/O ----------------------------------------------------Disc-IOs 4 # Cache-Hits 141 # Transactions Response-Time ------------ Messaging -------------------------------------------------
SETPROMPT The SETPROMPT command is similar to the TACL SETPROMPT command. SETPROMPT lets you redefine the MEASCOM input prompt to include environmental information, such as the current assumed object or the current log file. For more information on the TACL SETPROMPT command, see the TACL Reference Manual. Syntax SETPROMPT { { { { { { { { { ASSUME COMMENTS LOG NONE OUT SYSTEM SWAPVOL VOLUME WARNINGS } } } } } } } } } ASSUME displays the current assumed object.
+ SETPROMPT VOLUME $DATA MEAS • To set the MEASCOM prompt to display the current assumed object, then change the assumed object: + SETPROMPT ASSUME MEASUREMENT + ASSUME PLOT PLOT SHOW PLOT The SHOW PLOT command displays the setting of one or all plot attributes. Syntax SHOW [ / OUT filename / ] [ PLOT ] [ attribute ] OUT filename directs command output to filename. If filename does not exist, MEASCOM creates an EDIT file by that name and writes command output to it.
Example To display all plot attributes: + SHOW PLOT Set Plot Vert-Base on Set Plot Time-Base Set Plot Scale-From 0.000 Set Plot Scale-To 100.000 Set Plot Wide-Item off Set Plot From Set Plot To Set Plot For SHOW REPORT The SHOW REPORT command displays the setting of one or all report attributes. Syntax SHOW [ / OUT filename / ] [ REPORT ] [ attribute ] OUT filename directs command output to filename. If filename does not exist, MEASCOM creates an EDIT file by that name and writes command output to it.
Usage Note for G-Series PVUs In Measure G12 and later PVUs, code-space identifiers are displayed only for TNS code files (including accelerated TNS code files).
Example To specify a default schema for expanding partially qualified ANSI SQL names: + SQLSCHEMA ’SCHEMA_34’ START MEASSUBSYS The START MEASSUBSYS command starts the Measure subsystem. Syntax START [ MEASSUBSYS ] [ , cpu ] MEASSUBSYS is optional if you specified MEASSUBSYS as the command object by using the ASSUME command. cpu is the number of a CPU on your system. Specify cpu when the Measure subsystem is already running and you need to restart MEASCTL in a specific CPU.
These DEFINE statements do not impact the startup time for small measurements. • The CID table is an internal structure that determines how many entities can be measured. The table assigns one counter identifier (CID) to each measured entity. For example, if you request measurement of five different files, the CID table assigns one CID per file, no matter how many counters the measurement includes for each file. In Measure G05 and earlier PVUs, the CID table has an upper limit of 32,000 CIDs per processor.
that a warning be emitted when CID table usage exceeds 80 percent of the CID table size, enter: TACL> ADD DEFINE =_MEAS_EMIT_CID_USAGE, FILE WARN80 The DEFINE must be added prior to starting the Measure subsystem. The DEFINE is ignored for threshold values less than 01 or greater than 99. Warning event 4063 looks like this: 10-04-07 10:04:44 \MEASYOS.$XM00 TANDEM.MEASURE.H04 004063 Exceeded CID usage threshold.
START MEASUREMENT The START MEASUREMENT command specifies the file where the measurement data is to be written and starts the measurement. In Measure G09 and later PVUs, you can enable or disable OSS journal segment construction for the resolution of internal OSS path information in data records. In Measure H02 and later PVUs, you can use the FILESIZE and NOCOUNTERS options to select the measurement data file size and suppress counter data records.
hh:mm[:ss] where hh is the hour (0 through 23). mm is minutes (0 through 59). ss is seconds (0 through 59). If you omit start-time-of-day, the current time is used. FOR duration specifies the duration of the measurement. FOR and TO are mutually exclusive. If you omit both, you must use the STOP MEASUREMENT command to stop the measurement. duration is a time interval in one of these formats: n SECOND[S] n MINUTE[S] n HOUR[S] where n is an integer in the range 1 through 9999.
n SECOND[S] n MINUTE[S] n HOUR[S] where n is an integer in the range 1 through 9999. CAUTION: A collection interval increases the load on system resources and can affect performance during measurement. For information about collection intervals and performance, see Usage Notes for All PVUs and RVUs (page 126). OSS specifies that an OSS journal segment for OSS file pathname resolution is constructed as part of the measurement data.
Usage Notes for All PVUs and RVUs • If a system time change takes place during a measurement, the post processing may report incorrect results. You should consider stopping all measurements before any system time change takes place, for example, before using the TACL SETTIME command or the SETSYSTEMCLOCK procedure, in order to preserve measurement interval accuracy. A time change typically takes place at Daylight Savings Time (DST) transitions.
STATUS MEASSUBSYS The STATUS MEASSUBSYS command displays the following status information: • The number of active or configured measurements and a list of the measurement-data file names • The number of active MEASCTL processes and a list of the CPUs where they are active • In Measure G09 and later PVUs, the current setting of several subsystem environmental settings such as the number of CIDs supported, CID table locking, and the current default setting for OSS journal segment construction • In Meas
in CPU(s): 0, 1, 2, 3 Number of Active MEASIP Processes = 8 in CPU(s): 0, 1, 2, 3 OSS OFF, SQL OFF, Skew copy time OFF, Lock CID table ON, CIDs = 64K/CPU • In Measure H01 and later PVUs, to show information about the MEASIP processes: + STATUS MEASSUBSYS Number of Active (or Configured) Measurements = $SYSTEM.SYSTEM.ZASPAA $SYSTEM.SYSTEM.ZASPZOO $SYSTEM.MEASURE.
• Percent of the file used • In Measure G09 and later PVUs, the indication of whether OSS journal segment construction is ON or OFF • In Measure G11 and later PVUs, the indication of whether SQL journal segment construction is ON or OFF Syntax STATUS [ / OUT filename / ] [ MEASUREMENT ] data-file OUT filename directs command output to filename. If filename does not exist, MEASCOM creates an EDIT file by that name and writes command output to it.
Current EOF OSS Journal OFF • 8192 B SQL Journal OFF Maximum EOF Percentage Used 133693440 B 0.00 % In Measure D-series and later PVUs, to show information about the currently active measurement writing to OCT18: + STATUS MEASUREMENT OCT18 From 18 Oct 1994, 14:35:33 Cpu 1 Entities Disc 64 Entities Current EOF 178 B 46 Words 7552 Words Maximum EOF 1556480 B Percentage Used 0.3 % STOP MEASSUBSYS The STOP MEASSUBSYS command stops all measurements and the subsystem processes.
Syntax STOP [ MEASUREMENT ] data-file [ , NO ADD ] [ , { TO | AT } [ end-date , ] end-time-of-day ] MEASUREMENT is optional if you specified MEASUREMENT as the default command object by using the ASSUME command. data-file is an active measurement data file.
+ STOP MEASUREMENT $DATA.MEAS.MAY04 • To stop a measurement immediately but not start the MEASFH process; thus, the data is not available to the LIST command: + STOP MEASUREMENT $DATA.MEAS.MAY04, NO ADD SWAPVOL The SWAPVOL command specifies the swap volume where work files are created. The command applies for the remainder of the MEASCOM session or until another SWAPVOL command is executed. Syntax SWAPVOL [ $volume ] $volume is the name of a volume.
+ TIME 20 November 1994 18:02:25 • To display the time of the system \NY: + TIME \NY 20 November 1994 21:02:30 VOLUME The VOLUME command sets the default device name, subvolume name, or both for file-name expansions. Initially, the default device and subvolume are specified by the command interpreter. Syntax VOLUME [ $device ] [ [ $device.]subvolume ] $device is a device name. If you omit $device, the default device does not change.
ALL suppresses all warnings except those specified by one or more succeeding WARNINGS DISPLAY commands. Related Command Command Function Page COMMENTS Selectively displays or suppresses comment messages. 51 Example To suppress warning message 3013: + WARNINGS SUPPRESS 3013 ! (Exclamatory Symbol) The ! command retrieves a previously executed command from the history buffer and executes it again.
25+ ! SE • To reexecute the last command entered: 28+ ! ! (Exclamatory Symbol) 135
3 Entities and Counters This chapter describes the Measure entity types and their associated counters: Topic Page Counters Overview 137 Common Entity Header Fields 144 Measure Support for Open System Services (OSS) 147 Measure Support for ANSI SQL Names 149 Accessing ZMS Style Records (MEASDDLZ) 153 Entity Descriptions 154 Table 2 Measure Entity Types 136 Entity Type Measured System Resource Page CLUSTER Fiber Optic Extension (FOX) and ServerNet/FX cluster traffic.
Table 2 Measure Entity Types (continued) Entity Type Measured System Resource Page OSSNS OSS name server processes. 284 PROCESS Processes. 291 PROCESSH Process code range histograms. 313 SERVERNET ServerNet addressable controllers (SACs) and interprocessor communication (IPC) on HP NonStop S-series and NS-series servers. In G08 Measure, this entity applies to HP NonStop ServerNet Cluster and remote interprocessor communication (RIPC). 323 SQLPROC SQL/MP or SQL/MX processes.
Accumulating Counters Accumulating counters count the number of bytes passing to or from a file, disk, or other entity. A value is added to the counter whenever data is transferred to or from the entity. Attribute Setting Output Value Is... Units Displayed As... REPORT RATE is ON Number of bytes transferred for each second during the time interval. per-second values: /s, K/s, M/s, G/s, or T/s1 REPORT RATE is OFF Number of bytes transferred during the time interval.
Busy Counters Busy counters measure the time a resource is busy. The counter uses a busy or idle state word and a 64-bit busy-time accumulator to record the cumulative busy time in microseconds. Attribute Setting Output Value Is... Units Displayed As... REPORT RATE is ON. Percent of time busy during the interval. a percentage of the interval length REPORT RATE is OFF. Time busy during the interval. ms (milliseconds), sec, hr, day, or wk REPORT FORMAT is STRUCTURED.
Max Queue Counters (Legacy Style Only) Max queue counters track the maximum length of a queue. Report format options do not affect max queue counters. Due to the sharing of counter records among up to 64 measurement requests, there is no way to reset max queue values. As records exist in the system longer, the value of max queue counters diminishes. The reported value becomes irrelevant to the measured period. In Measure G11 and later PVUs, max queue counters are no longer tracked in the instrumentation.
Queue-Busy Counters Queue-busy counters measure the time during which a resource accessed through a queue is busy. They are used to measure device busy time in the ServerNet environment where multiple requests (I/Os) can be queued to a device at the same time. Attribute Setting Output Value Is... Units Displayed As... REPORT RATE is ON. Percent of time busy during the interval. a percentage of the interval length REPORT RATE is OFF.
Lockwait Counters Lockwait counters measure the time spent waiting for file locks. The lockwait counter is a special case of a response time counter where the transaction is a lock. Attribute Setting Output Value Is... Units Displayed As... REPORT RATE is ON. Average lock time for each transaction. ms, sec, hr, day, or wk REPORT RATE is OFF. Lock time. ms, sec, hr, day, or wk REPORT FORMAT is STRUCTURED. Lock time, expressed in microseconds.
Syslink Counters Syslink counters measure the time required to complete a message link to a remote system. Whenever a value is added to a syslink counter, a corresponding link counter is advanced. The syslink counter is a special case of a response time counter where the transaction is a message link. Attribute Setting Output Value Is... Units Displayed As... REPORT RATE is ON. Average link time for each transaction. ms, sec, hr, day, or wk REPORT RATE is OFF. Link time.
Table 3 ZMS Style Format Error Field Values (continued) Value Meaning .<2> Subsystem version mismatch. The Measure PVU that formatted the record is earlier than that of the instrumented product subsystem that produced this data record. There is more data in the internal record than MEASFH can format. Update the Measure PVU on the measured system. .<3> Template-version mismatch.
DDL Header (ZMS Style) NOTE: You should always use TEMPLATE-VERSION (page 147) to get the external records corresponding to a format that you can handle. There are three ways to do this: • Use MEASCOM, which uses the template version for the current version. • From a user program, call Measread_diff_() with the requested template version. • For structured files, use a pre-created structured file, with a record length corresponding to the requested template version for the entity in question.
02 loadid 02 load-id 03 prefix-id 03 interval-id 02 cpu-num type character 8. redefines loadid. type character 5. type character 3. type binary 16 unsigned. DDL Header Field Definitions Fields are listed alphabetically. CPU-NUM Number of the CPU on which the measurement was taken. For interprocessor communications (IPC) records, CPU-NUM is MYCPU.
LOADID Redefined. See LOAD-ID. OBJECT-UID (ZMS Style only) Reserved for future use. OS-VERSION Version ID of the operating system when the measurement was taken. This field is divided into two subfields: LETTER and NUMBER. SUBSYSTEM-VERSION (ZMS Style only) This value is set to 1 for the initial H-series and J-series versions of all subsystems. If the value is greater than 1, consult the subsystem documentation for information about changes that affect the Measure counters.
• OSS file pathnames can be long (up to 1023 characters in length). In displaying these long OSS file pathnames, Measure displays the first 80 characters and wraps any remaining characters to the next line. Measure displays the name in 80-character blocks until the entire OSS file pathname is displayed.
OSS Journal Segment The OSS journal segment is a set of OSS file pathname translations captured for Measure data collection. It provides OSS file pathname translations for data records in the Measure data file when such translations are no longer available from the OSS name server. The OSS name server cannot translate the OSS file pathname due to time constraints and the relocation of the data file to another node or network.
• Measure compresses any white space (one or more blanks) within single quotes that is not within double quotes. • Measure wraps displayed ANSI SQL names from one line to the next, as needed. If you cut and paste a wrapped ANSI SQL name from a Measure display, carriage return characters might be included that are not part of the name. To avoid this, cut and paste the name one line at a time.
translation must be available to reporting applications in structured Measure data output (see MEASLISTEXTNAMES (page 403)). Measure builds the SQL journal segment during the measurement and inserts it into the Measure data file during measurement shutdown. The journal segment is loaded into memory and attached to an application process or MEASCOM when a Measure data file is opened for analysis. This feature is optional. The default is no journal segment.
Aggregate Data Measurement In order to collect, analyze and display information about DLL execution, Measure now supports the collection and display of aggregate data about library code across all processes executing the same version of that code.
External Record Definitions • See the PROCESSH external record definition in DDL Record for PROCESSH Entities (ZMS Style) (page 316). • See the zmsproch-id. external record definition in DDL Record for PROCESSH Entities (ZMS Style) (page 316). Callable Procedures • MEAS_CODERANGENAME_DEMANGLE_ (page 382) Accessing ZMS Style Records (MEASDDLZ) The MEASDDLZ file lets applications access ZMS style data records with minimal impact.
CLUSTER The CLUSTER entity provides information about the number of FOX messages sent and received by all processes on the local system. In this manual, the term FOX refers to: Component Server FOX II Cyclone or VLX TorusNet NonStop K-series Servernet/FX NonStop S-series A FOX link involves at least two processes on different systems: the linker that initiated the link and the listener that accepted the link.
DDL Record for CLUSTER Entities (ZMS Style) The ZMS Style DDL record for CLUSTER entities is supported on Measure G11 and later PVUs. The fields included in BRIEF reports are in boldface type. ID Fields DDL Definition DEFINITION zmsclstr-id. 02 remote-system-name 02 remote-system 02 reserved-1 end type character 8. type binary 16 unsigned. type character 6. Counter Fields DDL Definitions DEFINITION zmsclstr-ctrs.
* F40 counter value items: 02 sent-bytes-f 02 returned-bytes-f 02 received-bytes-f 02 reply-bytes-f end type type type type binary binary binary binary 64. 64. 64. 64. Common ID Field DDL Definitions for CLUSTER Entities These ID fields are used in both styles of DDL records. Fields are listed alphabetically. REMOTE-SYSTEM System number of the remote system for which FOX traffic was measured. REMOTE-SYSTEM-NAME System name of the remote system for which FOX traffic was measured.
SENT-BYTES Number of FOX message bytes sent by all linker processes running on the local system. When a process sends or receives a FOX message, the MESSAGES-SENT or MESSAGES-RECEIVED counter of the PROCESS record for that process is also advanced. In D-series and G-series RVUs, this is a 32-bit counter. In G-series RVUs, SENT-BYTES-F field is a 64-bit version of SENT-BYTES. Counter type: Accumulating.
Usage Note for G-Series CLUSTER Entities The 64-bit byte-count fields (fields ending in -F) collect the same data as older 32-bit byte-count fields. For example, the 64-bit field SENT-BYTES-F collects the same data as the 32-bit field SENT-BYTES. The 64-bit fields are less subject to overflow caused by high levels of I/O activity. The 32-bit fields are still active and continue to return values. If no field overflow occurs, the 32-bit fields and the 64-bit fields return the same value.
* measures all controllers in all CPUs. type is the HP product number of the controller (such as 3128). The default is all controller types matching the specified cpu, channel, ctrl set. cpu is the number of the CPU in which the controller to be measured is configured. channel is the channel number of the controller to be measured. Use an asterisk (*) to indicate all channels. The default is all channels. ctrl is the controller number of the controller to be measured (0 through 31).
Legacy Style Only Counter Field DDL Definitions for CONTROLLER Entities These counter fields are used in Legacy Style DDL records only. Fields are listed alphabetically. IO-QLEN-MAX Maximum number of outstanding I/Os on the controller queue described by the IO-QTIME counter. Counter type: Max queue. IO-QTIME Time that the controller is busy performing I/Os. If the average number of I/Os is less than 1, the number can be interpreted as the controller busy percentage. Counter type: Queue.
CPU The CPU entity provides information about one or more CPUs in the local system. Topic Page Entity specification syntax for CPU entities 161 DDL records for CPU entities 161 DDL field definitions for CPU entities 165 Usage notes for CPU entities 176 Entity Specification Syntax for CPU Entities To describe CPU entities: CPU entity-spec [, entity-spec ] ... CPU collects information about one or more CPUs in the local system. entity-spec is specified as: { * | cpu } where: * indicates all CPUs.
02 02 02 02 02 02 02 02 02 02 02 02 02 02 02 02 02 02 02 02 02 02 02 02 02 02 02 02 02 02 02 02 02 process-ovhd disc-ios cache-hits transactions response-time comp-traps native-busy-time accel-busy-time tns-busy-time page-requests page-scans mm-page-scans Starting-free-mem Ending-free-mem Starting-UCME Ending-UCME Starting-UCL Ending-UCL Starting-SCL Ending-SCL Starting-Free-CIDs Ending-Free-CIDs unsp-pages-qtime unsp-pages-start unsp-pages-end link-prepush-msgs link-readlink-msgs link-large-msgs readlinkc
DDL Record for CPU Entities (Legacy Style) The Legacy Style DDL record for CPU entities will not change after the Measure G10 PVU. The fields included in BRIEF reports are in boldface type. The CPU DDL record for G-series entities is identical to the CPU record for D-series entities except: • Counters are added to support direct bulk I/O transfers. • The existing DISK-IOS counter also includes direct bulk I/O requests. • The SEND-BUSY-TIME field is no longer used.
02 cpu-subtype type binary 16 unsigned. * fields for TNS/R specific counters: 02 accel-busy-time type binary 64. 02 tns-busy-time type binary 64. 02 comp-traps type binary 32 unsigned. * Native Mode busy time 02 tnsr-busy-time type binary 64.
02 02 02 02 readlinkcache-ctrl readlinkcache-none replyctrlcache-msgs ending-free-cids 02 reserved type binary 64. type binary 64. type binary 64. type binary 32 unsigned. redefines ending-scl-lock type character 40. end Common and ZMS Style Only ID Field DDL Definitions for CPU Entities These ID fields are used in both styles of DDL records. Fields that are used in ZMS Style DDL records only are noted. Fields are listed alphabetically. CPU-SUBTYPE Binary value indicating the subtype for the CPU.
Common and ZMS Style Only Counter Field DDL Definitions for CPU Entities These counter fields are used in both styles of DDL records. Fields that are used in ZMS Style DDL records only are noted. Fields are listed alphabetically; subfields are listed in the order they occur in the field. ACCEL-BUSY-TIME Time that the CPU was busy executing accelerated code. This counter applies only to measurements taken on a TNS/R system or a TNS/E system. See Usage Notes for H-Series and J-Series CPU Entities (page 177).
CPU-QTIME Time (in microseconds) that processes spent on the ready list. In PVUs earlier than Measure G05, the ready list includes processes waiting on page faults. The CPU MEM-QTIME counter shows the time spent waiting for arrival of message system data. In Measure G05 and later PVUs, a process no longer remains active in the ready list timer, so the value of MEM-QTIME is zero. Counter type: Queue. DISC-IOS Number of I/O disk transfers performed by disk processes in the measured CPU.
INTR-BUSY-TIME Time that the CPU spent executing interrupt handlers. Most interrupts are caused by the processing of messages and I/O operations. The appropriate system description manual describes each type of interrupt. For TNS/E systems, this is the total time spend executing system Interrupt Processes (IPs) and system Auxiliary Processes (APs). Counter type: Busy. IPU (ZMS Style only) For NSMA systems in Measure J01 and later PVUs, IPU-specific counters.
This counter is valid and displayed by MEASCOM in ZMS style when the CPU subsystem version (ZMSCPU.HDR.SUBSYSTEM-VERSION), supplied by NonStop OS, is three or greater; otherwise the counter contains a zero value and is not displayed by MEASCOM. Counter type: Snapshot. MM-PAGE-SCANS (ZMS Style only) Number of memory page scans initiated by the memory manager to perform memory clean-up that was not caused by memory demand. The memory manager periodically scans memory for removable pages in the background.
READLINKCACHE-ALL Number of received messages for which both control and data bytes were cached in readlinkcache buffers. Counter type: Incrementing. READLINKCACHE-CTRL Number of received messages for which only control bytes were cached in readlinkcache buffers. Counter type: Incrementing. READLINKCACHE-NONE Number of received messages for which neither control nor data bytes were cached in readlinkcache buffers. Counter type: Incrementing.
STARTING-TIMER-CELLS (ZMS Style only) Number of instrumentation timer cells in use at the start of the measurement interval. Counter type: Snapshot. STARTING-UCL Number of user process code and library segment frames allocated in physical memory at the start of the measurement interval. This counter includes the number of frames allocated for user process TNS, accelerated, and native code segments for the short address spaces designated as user code (UC) and user library (UL). Counter type: Snapshot.
SWAPS Number of swap operations (both into and out of memory) performed by the memory manager. The memory manager swaps one or two pages at a time. The CPU SWAPS counter is advanced once for each swap regardless of the number of pages swapped. (The CPU SWAPS counter differs from the DISC SWAPS counter.) Only pages that are updated are swapped. Code pages are swapped into memory, but because they cannot be updated, they are never swapped out. Data pages are swapped both into and out of memory.
Legacy Style Only ID Field DDL Definitions for CPU Entities These ID fields are used in Legacy Style DDL records only. Fields are listed alphabetically. LCBS (Legacy Style only) No longer used; returns 0. MEMORY-PAGES (Legacy Style only) Redefined. See MEM-MB. MEMORY-PAGES32 (Legacy Style only) Redefined. See MEM-FRAMES.
ENDING-UCL-LOCK (Legacy Style only) No longer used. ENDING-UDS (Legacy Style only) Number of user process data segment frames allocated in physical memory at the end of the measurement interval. With SPRs to T9050G05 and later PVUs, ENDING-UDS frames are included in the ENDING-UCL counter. Counter type: Snapshot. ENDING-UDS-LOCK (Legacy Style only) No longer used. MEM-QLEN-MAX (Legacy style only) Maximum number of items on the queue described by the MEM-QTIME counter. In Measure G05 and later PVUs, zero.
STARTING-SDS (Legacy Style only) Number of system process data segment frames allocated in physical memory at the start of the measurement interval. This counter includes the number of frames allocated for system process stack segments, extended data segments, and process file segments (PFS) as well as the CPU system global data segment (SG) and any other extended data segments or frames allocated for operating system use.
Usage Notes for All CPU Entities • Unlike counters for other entity types, CPU counters are allocated and initialized only when the CPU is loaded. The counters are not reinitialized at the start of each measurement. Thus, the CPU counter provides an overview of system performance over an extended period. CPU counters can direct you to potential problems that you can examine in greater detail using other entities.
Usage Notes for H-Series and J-Series CPU Entities • In H-series, J-series, and later RVUs, all byte count fields are 64-bit counters. The suffix -F no longer appears in counter names in ZMS format; however, counters with names ending in -F remain available to applications requesting data in legacy style format. • In Measure H06/J04 and later PVUs, ZMS style CPU and ZMS style PROCESS support the LOCKED-PAGES-END and LOCKED-PAGES-START counters.
corresponding 64-bit field returns the correct value, and the 32-bit field returns a value of -1. The ERROR field for the measured entity also returns -1 to indicate an overflow condition. Convert your applications to use the 64-bit fields; 32-bit fields might be deactivated in a future release. In MEASCOM commands and in command (OBEY) files, use the names of the 32-bit fields. For example, issue the command LIST CPU BY DISC-IOS, not LIST CPU BY DISC-IOS-F.
DEVICE The DEVICE entity provides information about devices such as tape drives or printers. In G-series and later RVUs, the DEVICE entity applies only to tape drives. NOTE: Some devices are identified by specific entity types: DISC for disks, LINE and NETLINE for communication lines, and TERMINAL for subdevices and asynchronous terminals.
Devices connected by an FCSA or CLIM are dual fabric adapters, that is, they are connected to both the X and Y fabrics. For these disks, only "*" or "X" should be specified in the LIST command, and only "X" should be specified in the LISTACTIVE command. group (G-series and later RVUs) is the group number of the device to measure. (The group corresponds to the physical enclosure.) To indicate all groups, use an asterisk (*). The default is all groups.
03 lun 03 partition 03 target-id 02 config-name 02 adapter-name 02 SAC-name end type type type type type type binary 16 binary 16 binary 16 character character character type type type type type type type type type type type type binary binary binary binary binary binary binary binary binary binary binary binary unsigned. unsigned. insigned. 64. 64. 64. Counter Fields DDL Definition DEFINITION zmsdev-ctrs.
RECORD device. FILE is "device" ENTRY-SEQUENCED. . . . (error, time items, and measurement identification items; see Common Entity Header Fields (page 144)) . . . * entity identification items: 02 pin type binary 16 unsigned. 02 device-name type character 8. 02 logical-device type binary 16 unsigned. 02 ctrl type binary 16 unsigned. 02 unit type binary 16 unsigned. 02 device-type type binary 16 unsigned. 02 device-subtype type binary 16 unsigned.
Common and ZMS Style Only ID DDL Definitions for DEVICE Entities These ID fields are used in both styles of DDL records. Fields that are used in ZMS Style DDL records only are noted. Fields are listed alphabetically; subfields are listed in the order they occur in the field. ADAPTER-NAME Logical name associated with the adapter on which the SAC resides. Logical names are maintained by the system configuration database. The name is null-filled and null-terminated.
PLPT (ZMS Style only in Measure H03 and later PVUs) The PLPT structure replaces the SCSI-ID field. More attributes were necessary to identify a CLIM device, so the SCSI-ID field (which had a range of 0 to 999) was reduced to 16 bits and became the PLPT.TARGET-ID field. This structure is used in descriptor and data records. The PLPT structure consists of five subfields: PLPT-FLAGS MEAS_CLIM_REL Identifies this field as a PLPT structure rather than a SCSI-ID field.
Common and ZMS Style Only Counter Field DDL Definitions for DEVICE Entities These counter fields are used in both styles of DDL records. Fields that are used in ZMS Style DDL records only are noted. Fields are listed alphabetically. DBIO-READS Number of read operations to the measured device that were enabled for direct bulk I/O to a requesting application processor. This counter is a subset of READS. Counter type: Incrementing.
READS Number of read operations (from device to memory) performed by the I/O process. In addition to programmatic read operations, internal operations (such as retries on I/O operations) also modify this counter. A read operation to a key-sequenced file can result in multiple reads because the index blocks must be read before the data blocks. Counter type: Incrementing. REQUESTS Number of requests received by the I/O process.
Legacy Style Only Counter Field DDL Definitions for DEVICE Entities These counter fields are used in Legacy Style DDL records only. Fields are listed alphabetically. INPUT-BYTES-F (G-series Legacy Style only) Same as INPUT-BYTES but accommodates larger values (64 bits rather than 32). Counter type: Accumulating. OUTPUT-BYTES-F (G-series Legacy Style only) Same as OUTPUT-BYTES but accommodates larger values (64 bits rather than 32). Counter type: Accumulating.
Usage Note for H-Series and J-Series DEVICE Entities • For CLIM-attached devices in Measure H03/J01 and later PVUs, the DEVICE entity specification is updated to include lun and path and to not use scsi-id. As with FCSA devices, this entity specification can be used in the ADD, DELETE, LIST, LISTACTIVE, and LISTALL commands. NOTE: Although the path field still appears in PLPT structure in the ZMS Style DDL DEVICE record syntax, in H06.23/J06.
DISC The DISC entity provides information about one or more disks on the local system. NOTE: The keyword DISK is interchangeable with DISC in any MEASCOM command. Topic Page Entity specification syntax for DISC entities 189 DDL records for DISC entities 191 DDL field definitions for DISC entities 194 Usage notes for DISC entities 206 Examples of DISC measurements 208 Entity Specification Syntax for DISC Entities To describe DISC entities: DISC entity-spec [, entity-spec ] ...
Disks connected by an FCSA or CLIM are dual fabric adapters, that is, they are connected to both the X and Y fabrics. For these disks, only "*" or "X" should be specified in the LIST command, and only "X" should be specified in the LISTACTIVE command. group (G-series and later RVUs) is the group number of the disk to be measured. (The group number corresponds to the physical enclosure.) To indicate all groups, use an asterisk (*). The default is all groups.
DDL Record for DISC Entities (ZMS Style) The ZMS Style DDL record for DISC entities is supported on Measure G11 and later PVUs. The fields included in BRIEF reports are in boldface type. ID Fields DDL Definition DEFINITION zmsdisc-id. 02 pin 02 device-type 02 device-subtype 02 servernet 02 device-name 02 logical-device 02 GMS.
03 03 03 03 03 03 03 03 03 03 03 03 block-size blocks-inuse-start blocks-inuse-end block-splits hits misses faults audit-buf-forces blks-dirty-qtime write-cleans write-dirtys write-misses type type type type type type type type type type type type binary binary binary binary binary binary binary binary binary binary binary binary 32 unsigned. 32 unsigned. 32 unsigned. 64. 64. 64. 64. 64. 64. 64. 64. 64. end DDL Record Description Fields RECORD zmsdisc. FILE is "zmsdisc" ENTRY-SEQUENCED.
02 disc-process-type * counter value items: 02 request-qtime 02 request-qlen-max 02 requests 02 read-busy-time 02 write-busy-time 02 seek-busy-time 02 reads 02 writes 02 seeks 02 input-bytes 02 output-bytes 02 swaps 02 cblks-inuse-qtime 02 cblks-inuse-max 02 ablks-inuse-qtime 02 ablks-inuse-max 02 c 03 hits 03 misses 03 faults 03 audit-buf-forces 03 blks 03 blks-dirty-qtime 03 blks-dirty-max 02 control-points 02 control-point-writes 02 free-space-ios 02 requests-blocked type binary 16 unsigned.
02 02 02 02 02 02 02 read-qlen-max write-qbusy-time write-qtime write-qlen-max device-qbusy-time input-bytes-f output-bytes-f type type type type type type type binary binary binary binary binary binary binary 16 unsigned. 64. 64. 16 unsigned. 64. 64. 64. * New counters for G05 02 scsi-id type binary 64. * New counters in G05: 02 DBIO-reads 02 DBIO-writes type binary 32 unsigned. type binary 32 unsigned. * New counters in G07: 02 cn 03 blks 03 blks-inuse-start 03 blks-inuse-end occurs 4 times.
DISC-PROCESS-TYPE The disk process type, DP2. GMS Physical location address (group, module, slot). The GMS field is divided into three subfields: GROUP is the group number. MODULE is the module number. SLOT is the slot number. You can use the keywords GROUP, MODULE, and SLOT in the IF and BY clauses of the LIST and LISTALL commands. LOGICAL-DEVICE Logical device number of the device. PIN Process identification number of the disk process.
H03 PVU this value is stored in this smaller field in order to use the free upper 48 bits for lun and path. The valid values for target-id are 0 to 999. If target-id is in a descriptor, the wildcard is all 1s. SAC-NAME Logical name associated with the SAC used in this path to the physical volume. Logical names are maintained by the system configuration database. If the device is a CLIM device, SAC-NAME is the name of the CLIM(s). The default is all CLIMs. The name is null-filled and null-terminated.
Subfield Description Counter Type BLOCKS-INUSE-END (ZMS Style only) Number of cache blocks that were in use Snapshot at the end of the measurement or at measurement interval copy time. When TOTALS ONLY or TOTALS INCLUDE is used, this value is the total number of cache blocks in use at the end of the measurement or at measurement interval copy time. BLOCK-SPLITS (ZMS Style only) Number of block splits (per block size) for the volume.
CONTROL-POINT-WRITES Number of dirty cache blocks that the disk process wrote to disk at TMF control points. For a discussion of TMF auditing, see the TMF Operations and Recovery Guide. Counter type: Incrementing. CONTROL-POINTS Number of times the disk process wrote a TMF control point record to its audit-trail file. For a discussion of TMF auditing, see the TMF Operations and Recovery Guide. Counter type: Incrementing.
ENDING-FREE-SPACE Number of bytes free on the volume at the time indicated by the TO time. This counter is updated only for the primary path record. The RATE attribute has no effect on how this counter is displayed. The primary path can be obtained from the LIST DISC command (Measure H03 and later PVUs) or the SCF INFO DISCNAME, DETAIL command. Counter type: Snapshot. FREE-SPACE-IOS Number of I/O operations to the disk free space table. The disk free space table lists each available free block on the disk.
Because the I/O process modifies this counter before the I/O operation, if the write fails, the byte count might not be accurate. In D-series and G-series RVUs, this is a 32-bit counter. In H-series and J-series RVUs, this is a 64-bit counter. In G-series RVUs, the OUTPUT-BYTES-F field is a 64-bit version of OUTPUT-BYTES. Counter type: Accumulating. READ-QTIME Total time spent by read requests queued to this disk. Counter type: Queue.
This counter is updated only for the primary path record. The primary path can be obtained from the LIST DISC command (Measure H03 and later PVUs) or the SCF INFO DISCNAME, DETAIL command. Counter type: Snapshot. STARTING-FREE-SPACE Number of bytes free on the volume when the measurement starts. This counter is displayed by MEASCOM when the REPORT STYLE is ZMS. It can be used in the IF and BY clauses as well as with the PLOT commands. This counter is updated only for the primary path record.
Legacy Style Only ID Field DDL Definitions for DISC Entities These ID fields are used in Legacy Style DDL records only. Fields are listed alphabetically. CHANNEL (Legacy Style only) In D-series RVUs, channel number of the device. In G-series and later RVUs, redefined. See SERVERNET. CTRL (Legacy Style only) In D-series RVUs, controller number of the device. In G-series RVUs, no longer used; returns zero. UNIT (Legacy Style only) In D-series RVUs, unit number of the device.
Subfield Description Counter Type BLKS Number of cache blocks allocated. When TOTALS ONLY or TOTALS INCLUDE is used, this value is the maximum number of cache blocks allocated. BLKS is not a counter but a value obtained from an internal disk process data structure. N.A. BLKS-DIRTY-QTIME Time that cache blocks spent dirty. A cache block is dirty Queue when a buffered write operation writes data to it. After the disk process writes the dirty cache block to disk, the block is clean.
In Measure output, each subfield provides four values, one for each size of DP2 cache block: Field Label Cache Block Size C0 512 bytes C1 1024 bytes C2 2048 bytes C3 4096 bytes CW (Legacy Style only) Cache write metrics. This field is divided into these subfields: Subfield Description Counter Type WRITE-DIRTYS Number of times a required block is found dirty in cache Accumulating for a write operation.
OUTPUT-BYTES-F (Legacy Style only) In G-series RVUs, same as OUTPUT-BYTES but accommodates larger values (64 bits rather than 32 bits). Counter type: Accumulating. READ-BUSY-TIME (Legacy Style only) In D-series RVUs, the time spent reading from the disk. This counter includes the time spent reading data and positioning the disk heads to read the data. In G-series RVUs, no longer used; returns zero. Counter type: Busy. READ-QLEN-MAX (Legacy Style only) Maximum number of read requests queued to this disk.
Usage Notes for All DISC Entities • The DISC counters track all disk activity, including both file activity and internal activities such as writing to the free space table, writing to the undo area, and writing file and volume labels. Use the DISC counters to compare and balance the workload of two or more disks. For information on I/O and caching, see the appropriate system description manual. For information on TMF auditing and audit trails written to disk, see the TMF Operations and Recovery Guide.
ServerNet environment, you can have active requests in the read queue and write queue at the same time. Therefore, some overlap can exist between the READ-QBUSY-TIME and WRITE-QBUSY-TIME counters, and adding their values does not produce an accurate busy-time measurement. • The 64-bit byte-count fields (fields ending in -F) collect the same data as older 32-bit byte-count fields. For example, the 64-bit field INPUT-BYTES-F collects the same data as the 32-bit field INPUT-BYTES.
Future RVUs might adjust the mixed workload design to let applications take full advantage of the increase in CPU and disk capacity. The adjustment in design might require priority adjustments to the mixed workload environment. The most recent changes of the mixed workload algorithm include adjustments to let you increase the speed of the query but cause only moderate increases in response time.
Entity Specification Syntax for DISCOPEN Entities To describe DISCOPEN entities: DISCOPEN entity-spec [,entity-spec] ... DISCOPEN collects information about physical access to one or more disk file opens. The measured file must be local. However, the processes accessing that file can be local, remote, or both. entity-spec is defined as: diskfile [ ( [ \system-name, ] cpu, pin )] where: diskfile is the file to be measured. Specify diskfile as one of the following: { { { { { { { * [[$device.]subvolume.
NOTE: OSS file pathnames are case-sensitive and must be specified within double quotation marks (" "). Valid OSS file pathnames can refer to specific files or to a set of files within a specific directory. If a directory is specified, only files in that directory are included. Files in directories subordinate to the specified directory are not included. catalog, schema, object, partition (Measure H01 and later PVUs) are ANSI SQL identifiers.
02 02 02 02 02 02 02 02 02 02 end requests-blocked requests lockwait-time max-lockwait-time lock-timeouts lock-bounces cache-write-cleans DBIO-input-calls DBIO-output-calls SQL-operation-time type type type type type type type type type type binary binary binary binary binary binary binary binary binary binary 64. 64. 64. 64. 64. 64. 64. 64. 64. 64. DDL Record Description Fields RECORD zmsdopen. FILE is "zmsdopen" ENTRY-SEQUENCED. 02 hdr type zmsheader. 02 ctr type zmsdopen-ctrs.
* new counter value items for D10: 02 requests type 02 lockwait-time type 02 max-lockwait-time type 02 lock-timeouts type 02 lock-bounces type 02 cache-write-cleans type * SMS changes: 02 device-name 02 storage-pool 02 dbio-input-calls 02 dbio-output-calls * New counters for G08: 02 SQL-operation-time binary binary binary binary binary binary 32 unsigned. 64. 64. 32 unsigned. 32 unsigned. 32 unsigned. type character 8. type character 8. type binary 32 unsigned. type binary 32 unsigned. type binary 64.
FILE-TYPE One of these values, which indicates the file type: 0 Unstructured 1 Relative 2 Entry-sequenced 3 Key-sequenced OCB-NUMBER Identification number of the open control block (OCB). OPENER-CPU CPU number of the process that opened the measured file. OPENER-PIN Process identification number of the process that opened the measured file. OPENER-SYSTEM-NAME Name of the system that opened the measured file.
a new file, the disk process reads the new blocks from cache (incrementing this counter) unless no spare blocks are in cache. The DRIVER-INPUT-CALLS counter counts all read operations, both cache hits and misses. Counter type: Incrementing. CACHE-WRITE-CLEANS Number of times a required block is found clean in cache for a write operation, which saves a disk read to bring the block in. Counter type: Incrementing.
LOCKWAIT-TIME Time spent waiting for locks. The REQUESTS-BLOCKED counter counts the number of blocked requests—that is, the number of times requests are queued to locks. When the REPORT RATE value is on, this counter returns the average wait time (LOCKWAIT-TIME divided by REQUESTS-BLOCKED). Counter type: Lockwait. MAX-LOCKWAIT-TIME Maximum wait time per lock. This number can be used as a guide to set the timeout value in the application. REPORT RATE ON has no effect on this counter value.
• Both DISCOPEN and DISKFILE measure physical access to a file. When deciding when to use DISCOPEN and when to use DISKFILE, consider: ◦ DISCOPEN creates a separate counter record for each file open. Opener processes are identified individually. If the same process opens a file more than once, multiple DISCOPEN records are created. ◦ DISKFILE measures disk file access per file. DISKFILE creates one counter record for each measured file and does not track individual opener processes.
Example The report output for the DISCOPEN entity includes the ANSI SQL name for files that have an ANSI SQL name. Here is an example: Disc File Open $DATA.ZSD12345.Z1234567 ANSI SQL Name ‘TABLE Catalog_12.Schema_45.Table_56 PARTITION Partition_78’ Device Name $DATA Pool File Type Key Sequenced Local CPU 0 Opener Process: System \HURTS CPU 0 Pin 285 Format Version: H01 Data Version: H01 Subsystem Version: 1 Local System \HURTS From 12 Jan 2003, 19:35:28 For 46.
{ { { { { { { * [[$device.]subvolume.]filename[:crvsn] [$device.]#tempfile “pathname” ‘{TABLE|INDEX} [[catalog.]schema.]object[ PARTITION partition]’ ‘SCHEMA [catalog.]schema’ ‘CATALOG {catalog|*}’ } } } } } } } where: * measures local and remote access of all disk files. $device is the name of the volume (device) that contains the file to be measured. To indicate all volumes, use an asterisk (*). The default is the current default volume.
The fields included in BRIEF reports are in boldface type. ID Fields DDL Definition DEFINITION zmsdfile-id. 02 FCB-number 02 file-type 02 file-code 02 format 02 file-name. 03 volume 03 subvol 03 filename 02 device-name 02 storage-pool 02 file-name-MID. 03 PATHID 03 CRVSN 02 reserved-1 end type type type type binary binary binary binary 16 16 16 16 type type type type type character character character character character unsigned. unsigned. unsigned. unsigned. 8. 8. 8. 8. 8. type character 24.
DDL Record for DISKFILE Entities (Legacy Style) The Legacy Style DDL record for DISKFILE entities will not change after the Measure G10 PVU. The fields included in BRIEF reports are in boldface type. The DDL record for G-series DISKFILE entities is identical to the record for D-series DISKFILE entities except that counters are added to support direct bulk I/O and SQL actions. (Existing I/O counters such as DRIVER-INPUT-CALLS also include direct-bulk I/O operations.
02 OSS-Block-Write-Bytes 02 OSS-Cache-Callbacks 02 OSS-Callback-Writes end type binary 64. type binary 32 unsigned. type binary 32 unsigned. Common ID Field DDL Definitions for DISKFILE Entities These ID fields are used in both styles of DDL records. Fields are listed alphabetically; subfields are listed in the order they occur in the field. DEVICE-NAME Disk device on which the measured file is located. For SMF files, this field provides the physical location that corresponds to the logical file name.
STORAGE-POOL Name of the SMF storage pool to which the measured disk is assigned (process name of the storage pool process). Common and ZMS Style Only Counter Field DDL Definitions for DISKFILE Entities These counter fields are used in both styles of DDL records. Fields that are used in ZMS Style DDL records only are noted. Fields are listed alphabetically. BLOCK-SPLITS Number of block splits during writes to the file. Counter type: Incrementing.
ENDING-EOF EOF value of the file, in bytes, at measurement stop time or measurement interval copy time. The REPORT RATE value does not affect the display of this value in MEASCOM. This item is displayed in its absolute value in bytes, followed by the ratio of ENDING-EOF over STARTING-EOF. Counter type: Snapshot. ENDING-ROWS Number of rows in the SQL table at the end of a measurement interval. This counter is valid only if the table had zero rows at some time since the file open. Counter type: Snapshot.
OSS-BLOCK-WRITES Number of block write operations issued by DP2 for the OSS file system. Counter type: Incrementing. OSS-CACHE-CALLBACKS Number of cache callback requests sent by DP2 to the OSS file manager. Counter type: Incrementing. OSS-CALLBACK-WRITES Number of OSS-BLOCK-WRITES received by DP2 as a result of OSS-CACHE-CALLBACKS. Counter type: Incrementing. REQUESTS Number of requests for this file. Counter type: Incrementing.
Legacy Style Only Counter Field DDL Definition for DISKFILE Entities This counter field is used in Legacy Style DDL records only. OPEN-QLEN-MAX (Legacy Style only) Maximum number of opens on this file. Counter type: Max queue. Usage Notes for All DISKFILE Entities • Both DISCOPEN and DISKFILE measure physical access to a file. When you decide when to use DISCOPEN and when to use DISKFILE, consider: ◦ DISCOPEN creates a separate counter record for each file open.
• ◦ A file’s extent size exceeds 65535 pages. ◦ A format 2 file is explicitly requested at creation. In Measure G09 and later PVUs, the DISKFILE entity supports the use of OSS file pathnames as follows: ◦ The directory information in the OSS PATHID is ignored for DISKFILE usage of OSS file pathnames. ◦ Only one DISKFILE record exists for a file despite the number of accesses made. ◦ Any valid pathname that leads to the file can be used.
FILE The FILE entity measures the I/O operations performed by a user process on an explicitly opened file (logical file access). In Measure G09 and later PVUs, the FILE entity supports the use of OSS file pathnames in place of Guardian file names. It also displays OSS file pathnames in FILE reports and provides direct mapping between external structured records and OSSNAMES structured record output.
Record Creation OSS files are often opened within a parent process environment but accessed only in the context of a child process that inherits the open. Similarly, a child process might not actually use all file opens inherited from a parent process. To minimize the system impact of tracking OSS file opens in Measure, counter records for OSS file opens are initiated on first significant use of the file, not on the open of the file.
sockets. The actual process name (the one with the nn) is still shown as the Device Name in the LIST FILE display. Below is a sample LIST FILE display showing a bound socket: > meascom;add ossfidfb;list file * (osssocket) File Open $ZLSNN.Z00000.Z0001SQG:106899423347 Open Type OSSUNIXSTREAM OSSPath: "/tmp/kenm/mysock" Device Name $ZLS01 Device Type 0 (Process) Subdevice Type 0 Opener 1,576 ($Z01H) File Num 2 OSSPID: 64356356 Program $OSS2.ZYQ00000.Z000161Y:929432249 OSSPath: "/tmp/kenm/server.
{ “address port” { “address *” { “* port” } } } For more information about IP field values, see IP (page 237). Examples of use are in Command Examples: OSS File Opens (page 244). ANSI SQL Naming Conventions In Measure G11 and later PVUs, the FILE entity can provide ANSI SQL names in place of Guardian file names in its output. For complete details on SQL naming, see the SQL/MX Reference Manual. Entity Specification Syntax for FILE Entities To describe FILE entities: FILE entity-spec [,entity-spec] ...
filename is the name of the file to be measured. To indicate all files (except temporary files, which are specified separately), use an asterisk (*). :crvsn (Measure G09 and later PVUs) is the timestamp, creation version serial number (CRVSN), or file name extension necessary to form a unique file name. Use this option to guarantee file name uniqueness. The CRVSN is available from the Measure report and the LISTGNAME command. #tempfile is the file identification number of a temporary file to be measured.
"pname" (Measure G09 and later PVUs) is the OSS file pathname of the program file name of the process. An OSS file pathname that begins with /G is not supported, use the Guardian file name instead. file^open^type (Measure G10 and later PVUs) selects one, all, or several file open types for measurement. By default, only Guardian file opens are measured.
DDL Record for FILE Entities (ZMS Style) The ZMS Style DDL record for FILE entities is supported in Measure G11 and later PVUs. The fields included in BRIEF reports are in boldface type. ID Fields DDL Definition DEFINITION zmsfile-id. 02 opener-pin 02 file-number 02 file-type 02 device-type 02 file-name. 03 volume 03 subvol 03 filename 02 file-system-name 02 opener-processname 02 opener-program-filename. 03 volume 03 subvol 03 filename 02 device-name 02 opener-device-name 02 file-name-MID.
02 02 02 02 02 02 end OSS-cache-read-bytes OSS-cache-write-bytes OSS-block-reads OSS-block-read-bytes OSS-flow-controls misc-calls type type type type type type binary binary binary binary binary binary 64. 64. 64. 64. 64. 64. DDL Record Description Fields RECORD zmsfile. FILE is "zmsfile" ENTRY-SEQUENCED. 02 hdr type zmsheader. 02 ctr type zmsfile-ctrs. 02 id type zmsfile-id.
02 lock-waits 02 timeouts-or-cancels 02 escalations type binary 32 unsigned. type binary 32 unsigned. type binary 32 unsigned. * new entity identification items for 02 opener-processname type 02 opener-program-filename. 03 volume type 03 subvol type 03 filename type * SMS changes: 02 device-name 02 opener-device-name * New F40 counter value items: 02 message-bytes-f * New counters in G05: 02 dbio-reads 02 dbio-writes 02 dbio-read-bytes 02 dbio-write-bytes D10: character 8. character 8. character 8.
DEVICE-TYPE One of these values, which indicates the type of device associated with the file: 0 Process 1 Operator 2 $RECEIVE 3 Disk 4 Tape drive 5 Printer 6 Terminal 7 Communication line 8 SCSI FILE-NAME Name of the measured file. The name can apply to either a NonStop file or an SMF file. For NonStop files, if a primary disk file has associated partitions or alternate key files, all activity is recorded under the primary file name.
%H003 KEYFILE %H008 PROCFILE %H081 SQLFILE %H101 OSSFIFO %H102 OSSPIPE %H201 OSSDISK %H411 OSSUNIXSTREAM %H412 OSSUNIXDGRAM %H421 OSSINETSTREAM %H422 OSSINETDGRAM FILE-SYSTEM-NAME Name of the system from the complete file specification. FILE-TYPE File type, from the access control block: 1 Relative file 2 Entry-sequenced file 3 Key-sequenced file 0 Any other type of file IP IP address value. IP has three subfields: FAMILY specifies whether you are using IPv4 or IPv6 addressing.
OPENER-DEVICE-NAME Disk device on which the opener program file is located. For SMF files, this field provides the physical location that corresponds to the logical file name. For NonStop files, this field is the same as the OPENER-PROGRAM-FILENAME VOLUME subfield. OPENER-OSSPID If the opener process runs in the OSS environment, this field contains the OSS Process ID. If this process is not an OSS process, this field contains zeros.
DBIO-WRITE-BYTES Number of bytes transferred by direct bulk I/O write operations. Counter type: Incrementing. DBIO-WRITES Number of direct bulk I/O write operations to the measured file. This counter is a subset of WRITES. Counter type: Incrementing. DELETES-OR-WRITEREADS Number of calls to the WRITEUPDATE or WRITEUPDATEUNLOCK procedure with a buffer length of 0 (delete operations) or the number of calls to the WRITEREAD procedure (writeread operations).
INFO-CALLS Number of calls to these procedures that resulted in a message being dispatched to a process (IOP or other) to obtain the information requested: FILEINFO FILE_GETINFOBYNAME_ DEVICEINFO FILERECINFO FILE_GETINFOLISTBYNAME_ DEVICEINFO2 FILE_GETINFO_ FILEGETINFO_LIST For OSS, the number of fstat(), fstatvfs(), fpathconf(), getsockname(), getpeername(), getsockopt(), and sockatmark() API calls for this file. This counter includes both user and system calls to these procedures.
MISC-CALLS For OSS, this counter totals operations that increase FILE-BUSY-TIME but are not READS, WRITES, or INFO-CALLS. APIs that fall in this category are ioctl(), fsync(), ftruncate(), lseek(), and fcntl(), and the socket APIs connect(), setsockopt(), accept(), and shutdown(). Counter type: Incrementing. OSS-BLOCK-READ-BYTES For OSS regular files only, the number of bytes read to cache as a result of OSS-BLOCK-READS. Counter type: Accumulating.
RECORDS-ACCESSED Number of records read by the disk process or file system to perform the operations. Number of records accessed is greater than or equal to the number of records used. The ratio of RECORDS-USED to RECORDS-ACCESSED indicates the selectivity of the statement. The value of this counter is zero for Enscribe files. Counter type: Incrementing. RECORDS-USED Number of records returned to the SQL executor on reads, inserts, writes, updates, and deletes.
Counter type: Accumulating Usage Notes for All FILE Entities • FILE counters are useful for detecting potential problems in database access. • The Measure subsystem tracks only local entities. Therefore, the FILE entity includes I/O operations performed by local user processes only. The files being accessed can be local, remote, or both, depending on the measurement configuration. • The FILE entity measures all partitions of a partitioned file together, as a single entity.
• A read or write operation on an OSS regular file can be cached or noncached: ◦ Noncached operations result in messages sent directly to DP2 and transfer only as much information as the client specifies. ◦ Cached operations utilize the file-system cache in the client’s CPU and can result in a block read/write between the cache and DP2. Block reads or writes occur on cache block boundaries (except for the last block of the file) where a cache block is 4 KB.
The Measure file open type selects OSS file opens of OSS regular files and Guardian files opened via /G: LIST FILE * (OSSDISK) LIST FILE $VOL.*.* (OSSDISK) OSS Opens of FIFOs OSS file opens of the same OSS FIFO are selected by the Guardian-format file name or OSS file pathname. A LISTGNAME command on a FIFO file pathname returns the Guardian device name in the form of $ZPPNN. Measure records display the file open name in the same format ($ZPPNN.Znnnnn.Ziiiiiii:CRVSN) for all FIFO opens.
OSS Opens of AF_UNIX Sockets Using socketpair() OSS opens of AF_UNIX sockets due to socketpair() share the same Guardian format file name but have no OSS file pathname. Both socket opens created by socketpair() share the same Guardian format file name, which can select the shared OSS socket opens: LIST FILE $ZPLS.*.* LIST FILE $ZLSnn.*.
LIST FILE *, if IP-ADDR = "1.2.3.4 37" IP address = 1.2.3.4, port 37 LIST FILE *, if IP-ADDR = "1070::800:2006:4178 37" IPv6 address, port 37 OSS Opens of OSS AF_INET and AF_INET6 Datagram Sockets Bound AF_INET or AF_INET6 datagram socket opens have unique Guardian format names and use the socket address as IP address. Unbound AF_INET or AF_INET6 datagram socket opens have unique Guardian format names but no IP address.
LINE The LINE entity provides information about one or more communication lines.
line (G-series) is 0 or 1 to indicate a specific line managed by a CLIP in a SWAN concentrator. To indicate all lines, use an asterisk (*). type (G-series) is a specific line type. (To list the valid values for type, use SCF LISTDEV.) Specifying type selects only lines of a particular protocol, such as type 61 (X25AM). To indicate all types, use an asterisk (*). The default is all types. subtype (G-series) provides further information about the line type.
02 id end type zmsline-id. DDL Record for LINE Entities (Legacy Style) The Legacy Style DDL record for LINE entities will not change after the Measure G10 PVU. The fields included in BRIEF reports are in boldface type. The DDL record for G-series LINE entities is identical to the record for D-series LINE entities except: • Different entity identification fields are used. The CTRL, UNIT, and CHANNEL fields are no longer used.
Common ID Field DDL Definitions for LINE Entities These ID fields are used in both styles of DDL records. Fields are listed alphabetically. CLIP Number of counter ID control blocks available for new counter records in this processor at the end of the reported measurement interval. DEVICE-NAME Device name of the measured line. DEVICE-SUBTYPE Additional identifier for DEVICE-TYPE. For a list of subtype values for communications lines, see the System Generation Manual.
INPUT-DATA-BYTES Number of data bytes read from the communications line. This counter includes only user data. Data required by line protocol is ignored. All INPUT-DATA-BYTES events are recorded on the write unit of the line. The INPUT-DATA-BYTES count is always less than the INPUT-BYTES count. For SNAX lines, this counter includes only request/response header (RH) and RU data counts (3 + RU bytes) of RUs destined for an OPEN logical unit or a logical unit in a pass-through session.
READS Number of read operations (from the communications line to memory) performed by the I/O process. Because the I/O process modifies this counter before the I/O operation, this counter includes both successful and unsuccessful operations. The RETRIES counter counts the number of I/O operations retries due to failures. Counter type: Incrementing. REQUESTS Number of requests received by the I/O process for the communications line.
Fields are listed alphabetically. CHANNEL (Legacy Style only) In D-series RVUs, channel number of the line. In G-series RVUs, not used; returns zero. CTRL (Legacy Style only) In D-series RVUs, controller number of the measured line. In G-series RVUs, not used; returns zero. UNIT (Legacy Style only) In D-series RVUs, unit number of the measured line. In G-series RVUs, not used; returns zero.
Usage Notes for G-Series LINE Entities • The 64-bit byte-count fields (fields ending in -F) collect the same data as older 32-bit byte-count fields. For example, the 64-bit field INPUT-BYTES-F collects the same data as the 32-bit field INPUT-BYTES. The 64-bit fields are less subject to overflow caused by high levels of I/O activity. The 32-bit fields are currently active and continue to return values. If no field overflow exists, the 32-bit fields and the 64-bit fields return the same value.
* measures all network communications lines in all CPUs. $device is the device name of the line to be measured. Use an asterisk (*) to indicate all network lines. cpu is the number of the CPU in which the measured line is configured. To indicate all CPUs, use an asterisk (*). The default is all CPUs. %Htrackid is the identifier of a specific 3880 SWAN concentrator enclosure of three CLIPs. (Each CLIP controls two lines.) To indicate all TRACKIDs, use an asterisk (*).
Counter Fields DDL Definition DEFINITION zmsnline-ctrs.
* counter value items: 02 write-busy-time type binary 64. 02 read-busy-time type binary 64. 02 requests type binary 32 unsigned. 02 reads type binary 32 unsigned. 02 writes type binary 32 unsigned. 02 l2in-bytes type binary 32 unsigned. 02 l2out-bytes type binary 32 unsigned. 02 din4-bytes type binary 32 unsigned. 02 dout4-bytes type binary 32 unsigned. 02 cin4-bytes type binary 32 unsigned. 02 cout4-bytes type binary 32 unsigned. 02 u64-bytes type binary 32 unsigned. 02 u128-bytes type binary 32 unsigned.
DEVICE-TYPE A value of 63, which indicates an Expand line. LINE Specific line within a CLIP on a SWAN controller: either 0 or 1. LOGICAL-DEVICE Logical device number of the measured line. PIN Process identification number of the primary process. TRACKID 6-byte identifier returned by the 3880 controller SEEPROM and labeled externally on the enclosure. Common Counter Field DDL Definitions for NETLINE Entities These counter fields are used in both styles of DDL records. Fields are listed alphabetically.
DOUT4-BYTES Number of Level 4 data bytes written by the network line handler. This counter includes only user data and, on D-series, message system protocol bytes. The data associated with line protocol is ignored. In G-series and earlier RVUs, this is a 32-bit counter. In H-series and J-series RVUs, this is a 64-bit counter. In G-series RVUs, the DOUT4-BYTES-F field is a 64-bit version of DOUT4-BYTES. Counter type: Accumulating. L2IN-BYTES Number of Level 2 bytes read by the network line handler.
Counter type: Incrementing. REQUESTS Sum of all SYSTEM LINKS handled by this path plus the sum of all SYSTEM SENT FORWARDs (PASSTHRU REQUESTS) handled by this path. LINKS is a message-level number. SENT, RECEIVED, SENT FORWARD, and RECEIVE FORWARD are packet counters. One message can be several packets. REQUESTS counts both message-level events (LINKS) and packet-level events (SENT FORWARD) in the same counter.
This counter is only measured when Expand over SWAN is in use. Counter type: Busy. WRITES Number of write operations (from memory to the network communications line) performed by the line-handler process. Counter type: Incrementing. Legacy Style Only ID Field DDL Definitions for NETLINE Entities These ID fields are used in Legacy Style DDL records only. Fields are listed alphabetically. CHANNEL (Legacy Style only) In D-series RVUs, channel number of the measured line.
L2IN-BYTES-F (G-series, Legacy Style only) Same as L2IN-BYTES but accommodates larger values (64 bits rather than 32). Counter type: Accumulating. L2OUT-BYTES-F (G-series, Legacy Style only) Same as L2OUT-BYTES but accommodates larger values (64 bits rather than 32). Counter type: Accumulating. Usage Notes for All NETLINE Entities • In general, the data collected by the Measure L2 counters is very close to the actual line byte counts.
The optical disk system is an enclosure (familiarly called a jukebox) that contains one or more disk drives and multiple disk cartridges stored in cells. Each cartridge has two sides, and each side is a volume. G-series, H-series, and J-series RVUs do not support optical disks. However, Measure and later PVUs include an OPDISK DDL record for future use.
The fields included in BRIEF reports are in boldface type. ID Fields DDL Definition DEFINITION zmsodisk-id. 02 pin 02 device-type 02 device-subtype 02 servernet 02 device-name 02 logical-device 02 GMS. 03 group 03 module 03 slot 02 SCSI-id 02 config-name 02 adapter-name 02 SAC-name 02 volume-name 02 cell 02 side 02 reserved-1 end type type type type type type binary 16 binary 16 binary 16 binary 16 character binary 32 unsigned. unsigned. unsigned. unsigned. 8. unsigned.
02 logical-device type binary 16 unsigned. 02 ctrl type binary 16 unsigned. 02 unit type binary 16 unsigned. 02 device-type type binary 16 unsigned. 02 device-subtype type binary 16 unsigned. 02 volume-name type character 8. 02 cell type binary 16 unsigned. 02 side type binary 16 unsigned. * counter value items: 02 request-qtime type binary 64. 02 request-qlen-max type binary 16 unsigned. 02 requests type binary 32 unsigned. 02 read-busy-time type binary 64. 02 write-busy-time type binary 64.
GMS (ZMS Style only) Physical location address (group, module, slot). The GMS field is divided into three subfields: GROUP is the group number. MODULE is the module number. SLOT is the slot number. You can use the keywords GROUP, MODULE, and SLOT in the IF and BY clauses of the LIST and LISTALL commands. LOGICAL-DEVICE Logical device number of the optical disk enclosure. PIN Process identification number of the optical disk process.
Counter type: Queue Busy. INPUT-BYTES Number of bytes read from the optical disk. In addition to programmatic read operations, internal operations such as reading file labels also modify this counter. Because the optical disk process modifies this counter before the I/O operation, this counter might not be accurate if the read fails. Counter type: Accumulating. OUTPUT-BYTES Number of bytes written to the optical disk.
Counter type: Queue Busy. WRITE-QTIME (ZMS Style only) Total time spent by write requests queued to this optical disk. Counter type: Queue. WRITES Number of write operations from memory to device performed by the optical disk process. In addition to programmatic write operations, internal operations such as writing volume labels also modify this counter. Counter type: Incrementing. Legacy Style Only ID Field DDL Definitions for OPDISK Entities These ID fields are used in Legacy Style DDL records only.
WRITE-BUSY-TIME (Legacy Style only) Time spent writing to the disk. This counter measures the time from the instruction that starts the write operation to the interrupt that ends the operation. This counter includes the time spent writing data and positioning the disk heads to write the data. Counter type: Busy. Usage Notes for All OPDISK Entities • OPDISK does not measure optical disk volumes that have had no activity even if the REPORT ZERO-REPORTS attribute is set to INCLUDE.
cpu is the CPU on which the OSS elements be measured exist. Use a comma-separated list to specify multiple CPU numbers. Use an asterisk (*) to indicate all CPUs. The default is all CPUs. DDL Record for OSSCPU Entities (ZMS Style) The ZMS Style DDL record for OSSCPU entities is supported in Measure G11 and later PVUs. Fields included in BRIEF reports are in boldface type. ID Fields DDL Definition DEFINITION zmsosscp-id.
02 02 02 02 02 02 02 02 02 02 02 02 02 02 02 02 02 02 02 02 02 AF-INET-socket-opens AF-INET-socket-opens-limit AF-INET-socket-buf-bytes AF-INET-socket-buf-bytes-limit AF-UNIX-socket-opens AF-UNIX-socket-opens-limit AF-UNIX-socket-buf-bytes AF-UNIX-socket-buf-bytes-limit pipe-FIFO-opens pipe-FIFO-opens-limit pipe-FIFO-buf-bytes pipe-FIFO-buf-bytes-limit TTY-opens TTY-opens-limit NFS-opens NFS-opens-limit dir-opens dir-opens-limit dir-streams dir-streams-limit LCL 03 PS-proxy-reads 03 PS-proxy-writes 03 PS-p
. . .
Common ID Field DDL Definitions for OSSCPU Entities These ID fields are used in both styles of DDL records. Fields that are used in ZMS Style DDL records only are noted. Fields are listed alphabetically. CACHE-BLOCK-SIZE Size of an OSS file-system cache block in bytes. CACHE-BLOCKS-ALLOWED Maximum number of blocks configured for the OSS file-system cache. PP-BLOCK-SIZE Size of an OSS pipe pool block in bytes. PP-BLOCKS-ALLOWED Maximum number of blocks allowed for the OSS pipe pool.
AF-UNIX-SOCKET-OPENS-LIMIT (ZMS Style only) In H06.25/J06.14 and later RVUs, limit of AF_UNIX socket opens on a CPU. Counter type: Snapshot. AF-UNIX-SOCKET-BUF-BYTES (ZMS Style only) In H06.25/J06.14 and later RVUs, current number of bytes allocated for AF_UNIX socket data buffers on a CPU. Counter type: Snapshot. AF-UNIX-SOCKET-BUF-BYTES-LIMIT (ZMS Style only) In H06.25/J06.14 and later RVUs, maximum size (in bytes) for AF_UNIX socket data buffers on a CPU. Counter type: Snapshot.
DISK-OPENS-LIMIT (ZMS Style only) In H06.25/J06.14 and later RVUs, soft limit of OSS opens of regular files on a CPU. See also Usage Notes for OSSCPU Entities (page 282). Counter type: Snapshot. FM-CACHE-INFOS Number of OSS file manager cache info requests received from DP2. The OSS file manager allocates PXS and disk cache segments and acts as an agent for DP2 and the NonStop memory manager for cache management requests. Counter type: Incrementing.
FS-CACHE-RD-READ-REQS Number of block read requests to DP2 from the OSS file-system cache to satisfy an application read API call. A block read logical operation can return multiple blocks. Counter type: Incrementing. FS-CACHE-VALID-QTIME Amount of time the OSS file-system cache blocks are valid. Valid cache for open files includes active cache blocks and inactive (unowned) file cache blocks for recently closed files.
FS-MAP-FAILURES Number of times the OSS file-system cache failed to obtain a map buffer for prefetching cache blocks. The map buffer is a cache I/O buffer that is used for I/O transfers with DP2. A map failure can be a symptom of performance degradation. To remedy the failure, move applications. Counter type: Incrementing. FS-PREFETCH-BLOCKS Number of blocks prefetched by the OSS file-system cache. Counter type: Incrementing. FS-PREFETCH-USED Number of prefetched blocks actually used by the application.
Subfield Description Counter Type LS-SEND-BYTES Number of bytes received in SEND requests from satellite Accumulating sockets in CPU n where n is the LCL array index. Socket APIs, write(), send(), sendto() and sendmsg(), update this counter when the client is using a satellite socket (the socket is mastered in a different CPU). (See Note under LS-SENDS counter.
OSS-OPENS (ZMS Style only) In H06.25/J06.14 and later RVUs, current number of OSS file opens of all file types (regular file opens, pipe/FIFO opens, directory opens, directory streams, TTY opens, NFS opens, AF_INET socket opens, and AF_UNIX socket opens) on a CPU. Counter type: Snapshot. OSS-OPENS-LIMIT (ZMS Style only) In H06.25/J06.
PXS-BUF-BYTES (ZMS Style only) In H06.25/J06.14 and later RVUs, current number of bytes allocated within the POSIX extended segment (PXS) from 32-bit addressable memory on a CPU. Counter type: Snapshot. PXS-BUF-BYTES-LIMIT (ZMS Style only) In H06.25/J06.14 and later RVUs, maximum size (in bytes) for the POSIX extended segment (PXS) that is allocated from 32-bit addressable memory on a CPU. Counter type: Snapshot. PXS-ENDING-BYTES Current number of bytes allocated within the POSIX extended segment (PXS).
These counters occur once for remote systems. TTY-OPENS (ZMS Style only) In H06.25/J06.14 and later RVUs, current number of TTY opens (opens of tty devices using the /G name and the /dev/tty name) on a CPU. Counter type: Snapshot. TTY-OPENS-LIMIT (ZMS Style only) In H06.25/J06.14 and later RVUs, limit of TTY opens (opens of tty devices using the /G name and the /dev/tty name) on a CPU. Counter type: Snapshot.
Measure and the ALL_OPENS resource limit in the OSS SCF interface represent the same value.) For more information on OSS environmental limits, see the Open System Services Management and Operations Guide. • A read or write operation on a regular file can be sent directly to DP2 or satisfied via the OSS file system cache. ◦ Noncached operations (FS-DIRECT-READS, FS-DIRECT-WRITES) result in messages to DP2 and transfer only as much information as specified by the client.
OSSNS The OSSNS entity provides information about the operation and performance of OSS name server processes. There is one record for the primary and backup processes configured for each OSS name server defined on a system. Any OSS operation involving an OSS file pathname uses the OSS name server. OSSNS entities can only be used on systems running Measure G10 and later PVUs.
subvolume is the subvolume in which the object file of the OSS name server process to be measured is located. To indicate all subvolumes, use an asterisk (*). filename is the name of the object file of the OSS name server process to be measured. To indicate all files (except temporary files), use an asterisk (*). DDL Record for OSSNS Entities (ZMS Style) The ZMS Style DDL record for OSSNS entities is supported in Measure G11 and later PVUs. The fields included in BRIEF reports are in boldface type.
DDL Record for OSSNS Entities (Legacy Style) The Legacy Style DDL record for OSSNS entities will not change after the Measure G10 PVU. The fields included in BRIEF reports are in boldface type. RECORD ossns. FILE is "ossns" ENTRY-SEQUENCED. . . . (error, time items, and measurement identification items; see Common Entity Header Fields (page 144)) . . . * Entity identification items: 02 pin type binary 16 unsigned. 02 process-name type character 8. 02 program-file-name. 03 volume type character 8.
LC-ENTRIES Configured number of link cache records available on an OSS name server. PIN Process identification number of the measured OSS name server process. PROCESS-NAME Name of the OSS name server process. PROGRAM-FILE-NAME Object file of the measured OSS name server process. The name can apply to either a NonStop file or an SMF file. PROGRAM-FILE-NAME has three subfields: VOLUME For NonStop files, is the device location of the physical file.
DP2-MSG-QTIME Accumulated service time spent in a thread waiting while DP2 messages were outstanding on an OSS name server. Counter type: Queue. GETTIME-REQS Number of gettime requests sent by the OSS name server to a pipe server or local server. Counter type: Incrementing. IC-DIRTY-QTIME Amount of time inode cache entries were dirty on an OSS name server. Counter type: Queue. IC-HITS Number of inode cache hits encountered by an OSS name server.
PS-MESSAGES Number of messages an OSS name server sent to pipe server processes. The pipe server runs in each CPU ($ZPPnn, where nn = CPU number) and provides pipe services for applications using OSS pipes and FIFOs. Counter type: Incrementing. PS-MSG-QTIME Amount of time pipe server messages from an OSS name server were outstanding. Counter type: Queue. RR-PROCESSED Number of name resolution requests processed by the OSS name server under measurement. Counter type: Incrementing.
LS-MSG-QLEN-MAX (Measure G10 PVU only) Maximum number of concurrent local server messages outstanding on an OSS name server since it began execution. In H-series and J-series RVUs, this counter has a value of 1. Counter type: Max queue. PS-MSG-QLEN-MAX (Measure G10 PVU only) Maximum number of concurrent pipe server messages outstanding on an OSS name server since it began execution. In H-series and J-series RVUs, this counter has a value of 1. Counter type: Max queue.
process has a PXLINK cache. The PXLINK cache not only keeps copies of PXLINK records, it also remembers unsuccessful name lookups. If the name is looked up again, it is not necessary to reread the PXLINK file to discover again that the name does not exist. PROCESS The PROCESS entity provides information about one or more processes on a local system. In Measure G09 and later PVUs, the PROCESS entity type handles OSS file pathnames.
$device is the volume (device) on which the object file of the process to be measured is located. Use an asterisk (*) to indicate all volumes. The default is the current default volume. subvolume is the subvolume in which the object file of the process to be measured is located. Use an asterisk (*) to indicate all subvolumes. filename is the name of the object file of an executing process to be measured. Use an asterisk (*) to indicate all files (except temporary files).
03 GMOM-pin 03 GMOM-jobid 02 GMOM-full-id 02 02 02 02 02 02 02 02 02 GMOM-sysname GMOM-process-name program-accelerated native-process system-process ipus reserved-2 hometerm-sysname hometerm-name. 03 device 03 subdevice 03 qualifier type binary 16 unsigned. type binary 16 unsigned. type binary 64 redefines GMOM. type character 8. type character 8. type binary 16 unsigned. type binary 16 unsigned. type binary 16 unsigned. type binary 16 unsigned. type character 2. type character 8. type character 8.
02 02 02 02 02 02 02 02 02 OSSNS-DD-calls OSSNS-requests OSSNS-message-bytes OSSNS-wait-time OSSNS-redirects launches launch-wait-time open-close-wait-time ipu-switches * New counters for H05: 02 ipu-num 02 ipu-num-prev * New counters for H06 02 initial-priority-start 02 initial-priority-end 02 current-priority-start 02 current-priority-end 02 locked-pages-start 02 locked-pages-end type type type type type type type type type binary binary binary binary binary binary binary binary binary 64. 64. 64.
03 subvol 03 filename 02 priority * counter value items: 02 cpu-busy-time 02 ready-time 02 mem-qtime 02 dispatches 02 page-faults 02 pres-pages-qtime 02 pres-pages-max 02 ext-segs-qtime 02 ext-segs-max 02 vsems 02 recv-qtime 02 recv-qlen-max 02 messages-sent 02 sent-bytes 02 returned-bytes 02 messages-received 02 received-bytes 02 reply-bytes 02 lcb-allocations 02 mqc-allocations type character 8. type character 8. type binary 16 unsigned. type binary 64. type binary 64. type binary 64.
02 UCL-qtime 02 UCL-max type binary 64. type binary 16 unsigned. 02 UCL-lock-qtime 02 UCL-lock-max type binary 64. type binary 16 unsigned. 02 file-open-calls 02 info-calls * D30 changes: 02 begin-trans 02 abort-trans type binary 32 unsigned. type binary 32 unsigned.
02 open-close-wait-time end type binary 64. For descriptions of the header fields used by all entities, see Common Entity Header Fields (page 144). Common and ZMS Style Only ID Field DDL Definitions for PROCESS Entities These ID fields are used in both styles of DDL records. Fields that are used in ZMS Style DDL records only are noted. Fields are listed alphabetically; subfields are listed in the order they occur in the field. ANCESTOR-CPU Number of the CPU on which the ancestor process resides.
GMOM-FULL-ID 64-bit redefinition of the individual GMOM fields as a single value. A value of zero indicates that the process is not part of a NetBatch job stream. GMOM-PROCESS-NAME Name of the GMOM process that initiated the NetBatch job that initiated this process; otherwise spaces. GMOM-SYSNAME Expand system name of the GMOM node if the GMOM process is remote; otherwise spaces.
PROGRAM-FILE-NAME Object file of the measured process. The name can apply to either a NonStop file or an SMF file. PROGRAM-FILE-NAME has three subfields: VOLUME For NonStop files, is the device location of the physical file. For SMF files, is the device location of the physical file that corresponds to the logical file name is stored in the DEVICE-NAME field. SUBVOLUME For NonStop files, is the device name of the physical volume on which the program file is located.
ABS-SEGS-START Number of items on the queue described by ABS-SEGS-QTIME at the start of the measurement interval. Counter type: Snapshot. ACCEL-BUSY-TIME Time that the CPU was busy executing accelerated code in this process. This counter applies only to measurements taken on a TNS/R or TNS/E system. See Usage Notes for H-Series and J-Series PROCESS Entities (page 310). Counter type: Busy. ALLOC-SEG-CALLS Number of successful calls made to the SEGMENT_ALLOCATE_ or ALLOCATESEGMENT procedures.
NOTE: The CPU-BUSY-TIME counter is enhanced for more precise measurement of processor utilization. Measurements on systems running D-series and G-series RVUs might show slightly higher utilization than comparable measurements on systems running C-series RVUs. The difference appears mainly on systems that run at low utilization levels. Related Counter Description Page CPU CPU-BUSY-TIME Similar, but is the busy time for the entire processor.
FILE_GETINFO_ fpathconf() select() FILE_GETINFOBYNAME_ fstat() sockatmark() FILE_GETINFOLISTBYNAME_ fstatvfs() stat() FILEGETINFO_LIST getpeername() statvfs() DEVICEINFO getsockname() Counter type: Incrementing. INITIAL-PRIORITY-END (ZMS Style only) In H06.25/J06.14 and later RVUs, the target (requested) process priority in effect at the end of the measurement interval.
LAUNCH-WAIT-TIME Amount of time the process waits on requests to PROCESS_CREATE_ or the fork()/exec*() family of process creation APIs. Counter type: Busy. LAUNCHES Number of calls made by the process to process creation procedures. For Guardian process creation, the counter includes unsuccessful creation attempts. Counter type: Incrementing.
MESSAGES-SENT Number of messages sent by the linker process. User I/O operations (except on $RECEIVE) and calls to some system procedures increment this counter. Counter type: Incrementing. MQC-ALLOC-FAILURES Number of MQC requests that failed. Counter type: Incrementing. MQC-ALLOCATIONS Number of MQCs currently allocated to the process. Counter type: Incrementing. MQCS-INUSE-QTIME Time for which MQCs have been allocated to the process. Counter type: Queue.
OSS-TTY-WRITE-BYTES Number of bytes written by the process to all files opened as /dev/tty. Counter type: Accumulating. OSS-TTY-WRITES Number of WRITE operations performed by the process to all OSS file opens of /dev/tty. Counter type: Incrementing. OSSNS-DD-CALLS Number of data definition requests sent to OSS name servers.
Counter type: Queue. PRES-PAGES-START Number of pages in main memory sponsored by the measured process at the start of the measurement interval. Counter type: Snapshot. READY-TIME The time that the measured process spent on the ready list or executing. To keep the measurement overhead to a minimum, this counter is not turned off during certain system activities. This counter includes: • Time required to dispatch the process.
Counter type: Accumulating. RECV-QTIME Time that messages spent waiting on the process message input queue. For user processes, the message input queue is $RECEIVE. Counter type: Queue. REPLY-BYTES Number of message bytes sent by a listener process in reply to a message from a linker process. The message bytes are also included in the RETURNED-BYTES counter of the linker process that received the message. In G-series and earlier RVUs, this is a 32-bit counter.
SENT-CBYTES Number of protocol control bytes sent by the message system on behalf of this process. This counter is updated when the process is acting as a linker process. The protocol control bytes are not included in the counter SENT-BYTES. Counter type: Accumulating. TNS-BUSY-TIME Time that the CPU was busy executing TNS code in this process. This counter applies only to measurements taken on a TNS/R or TNS/E system. See Usage Notes for H-Series and J-Series PROCESS Entities (page 310).
LCBS-INUSE-QTIME (Legacy Style only) Redefined. See MQCS-INUSE-QTIME. MAX-LCBS-INUSE (Legacy Style only) Redefined. See MAX-MQCS-INUSE. MAX-MQCS-INUSE (Legacy Style only) Maximum number of MQCs on the queue. Counter type: Max queue. MSGS-SENT-QLEN-MAX (Legacy Style only) Maximum number of items on the queue described by the MSGS-SENT-QTIME counter. In H-series and J-series RVUs, the value is always 1. Counter type: Max queue.
Counter type: Busy. UCL-LOCK-MAX (Legacy Style only) No longer used. UCL-LOCK-QTIME (Legacy Style only) No longer used. UCL-MAX (Legacy Style only) In Measure G10 and later PVUs, zero. In Measure G09 and earlier PVUs, maximum number of items on the queue described by the UCL-QTIME counter. Counter type: Max queue. UCL-QTIME (Legacy Style only) In Measure G10 and later PVUs, zero.
priority of 100. If the priority for that process were changed to 150 using one of the procedures listed above, its initial priority would thereafter be 150. NOTE: The original initial priority value of a process (also known as the creation priority) is displayed in the first line of the PROCESS header and stored in the PRIORITY field in the ID section of the ZMSPROC record. This PRIORITY value is static during a measurement.
by the memory manager process, PIN 1, which also sponsors pages allocated by and for the system. ◦ • An unsponsored locked page frame remains unsponsored (associated with no process) until the next routine visit by the Memory Manager process. The initial record of the PROCESS entity is created when the process is started and not at the measurement start time.
PROCESSH The PROCESSH entity provides information about the relative execution time of one or more code ranges within a program. You can measure code ranges by specifying procedure names and address ranges or by specifying a code-space category (user code, user library, system code, or system library). If you use a wildcard in a file name that applies to multiple disk files (such as $SYSTEM.WORK.*), you must specify a code-space category rather than a procedure name and address range.
{ [[$device.]subvolume.]filename[:crvsn] [ ( pid ) ] } { "pname" [ ( pid ) ] } where * measures all processes on the local system. pid is the process identifier of the process to be measured. Specify pid as one of: { cpu,pin } { SYSTEM-PROCESSES } cpu is the number of the CPU on which the process to be measured is running. To indicate all CPUs, use an asterisk (*). The default is all CPUs. pin is the process identification number of the process to be measured.
where code-space is a code-space specification. In Measure G12 and later PVUs, code-space is required only for TNS and accelerated TNS code; otherwise this parameter is accepted but ignored. For accelerated and TNS code, the code-space specification is: Space Specification Accelerated TNS User code UC[.n], where n is in the range: 0-31 0-15 User library UL[.n], where n is in the range: 0-31 0-15 System code SC[.n], where n is in the range: 0-31 0 System library SL[.
code-range-address In an EDIT (code 101) file describing TNS code, code-range-address is an octal address in the range 0 through 177777 or an address offset. For EDIT (code 101) files describing TNS/R or TNS/E code, code-range-address is a valid virtual address for the systems, in hexadecimal, or an address offset. To indicate an offset from the previous code-range-address, precede the octal or hexadecimal number with a plus sign (+).
02 process-name 02 program-file-name. 03 volume 03 subvol 03 filename 02 osspid 02 ancestor-cpu 02 ancestor-pin 02 ancestor-sysname 02 ancestor-process-name 02 device-name 02 program-file-name-MID. 03 PATHID 03 CRVSN 02 code-space 02 object-device-name 02 object-file-name-MID. 03 PATHID 03 CRVSN 02 code-range-flags 02 code-range-length 02 reserved-1 02 code-range end type character 8.
02 process-busy-samples type binary 32 unsigned. 02 code-space type binary 16 unsigned. 02 code-space-busy-samples type binary 32 unsigned. 02 code-range type character 32. 02 code-range-busy-samples type binary 32 unsigned. * new entity identification field: 02 userid. 03 group type binary 8 unsigned. 03 user type binary 8 unsigned. 02 creatorid. 03 group type binary 8 unsigned. 03 user type binary 8 unsigned. * fields for Liberty specific counters as of C30: 02 accel-busy-samples type binary 32 unsigned.
CODE-RANGE Name of the measured code range. CODE-RANGE-FLAGS Reserved for future use. CODE-RANGE-LENGTH Reserved for future use. CODE-SPACE One of these code space identifiers: 0 UC.n (user code—TNS or accelerated) 1 UCR (user code—TNS/R native) 2 UL.n (user library—TNS or accelerated) 3 ULR (user library—TNS/R native) 4 SC.0 (system code—TNS or accelerated) 5 SCR (system code—TNS/R native) 6 SL.
PROGRAM-FILE-NAME Name of the object file the process is executing. The name can apply to either a NonStop file or an SMF file. PROGRAM-FILE-NAME has three subfields: VOLUME For NonStop files, is the device location of the physical file. For SMF files, is the device location of the physical file that corresponds to the logical file name is stored in the DEVICE-NAME field. SUBVOLUME For NonStop files, is the device name of the physical volume on which the program file is located.
Counter type: Sampling. NATIVE-BUSY-SAMPLES (ZMS Style only) Number of times the sampling operation found the measured code executing native code. Counter type: Sampling. PROCESS-BUSY-SAMPLES Number of times the sampling operation found the process executing. Counter type: Sampling. TNS-BUSY-SAMPLES Number of times the sampling operation found the measured code executing TNS code. This counter applies only to measurements taken on a TNS/R or TNS/E system.
DEFINE Statement Scaling Factor ADD DEFINE =_SET_PROCESSH_RATE, file SCALE16X 16 time ADD DEFINE =_SET_PROCESSH_RATE, file SCALE32X 32 times To determine the approximate sampling rate, issue successive PEEK INT requests in a processor to view the accumulated sample interrupts, or examine the counter PROCESSH-SAMPLES in the ZMSCPU report or through MEASCOM LIST CPU command.
• This example describes how to check for shared run-time libraries (SRLs) in a program and how to include them in your program’s PROCESSH measurement: 1. Identify all the SRLs associated with a particular program using the Native Object File Tool (NOFT) utility. 2. Add the SRLs as user library files (URLs) in MEASCOM. The accounting and mapping of the PROCESSH total samples count for the program and the sum of the individual code-space sample counts now match as closely as currently possible.
NOTE: You can measure ServerNet addressable controllers with a D-series RVU measurement application if it specifies all controllers (ADD CONTROLLER *) or only a CPU number (ADD CONTROLLER 2). To measure specific SACs, you must modify the entity identifiers.
slot is the slot number of the SAC to be measured. Use an asterisk (*) to indicate all slots. The default is all slots. subdevice is a specific subdevice identifier in cases where the group, module, and slot identifiers or the configuration name indicate more than one SAC. For RIPC, subdevice is equal to the CPU number of a processor in a remote node, system, or ServerNet cluster. To indicate all remote processors on the specified remote node, system, or ServerNet cluster, use an asterisk (*).
Counter Fields DDL Definition DEFINITION zmssvnet-ctrs.
03 remote-cluster 03 slot 02 subdevice 02 remote-CPU * F40 new counter value items: 02 io-qbusy-time * G01 counter value items: 02 read-requests 02 read-bytes 02 read-qtime 02 read-qlen-max 02 read-qbusy-time 02 write-requests 02 write-bytes 02 write-qtime 02 write-qlen-max 02 write-qbusy-time * G08 new counters: 02 read-cbytes 02 write-cbytes 02 server-qtime 02 server-qlen-max 02 retries 02 acks 02 x-defrd-busy-time 02 y-defrd-busy-time end type binary 32 unsigned redefines module.
For RIPC records, GROUP and SLOT both have values of 0. You can use the keywords GROUP, MODULE, REMOTE-CLUSTER, and SLOT in the IF and BY clauses of the LIST and LISTALL commands. NODE-CLASS SAC type as defined in the system configuration database. NODE-CLASS-S Redefines NODE-CLASS as a four-character alphanumeric field.
SUBDEVICE Subdevice identifier for a particular SCSI device on a multifunction I/O board (MFIOB): Identifier Device type 0 ServerNet bus interface (SBI) 1 Internal SCSI Controller #1 2 Internal SCSI Controller #2 3 COMM and SP 4 SP to SP communications 5 External SCSI port controller For IPC records, SUBDEVICE applies to the destination (listener) processor. For more information, see Usage Notes for SERVERNET Entities (page 332).
READ-CREQUESTS (ZMS Style only) For a CLIM, number of incoming control requests from this entity into the CPU. This counter appears in the NORMAL but not the BRIEF display. Counter type: Incrementing. READ-QBUSY-TIME Time spent in a state in which requests for data transfer from this entity to memory were queued. Counter type: Queue Busy. READ-QTIME Total time spent by read requests queued for this entity. With REPORT RATE ON, the average number of read requests queued for this entity.
WRITE-CREQUESTS (ZMS Style only) For a CLIM, number of outgoing control requests from the CPU into this entity. This counter appears in the NORMAL but not the BRIEF display. Counter type: Incrementing. WRITE-QBUSY-TIME Time spent in a state in which requests for data transfer from memory to this entity were queued. Counter type: Queue Busy. WRITE-QTIME Total time spent by write requests queued for this entity. With REPORT RATE ON, the average number of write requests queued for this entity.
IO-QLEN-MAX (Legacy Style only) Maximum number of outstanding I/Os on the queue described by the IO-QTIME counter. Counter type: Max queue. READ-QLEN-MAX (Legacy Style only) Maximum number of requests on the read request queue since the counter record was allocated. Counter type: Max queue. SERVER-QLEN-MAX (Legacy Style only) Maximum number of items on the queue described by the SERVER-QTIME counter. This counter applies only to IPC and RIPC records.
where xx is the linker CPU number and yy is the listener CPU number. • To support ServerNet cluster connections, the data in the linker and listener records have changed to let the linker record (initiating processor) fully account for message traffic. This support is accomplished by ensuring that both the linker and listener records count the same number of message exchanges and corresponding bytes counts.
Except for NonStop BladeSystems, each CLIM is uniquely identified by group-module-slot-port, which identifies the CLIM in terms of the physical location of the P-Switch it is connected to and the port on which it is connected. The group (system enclosure) and module (subset of a group) identifies the P-Switch. The slot (physical, labeled space in a module) identifies a particular ServerNet PIC within that P-Switch and the port.
Examples of Configuring Measurements for ServerNet Cluster The ADD, DELETE, LIST, LISTACTIVE and LISTALL commands can refer to one or more SERVERNET entities. Although this example uses the ADD command, the DELETE, LIST, LISTACTIVE and LISTALL commands all share the same command syntax for configuring or displaying measurements on remote ServerNet cluster activity. Command Measures All Remote IPC Activity Between...
SQLPROC entity-spec [,entity-spec] ... SQLPROC collects information about one or more SQL processes on the local system. entity-spec is specified as: { { { { { * } $process-name [ ( pid ) ] } disk-filename [ ( pid ) ] } "pname" [ ( pid ) ] } pid } where: * measures all SQL processes on the local system. $process-name is the name of the SQL process to be measured. pid is a process identifier. The process identifier identifies the owner of the SQL process to be measured.
"pname" (Measure G09 and later PVUs) is either a fully qualified or partial OSS file pathname. An OSS file pathname that does not begin with a slash (/) is considered to be a partial pathname and is expanded by prefacing it with the current setting for OSSPATH. An OSS file pathname that begins with /G is not supported, use the Guardian file name instead. NOTE: OSS file pathnames are case-sensitive and must be specified within double quotation marks (" ").
02 open-time end type binary 64. DDL Record Description Fields RECORD zmssqlp. FILE is "zmssqlp" ENTRY-SEQUENCED. 02 hdr type zmsheader. 02 ctr type zmssqlp-ctrs. 02 id type zmssqlp-id. end DDL Record for SQLPROC Entities (Legacy Style) The Legacy Style DDL record for SQLPROC entities will not change after the Measure G10 PVU. The fields included in BRIEF reports are in boldface type. RECORD sqlproc. FILE is "sqlproc" ENTRY-SEQUENCED. . . .
03 GMOM-jobid 02 GMOM-full-id 02 GMOM-sysname 02 GMOM-process-name end type binary 16 unsigned. type binary 64 redefines GMOM. type character 8. type character 8. For descriptions of the header fields used by all entities, see Common Entity Header Fields (page 144). Common ID Field DDL Definitions for SQLPROC Entities These ID fields are used in both styles of DDL records. Fields that are used in ZMS Style DDL records only are noted.
GMOM-PROCESS-NAME Name of the GMOM process that initiated the NETBATCH job that in turn initiated this process; otherwise spaces. GMOM-SYSNAME Expand system name of the GMOM node if the GMOM process is remote; otherwise spaces. OSSPID If the process runs in the OSS environment, this field contains the OSS process ID. For non-OSS processes, this field contains zeros. PIN Process identification number of the measured SQL process. PRIORITY Creation priority of the process.
OPEN-TIME Time this process spent executing OPENs. Counter type: Elapsed. OPENS Number of OPEN calls performed by the SQL executor on behalf of this process. Counter type: Incrementing. SQL-NEWPROCESS-TIME Time this process spent executing SQL-NEWPROCESS. Counter type: Elapsed. SQL-NEWPROCESSES Number of NEWPROCESS calls generated by the SQL executor on behalf of this process. Counter type: Incrementing. SQL-OBJ-RECOMPILE-TIME Time this process spent recompiling objects.
executes them. In SQL/MX, however, SQL procedures and modules are compiled and stored separately from the program object file that executes them. In Measure G11 and later PVUs, the SQLSTMT entity displays ANSI SQL names in SQLSTMT reports. In Measure H01 and later PVUs, MEASCOM LIST commands accept ANSI SQL/MX release 2 module object names (run unit names) in SQLSTMT entity specifications.
{ { { { { { disk-filename ( pid ) "run-unit", #index disk-filename ( pid ) "run-unit", [Index] #index "pname" (pid) "run-unit", #index "pname" (pid) "run-unit" [Index] #index pid "run-unit", #index pid "run-unit" [Index] #index } } } } } } where: * measures all SQL statements in all processes on the local system. $process-name is the name of the process that contains the SQL statements to be measured. pid is a process identifier. The process identifier identifies the owner of the process to be measured.
index is the index number of the statement to be measured. If omitted, all current indexed statements are measured. disk-filename is the name of an object file that contains SQL code to be measured. This file must be the object file of a running process. Specify the disk-file name as: { [ \system.]$device.subvol.filename[:crvsn]} { subvol.filename[:crvsn] } { filename[:crvsn] } \system is the system name. To indicate all devices, use an asterisk (*). The default is the current default system.
02 process-name 02 program-file-name. 03 volume 03 subvol 03 filename 02 osspid 02 ancestor-cpu 02 ancestor-pin 02 ancestor-sysname 02 ancestor-process-name 02 device-name 02 program-file-name-MID. 03 PATHID 03 CRVSN 02 statement-index 02 GMOM. 03 GMOM-node 03 GMOM-cpu 03 GMOM-pin 03 GMOM-jobid 02 GMOM-full-id 02 02 02 02 02 02 02 end GMOM-sysname GMOM-process-name version creator-type run-unit-len reserved-1 run-unit type character 8.
. (error, time items, and measurement identification items; see Common Entity Header Fields (page 144)) . . . * entity identification items: 02 pin type binary 16 unsigned. 02 process-name type character 8. 02 program-file-name. 03 volume type character 8. 03 subvol type character 8. 03 filename type character 8. 02 priority type binary 16 unsigned. 02 userid. 03 group type binary 8 unsigned. 03 user type binary 8 unsigned. * counter record identifiers and counter values: 02 run-unit type character 32.
02 GMOM-sysname 02 GMOM-process-name * New Version Field 02 VERSION 02 CREATOR-TYPE type character 8. type character 8. type binary 16 unsigned. type binary 16 unsigned. * New identifiers for full ANSI SQL name 02 FULL-NAME-OFFSET type binary 16 unsigned. 02 FULL-NAME-LEN type binary 16 unsigned. * Variable data goes here, pointed to by VARSTRING (offset/length) * VAR-DATA needs to be defined for the maximum expected variable length data. 02 FULL-NAME end type character 784.
GMOM-PIN PIN of the GMOM process if this process is part of a NetBatch job; otherwise zero. GMOM-JOBID Job ID of the NetBatch job that initiated this process, assigned by the indicated GMOM; otherwise zero. GMOM-FULL-ID 64-bit redefinition of the individual GMOM fields as a single value. A value of zero indicates the process is not part of a NETBATCH job stream. GMOM-PROCESS-NAME Name of the GMOM process that initiated the NETBATCH job that in turn initiated this process; otherwise spaces.
PROGRAM-FILE-NAME-MID Internal format representation of an OSS file pathname. PROGRAM-FILE-NAME-MID has two subfields: PATHID is an internal format representation of an OSS file pathname. For non-OSS files, the field contains zeros. CRVSN is a creation version serial number that identifies a unique instance of an OSS file. For non-OSS files, the field contains zeros. RUN-UNIT (ZMS Style only) Left-justified external-format ANSI SQL name.
In G-series and earlier RVUs, this is a 32-bit counter. In H-series and J-series RVUs, it is a 64-bit counter. In G-series RVUs, the DISC-READS-F field is a 64-bit version of DISC-READS. Counter type: Incrementing. ELAPSED-BUSY-TIME Time the process spent executing the measured SQL statement. If REPORT RATE is on, the output value is the percent of time busy during the interval.
RECORDS-ACCESSED Number of records accessed by the disk process or file system to evaluate this statement. In G-series and earlier RVUs, this is a 32-bit counter. In H-series and J-series RVUs, it is a 64-bit counter. In G-series RVUs, the RECORDS-ACCESSED-F field is a 64-bit version of RECORDS-ACCESSED. Counter type: Accumulating. RECORDS-USED Number of records inserted, updated, deleted, or read by the SQL executor through this statement. In G-series and earlier RVUs, this is a 32-bit counter.
RUN-UNIT-128 (Legacy Style only) Name of the SQL/MP procedure or SQL/MX module that contains the measured SQL statement. Legacy Style Only Counter Field DDL Definitions for SQLSTMT Entities These counter fields are used in Legacy Style DDL records only. Fields are listed alphabetically. CALLS-F (G-series, Legacy Style only) Same as CALLS but accommodates larger values (64 bits rather than 32).
$*.*.file means the specified file in any subvolume in any volume. • To ensure that the SQLSTMT entity is captured when monitoring a SQL/MP application, start the SQLSTMT entity measurement before starting the SQL application process and stop the SQLSTMT entity measurement after the SQL process stops (that is, start and stop the SQLSMT entity measurement so that it spans the entire execution of the SQL application.
Which network links are measured depends on the entity specification in the measurement configuration: • If you specify all systems, the Measure subsystem counts the local end of all messages passed between the local system and any remote system. • If you specify a local CPU, the Measure subsystem counts the local end of all messages passed between the local system and any remote system through the line-handler processes (primary or backup) in the specified CPU.
Counter Fields DDL Definition DEFINITION zmssys-ctrs. 02 links 02 link-time 02 sent 02 received 02 sent-forward 02 received-forward end type type type type type type binary binary binary binary binary binary 64. 64. 64. 64. 64. 64. DDL Record Description Fields RECORD zmssys. FILE is "zmssys" ENTRY-SEQUENCED. 02 hdr type zmsheader. 02 ctr type zmssys-ctrs. 02 id type zmssys-id.
LINK-TIME Time spent waiting on acknowledgment from the remote system of messages sent to it by the local system. Counter type: Syslink. LINKS Number of messages (each consisting of one or more packets) sent from the local system to the remote system. This counter does not include level 4, file-system interface, acknowledgment messages. Counter type: Incrementing. RECEIVED Number of packets sent from the remote system to the local system with the local system as their final destination.
TERMINAL The TERMINAL entity provides information about terminal I/O. Topic Page Entity specification syntax for TERMINAL entities 357 DDL records for TERMINAL entities 358 DDL field definitions for TERMINAL entities 359 Usage notes for TERMINAL entities 361 Entity Specification Syntax for TERMINAL Entities To describe TERMINAL entities: TERMINAL entity-spec [,entity-spec] ...
line (G-series) is 0 or 1 to indicate a specific line managed by a CLIP in a 3880 SWAN concentrator. To indicate all lines, use an asterisk (*). type (G-series) is a type number to indicate a particular protocol, such as 61 (X.25). Use SCF LISTDEV to see valid values for line types. To indicate all line types, use an asterisk (*). subtype (G-series) is a subtype number, such as 62 (X25AM). To see valid subtype numbers, use SCF LISTDEV. The default is all subtypes that apply to type.
The Legacy Style DDL record for G-series TERMINAL entities through the G10 Measure PVU is identical to the record for D-series TERMINAL entities except: • Longer byte-count fields are provided to reduce the possibility of field overflow. In G-series RVUs, each 32-bit byte-count field has a 64-bit counterpart. For information on using the 64-bit fields, see Usage Notes for G-Series TERMINAL Entities (page 361). • The ERROR field can signal a field overflow in a 32-bit byte-count field. RECORD terminal.
In G-series RVUs, the INPUT-BYTES-F field is a 64-bit version of INPUT-BYTES. Counter type: Accumulating. OUTPUT-BYTES Number of bytes written to the terminal. The total number of bytes transferred to and from the terminal is INPUT-BYTES plus this counter. Because the terminal process modifies this counter before the I/O operation, this counter might not be accurate if the write instruction fails.
Fields are listed alphabetically. INPUT-BYTES-F (G-series, Legacy Style only) Same as INPUT-BYTES but accommodates larger values (64 bits rather than 32 bits). Counter type: Accumulating. OUTPUT-BYTES-F (G-series, Legacy Style only) Same as OUTPUT-BYTES but accommodates larger values (64 bits rather than 32 bits). Counter type: Accumulating.
Entity Specification Syntax for TMF Entities To describe TMF entities: TMF entity-spec [,entity-spec] ... TMF collects information about TMF transactions in one or more CPUs on the local system. entity-spec is specified as: { * | cpu } where: * measures TMF activity on all CPUs. cpu is the number of the CPU on which TMF transactions are measured. Use a comma-separated list to specify multiple CPUs.
. . . * entity identification items: (none) * counter value items: 02 home-trans 02 home-trans-qtime 02 home-trans-qmax 02 remote-trans 02 remote-trans-qtime 02 remote-trans-qmax 02 home-net-trans 02 home-net-trans-qtime 02 home-net-trans-qmax 02 aborting-trans 02 trans-backout-qtime 02 trans-backout-qmax end type type type type type type type type type type type type binary binary binary binary binary binary binary binary binary binary binary binary 32 unsigned. 64. 16 unsigned. 32 unsigned. 64.
HOME-NET-TRANS-QTIME Time that transactions that began on the local system and were subsequently distributed to a remote system spent on the remote system. This counter is advanced only in the TMF record of the CPU containing the primary $TMP process. If the primary $TMP switches CPUs in response to a PRIMARY command, the new $TMP increments the HOME-NET-TRANS-QTIME counter in the TMF record of its CPU until it accurately reflects the number of HOME NETWORK transactions on the system.
number of transactions in the local system waiting for BACKOUT. The TRANS-BACKOUT-QTIME counter in the TMF record of the CPU that contained the old primary becomes zero. Counter type: Queue. Legacy Style Only Counter Field DDL Definitions for TMF Entities These counter fields are used in Legacy Style DDL records only. Fields are listed alphabetically. HOME-NET-TRANS-QMAX (Legacy Style only) Maximum number of items on the queue described by the HOME-NET-TRANS-QTIME counter. Counter type: Max queue.
Counter type: Response time. USERDEF The USERDEF entity provides information about user-defined counters. In Measure G09 and later PVUs, the USERDEF entity type handles OSS file pathnames. In the ZMS Style interface (Measure G11 and later PVUs), the record template, report, and structured data file name for USERDEF records is ZMSUDEF.
$device is the name of the volume (device) that contains the process to be measured. To indicate all devices, use an asterisk (*). The default is the current default device. subvolume is the name of the subvolume that contains the process to be measured. To indicate all subvolumes, use an asterisk (*). The default is the current default subvolume. filename is the name of the object file for the process. filename must be the object file of an executing process.
03 GMOM-pin 03 GMOM-jobid 02 GMOM-full-id 02 02 02 02 02 02 end GMOM-sysname GMOM-process-name type index reserved-2 name type binary 16 unsigned. type binary 16 unsigned. type binary 64 redefines GMOM. type character 8. type character 8. type binary 16 unsigned. type binary 16 unsigned. type character 4. type character 16. Counter Fields DDL Definition DEFINITION zmsudef-ctrs. 02 value end type binary 64. DDL Record Description Fields RECORD zmsudef. FILE is "zmsudef" ENTRY-SEQUENCED.
02 ancestor-sysname type character 8. 02 ancestor-process-name type character 8. * SMS changes: 02 device-name type character 8. * Identifiers for OSS file pathname support: 02 osspid type binary 32 unsigned. 02 program-file-name-mid. 03 pathid type binary 64. 03 crvsn type binary 64. * New identifiers for NetBatch Job Control: 02 GMOM. 03 GMOM-node type binary 16 unsigned. 03 GMOM-cpu type binary 16 unsigned. 03 GMOM-pin type binary 16 unsigned. 03 GMOM-jobid type binary 16 unsigned.
GMOM-PIN PIN of the GMOM process if this process is part of a NetBatch job; otherwise zero. GMOM-JOBID Job ID of the NetBatch job that initiated this process, assigned by the indicated GMOM; otherwise zero. GMOM-FULL-ID 64-bit redefinition of the individual GMOM fields as a single value. A value of zero indicates the process is not part of a NETBATCH job stream. GMOM-PROCESS-NAME Name of the GMOM process that initiated the NETBATCH job that in turn initiated this process; otherwise spaces.
PROGRAM-FILE-NAME-MID Internal format representation of an OSS file pathname. PROGRAM-FILE-NAME-MID has two subfields: PATHID is an internal format representation of an OSS file pathname. For other files, the field contains zeros. CRVSN is a creation version serial number that identifies a unique instance of an OSS file. For other files, the field contains zeros.
Usage Notes for All USERDEF Entities • To define a counter in an application, modify the source code to call the MEASCOUNTERBUMPINIT and MEASCOUNTERBUMP procedures at appropriate times. For descriptions, see MEASCOUNTERBUMP (page 387) and MEASCOUNTERBUMPINIT (page 388). • To collect information from a user-defined counter, use the ADD USERDEF command to specify the process that contains the user-defined counter and the ADD COUNTER command to specify the name of the counter.
4 Measure Callable Procedures This chapter describes the Measure callable procedures, which let an application process control the Measure performance monitor and access measurement data. For general information on Measure procedures, see Measure Procedures Overview (page 375). Summary of Measure Procedures Table 5 lists the Measure procedures by function.
Table 5 Measure Callable Procedures (continued) Function Procedure Description Page MEASCOUNTERBUMPINIT Returns the offset of a user-defined counter 388 Controlling the MEASMONCONTROL Measure subsystem MEASMONSTATUS Returns the name and number 412 of currently active measurement Getting system configuration information MEASLISTCONFIG Returns system configuration information supplied by the MEASCTL process Translating file names MEAS_CODERANGENAME_DEMANGLE_ Translates the internal 382 (mangled)
Table 5 Measure Callable Procedures (continued) Function Procedure Description Page Adjusting record formats MEAS_ADJUSTZMSRECORD_ Adjusts ZMS style structure records to the MEASDDLS format with which an application was compiled 380 Maintaining fine granularity timers MEAS_ALLOCATE_TIMERCELLS_ Allocates timer cells for maintaining TCELLBUSY, TCELLQUEUE, and TCELLQBUSY timers 381 MEAS_BUMP_TIMERCELL_ Sets, resets, increments, or 381 decrements a TCELLBUSY, TCELLQUEUE, or TCELLQBUSY timer MEAS_D
Reading in the Declaration Files Use the TAL compiler ?SOURCE command to read these declaration files into the source code global declarations: • $SYSTEM.SYSTEM.EXTDECS0 contains the external declarations for the Measure procedures. Each Measure procedure used in your program must be specified in the ?SOURCE command. • $SYSTEM.SYSnn.MEASDECS contains the template structures for the Measure control block, configuration table, and entity descriptors.
• In name fields, an asterisk (*) specifies all. Such fields must be left justified and blank filled. Include, as the left-most character, any appropriate leading character, such as a backslash (\) for a node, a dollar sign ($) for a volume, or a pound sign (#) for a subdevice. • To designate all system processes, set the PIN to -2. Wildcards are not always valid. For example, when Measure reads from active counters, the entity descriptor you pass to MEASREADACTIVE must specify a single entity.
Header Record Use the template structure CONTAB^HDR to define the header record, which contains: Variable Description Type The numeric value 50 or the literal CONTAB^T (declared in the MEASDECS file). This field identifies this record as the header record. Len The length of the entire configuration table, in bytes. Sections An array of offsets in bytes that point to each entity type’s descriptor section within the table. The first word of the array is always zero.
try allocating more data pages with the TAL compiler ?DATAPAGES command or the RUN command MEM parameter. Maintaining Compatibility With New Structures in MEASDDLS and MEASCHMA In some instances, a recompile is required for C and C++ applications. For example, G08 Measure redefines the MODULE and SUBDEVICE fields of the SERVERNET entity descriptor as the ServerNet Cluster NODE number and CPU number of the remote system's processor.
MEAS_ADJUSTZMSRECORD_ Takes ZMS style structure records in any published format and adjusts them to the MEASDDLS format with which an application was compiled. error := MEAS_ADJUSTZMSRECORD_ ( in^record ,out^record ,templateversion ); ! i ! o ! i in^record input INT .EXT:ref:n is a buffer containing a ZMS style external record to be padded or truncated to a new desired format.
MEAS_ALLOCATE_TIMERCELLS_ Allocates the specified number of timer cells and returns an array of indexes to those timers. Use this procedure if your application requires fine granularity timers to maintain counters outside of Measure. Measure calls this procedure implicitly when you define a timer-cell counter for the USERDEF entity. error := MEAS_ALLOCATE_TIMERCELLS_ ( count ,types ,indexes); ! i ! i ! i,o count input INT:value is the number of timer cells to allocate.
• INCQUEUE (5) to increase a TCELLQUEUE counter. Use this value when an item is placed in the queue. • DECQUEUE (6) to decrease a TCELLQUEUE counter. Use this value when an item is removed from the queue. • INCQBUSY (7) to increase a TCELLQBUSY counter. Use this value when the first item is placed in a queue after the queue has been empty. • DECQBUSY (8) to decrease a TCELLQBUSY counter. Use this value when the last item is removed from the queue, rendering the queue empty.
OutputNameMaxLen input INT:value is the size of the input buffer. OutputNameLen output INT .EXT:ref:* is the count of bytes in the output buffer. NameType output INT .EXT:ref:1 is 1 for a non-C++ name; 2 for a C++ name. MEASCONFIGURE Configures a measurement. Your measurement application must call the MEASOPEN procedure (to initialize a Measure data file) before calling MEASCONFIGURE.
contab input INT .EXT:ref:* is an array that defines the measurement configuration. The contab array contains a header section, an entity descriptor section for each entity type, and a trailer section: CONTAB^HDR ... maxents entitydesc_1 [ entitydesc_2 ... entitydesc_n ] CONTAB^TRAILER ... len Structure definitions for use in the contab array are defined in the MEASDECS file. MEASDECS declares both a literal and a numeric identifier for each entity type and for the header and trailer records.
entitydesc_1 [ entitydesc_2 ... entitydesc_n ] defines the entities in the measurement. Specify each entitydesc as one of the entity descriptors described in Chapter 5: Entity Descriptors (page 451). In Measure H01 and later PVUs, MEASCONFIGURE accepts DISCOPEN, DISKFILE, FILE, or SQLSTMT entity descriptors for ANSI SQL objects or partitions. CONTAB^TRAILER marks the end of the contab array.
MEASCONTROL Starts and stops a measurement, including a measurement interval as an option. Before calling MEASCONTROL, your program must call MEASCONFIGURE and configure the measurement. error := MEASCONTROL ( meascb ,measnum ,[ starttime ] ,[ stoptime ] ,[ interval ] ); ! ! ! ! ! i,o i i i i meascb input, output INT:ref:$LEN(MEASCB^DEF) / 2 is a control block where the Measure subsystem stores data for subsequent procedure calls.
MEASCOUNTERBUMP Modifies the counter identified by offset provided that: • The counter is part of a currently active measurement. If the counter is not currently being measured, MEASCOUNTERBUMP returns ERR^UDCNOTPRESENT. • The counter type is compatible with bumptype. You specify the counter type when you add the counter to the measurement configuration. For more information about user-defined counters, see MEASCOUNTERBUMPINIT (page 388).
doubleadd input INT(32):value is the value (positive or negative) to add to the counter. Applicable only if you specified bumptype as ADD and the counter type as FACCUM. doubleadd cannot be specified if addvalue is also specified. quadadd input FIXED:value is the value (positive or negative) to add to a 64-bit counter. Applicable only if you specified bumptype as ADD and the counter type as FACCUM. quadadd cannot be specified if addvalue is also specified.
MEAS_DEALLOCATE_TIMERCELLS_ Deallocates timer cells previously allocated by the application. Use this procedure if your application requires fine granularity timers to maintain counters outside of Measure. Measure calls this procedure implicitly for timer-cell counters defined for the USERDEF entity. Measure also calls this procedure implicitly if your program ends without invoking the procedure.
dfnum input INT:value is the data file access number, or -1, which is also the default value. To access the journal segment of an open measurement data file containing a journal segment, use the dfnum value returned by the MEASOPEN procedure. If -1 is specified, or if no journal segment is available for the specified data file but the file is from the current system, resolution is from the SQL/MX subsystem. If the descriptor is used for an active measurement, -1 should be passed. name input STRING .
connectionInfo input INT(32) .EXT:ref:25 allocated by caller to be at least 100 bytes, this field is filled in by a call to MEAS_SQL_MAP_INIT_. This value is only needed when the SQL journal is not being used. Usage Notes • When used in conjunction with the SQL journal, each ANSI SQL name can correspond to several instances of the same object. • If this API is called with a name_len of 0, the pathid, crvsn and uid fields are filled with wildcard values.
MEASGETVERSION Returns the Measure version of the specified data file. If the call provides a buffer, the procedure also returns an array of external entity record lengths for the version, indexed by entity type. You can use MEASGETVERSION for two purposes: • You can use the Measure version of the data file to decide which MEASFH version to use. • You can use the record length returned to decide the address of the records returned by other procedures (such as MEASREAD).
is an integer indicating the legacy style byte length of the buffer array. This value should typically be equal to (MAX^T+1) * 2. sysname output INT .EXT:ref:4 is an 8-byte string that returns the name of the system on which the data file was created. release output INT .EXT:ref:2 is a two-word array containing the Measure version identification: RELEASE[0] = F40 RELEASE[1] = 1 processor^type output INT .
MEASINFO Returns the measurement configuration from a data file. Unlike the MEASREADCONF procedure, this procedure does not require that the calling process has already opened the file (using MEASOPEN). Because MEASINFO uses less disk space than MEASREADCONF, HP recommends it for tasks that do not require retrieval of actual data records. Use it to help you decide whether a data file contains the needed information before creating MEASFH.
bytesret output INT:ref:1 is the byte size of the contab array (and the number of bytes returned to the destination buffer, contab.) If bytesret is larger than bufsize, error 3204 (ERR^BUFTOOSMALL) is returned. starttime output FIXED:ref:1 is the start time of the measurement. stoptime output FIXED:ref:1 is the stop time of the measurement. A value of -1 indicates that no stop time was specified. interval output FIXED:ref:1 is the collection interval of the measurement.
Entity Type Literal Value Numeric Identifier TMF TMF^T 14 SQLPROC SQLPROC^T 15 SQLSTMT SQLSTMT^T 16 OPDISK OPDISK^T 17 CONTROLLER CTRL^T 18 SERVERNET SVNET^T 18 DISKFILE DISKFILE^T 19 OSSCPU OSSCPU^T 20 OSSNS OSSNS^T 21 (Max value) MAX^T 24 ctrspace output INT(32) .EXT:ref:MAX^T+1 is an array that contains the maximum counter space in words used by each entity type. The array elements are in order by numeric identifier.
If the SQL journal file is under construction, bits 12 and 13 are both reported as ON. Usage Notes • MEASINFO calls ALLOCATESEGMENT to obtain a private extended data segment (segment ID 10) for read traversal of the data file. The extended segment size is set to accommodate two 30 KB (30,000-byte) read buffers and a 32 KB (32,000-byte) temporary buffer for reconstructing records that span the read buffers.
MEASLISTCONFIG Gets system configuration information from the MEASCTL process in a specified CPU. Each MEASLISTCONFIG call gets configuration information for one type of device associated with one CPU. error := MEASLISTCONFIG ( cpunum ,entity ,buf ,bufsize ,bytesret ,firstcall ); ! ! ! ! ! ! i i o i o i,o error INT is the error code indicating the outcome of the operation. Possible error codes include: Error Code Description 0 Successful completion.
firstcall input, output FIXED:ref:1 is a context value. The first time you pass entity to MEASLISTCONFIG, specify firstcall as 0. MEASLISTCONFIG modifies and returns firstcall. A firstcall value other than -1 indicates that more descriptors are available. Call MEASLISTCONFIG again with the same entity value and the returned firstcall value. When no more descriptors are available, MEASLISTCONFIG sets firstcall to -1 and returns error 3022 (WARN^NO^MORE^DATA).
Error Code Description 3295 err^sql^api^internal The ANSI SQL name specified could not be parsed. 3296 err^sqlmx^map^process Either connectionInfo was not passed, or connectionInfo had incorrect contents, probably because it had not been filled in by a call to MEAS_SQL_MAP_INIT. dfnum input INT:value is the data file access number, or -1. To access the journal segment of an open Measure data file that contains a journal segment, use the dfnum value returned by the MEASOPEN procedure.
extname_len output INT:ref:1 is the length, in bytes, of the name returned in extname, including the null terminator. extname_len is required when extname is specified. index input, output FIXED:ref:1 is a context value for iterative calls. On the initial call of MEASLISTENAME, index must be set to -1f. If the translation resolves to a single name, index is returned as -1f. If the translation resolves to multiple names, index contains an internal value that must be passed to subsequent MEASLISTENAME calls.
input_type input INT:ref:1 specifies the type of the data in the fname_pathid_uid field. input_type is a required parameter if the data is an ANS UID, otherwise it is optional. For ANSI SQL names: 0 File name data type 1 Path ID data type 2 ANS UID data type connectionInfo input INT(32) .EXT:ref:25 is allocated by the caller to be at least 100 bytes. connectionInfo must have been filled in by a call to MEAS_SQL_MAP_INIT_. connectionInfo is only needed when the SQL journal is not being used.
MEASLISTEXTNAMES Makes the Measure subsystem list structured OSS and ANSI SQL name information to the EXTNAMES file in the specified subvolume. error := MEASLISTEXTNAMES ( dfnum ,volume^subvol ); ! i ! i error INT is the error code indicating the outcome of the operation. Possible error codes include: Error Code Description 0 Successful completion. 3235 err^buildingossjournal The OSS journal segment for this file is still under construction.
MEASOPEN(). You must use the MEASOPEN() option parameter to open and "attach" a journal segment. For more information, refer to the MEASOPEN (page 414) Usage Notes. • If the EXTNAMES file already exists in the specified volume^subvol, new entries are written to the existing file, and all instances of error 10 (duplicate record) are ignored. A single EXTNAMES file can contain information for many measurements from many systems.
the specified data file and the file is from the current system, dfnum is retrieved from the SQL/MX or OSS subsystem. name input STRING .EXT:ref:* is a buffer containing the OSS file pathname or ANSI SQL object name to be translated. If passed, an ANSI SQL object name must be fully qualified and not contain wildcards. name_len input INT:value is the length, in bytes, of the OSS file pathname or ANSI SQL object name specified in name. The length includes a terminating null byte if it is an OSS path name.
extname_max input INT:value is the maximum length, in bytes, that the extname buffer can hold. Required when extname is specified. extname_len input INT:ref:12 for iterative calls, is the length, in bytes, of the OSS file pathname or ANSI SQL partition name returned in extname. The length includes the null terminator. index input, output FIXED:ref:1 is a context value for iterative calls. On the initial call of MEASLISTGNAME, index must be set to -1f.
Example These are simplified examples in pTAL pseudo code of how LISTGNAME could handle ANSI SQL names: Example of iterative calls: index := -1f; do begin if (error := measlistgname(dfnum, ansi_name, ansi_name_len, guardianname, sys, !pathid!, !crvsn!, !partname!, !partname_max!, !partname_len!, index, MEAS_ANSI_SQL_NAME )) then ! handle error else if guardianname <> " " then begin ! display guardian filename if partname_len then ! display the partition name end else ! display the TABLE or INDEX name end un
MEASLISTOSSNAMES Causes the Measure subsystem to list structured OSS file pathname translation information to the file OSSNAMES in the default subvolume. In Measure G11 and later PVUs, MEASLISTOSSNAMES is still available but is superseded by MEASLISTEXTNAMES (page 403), which supports both OSS file pathnames and ANSI SQL names. error := MEASLISTOSSNAMES ( dfnum ,volume^subvol ); ! i ! i error INT is the error code that indicates the outcome of the operation.
MEASLISTPNAME Translates a Guardian file name or an OSS pathid to its OSS file pathname equivalent. The OSS file pathname must be valid at the time of the call. If CRVSN is specified, only a translation that matches the specified CRVSN returns a pathname. In Measure G11 and later PVUs, MEASLISTPNAME is still available but is superseded by MEASLISTENAME (page 399), which supports both OSS file pathnames and ANSI SQL names.
sysname input INT .EXT:ref:4 is an array of four words. If the file is remote, the array should contain the EXPAND system name of the file specified in filename. crvsn input INT .EXT:ref:3 is an array of three words that can contain a CRVSN value for qualifying the pathid or file name specified. pathid output INT .EXT:ref:12 is the internal format pathid associated with the returned OSS file pathname. pathname output STRING .EXT:ref:* is a buffer where the OSS file pathname is returned.
MEASMONCONTROL Starts or stops the Measure subsystem. Alternatively, adds a CPU to the Measure subsystem by starting a measurement control process (MEASCTL) in the specified CPU. NOTE: The calling process must have a super-group user (255,*) process accessor ID to invoke this procedure. error := MEASMONCONTROL ( meascb ,start ,[ cpunum ] ); ,[ errDetail ] ); ! ! ! ! i,o i i o error INT:value is the error code indicating the outcome of the operation.
this parameter should not be used. The Measure subsystem restarts a MEASCTL automatically in a reloaded CPU. errDetail output INT:ref:1 (H06.20/J06.09 and later RVUs) contains the PROCESS_CREATE_ error detail number if: START is true, MEASMON could not be started, and error contains a PROCESS_CREATE_ error number. For a complete list of error numbers and their definitions for PROCESS_CREATE_, see the Guardian Procedure Calls Reference Manual. Usage Notes To set the CID Table limit to 1,024,000 CIDs in H06.
is an array of measurement data file names in 12-word format. The array index of a measurement data file name is its measurement number. The array size is 64 files (64 * 12 words). settings output INT:ref:1 is an array of flags that indicates the current setting of Measure subsystem parameters controlled through defines: settings.0:2 Maximum CIDs setting: 0 = 1,024,000 1 ... 7 = reserved for future use settings.3:7 Reserved for future use. settings.8 SQL/MX journal segment default ON/OFF. settings.
MEASOPEN Obtains read, write, or read and write access to a measurement data file; opens the file: and creates a Measure file-handling (MEASFH) process to access the file. Use the write and read parameters to specify access type. To close a data file, use the MEASCLOSE procedure. The best way to access a remote data file is to use the measfh parameter. You specify an object file (on the remote system) that contains a Measure file-handling (MEASFH) program.
read input INT:value is used with write to specify access type. For possible values and results, see write. measfh input INT:ref:12 is an array containing the name of an object file for a Measure file-handling (MEASFH) program. By default, MEASFH object files are named $SYSTEM.SYSnn.MEASFH. swapvol input INT:ref:4 is an array containing the volume name for MEASFH swap files. If swapvol is not specified, swap files are created on the swap volume of the calling process.
FIXED:value is the desired file size (capacity) in bytes. The minimum value for filesize is 133169152 (127 MB) and the maximum is 1099507433472 (1048572 MB) or the maximum disk size (if less than 1048572 MB). The default value is 1073741824 (1024 MB). A value of -1 indicates that no file size was specified. If filesize is less than the minimum allowed value or greater than the maximum allowed value, error 3203 (ERR^BADPARAMS) is returned.
function, and MEASOPEN returns a new error code if the SQL/MX journal segment is under construction. • In Measure H02 and later PVUs, the MEASOPEN callable procedure allows the caller to select the measurement data file size, suppress counter data records in the measurement data file, or both. MEASREAD Reads one or more counter records from a measurement data file. The maximum number of records returned in a single call depends on the size of the destination buffer, loc.
INT:ref:1 is the number of bytes returned to the destination buffer, loc. firstcall input, output FIXED:ref:1 is a context value. The first time you pass entitydesc to MEASREAD, specify firstcall as 0. MEASREAD modifies and returns firstcall. As long as MEASREAD returns a nonzero value in firstcall, more counter records are available. Call MEASREAD again with the same entitydesc value and the firstcall value that was returned to you. nomtime input FIXED:value is a nominal time.
MEASREAD_DIFF_ This procedure is an enhanced version of MEASREAD that reads a window of one or more counter records from a measurement data file. Before reading from a measurement data file, you must obtain read access to the file using the MEASOPEN procedure. The data file can be associated with a currently active measurement, or it can contain data from an inactive measurement. Unlike MEASREAD, the caller specifies both start time and stop time.
loc output INT .EXT:ref:* is the destination buffer for the counter records. Counter records are written to the buffer in DDL record format. For the DDL record definitions for each entity type, see Chapter 3: Entities and Counters (page 136). bufsize input INT:value is the size in bytes of the destination buffer, loc. If bufsize is greater than 32,000, error 3203 (ERR^BADPARAMS) is returned. bytesret output INT .EXT:ref:1 is the number of bytes returned to the destination buffer, loc.
INT:value is the Measure product version when the data file was created. The value returned consists of two parts: bits 0:7 contain an alphabetic character (for example, D), and bits 8:15 contain a numeric value (for example, 30). Together they express the product version (for example, D30). all^recs input INT:value requests all interval records within the specified window if the value 1 is passed. If omitted or 0, only summary records are returned.
Int Int End; port = SQLIndex; IP^wildcard^flags = value; DEFINE IP^addr^wildcard^flag = IP^wildcard^flags.<0>#, port^wildcard^flag = IP^wildcard^flags.<1>#; or if^item input INT .
sqlname input is a 16-word array specifying the SQL run unit. sqlindex specifies the index number. These two fields are used in conjunction with the SQLSTMT entity. If passed, MEASREAD_DIFF_ returns the records with the specified run-unit and index numbers. sqlindex input is the SQL index. See sqlname. A value of -1 specifies all SQL indexes. NOTE: If the if^item^def structure is passed, it must first be initialized. To pass the structure without a value, IF^ITEM^DEF.Item^Name must be initialized to 0.
Example 3: IP address = 11.22.33.44, port 37 if^item = {.... IP^ADDR = {0xb, 0x16, 0x21, 0x2c, 0x0 }, PORT = 0x25, IP^WILDCARD^FLAGS = 0x0 } Example 4: IPv6 address = 1070::800:2006:4178, port 37 if^item = {.... IP^ADDR = IP^ADDR = {0x10, 0x70, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x8, 0x0, 0x20, 0x6, 0x41, 0x78}, PORT = 0x25 IP^WILDCARD^FLAGS = 0x0 } totals^num output INT(32) .EXT:ref:1 is a count of the records used to accumulate the summary record.
MEASREADACTIVE Reads data from a currently active counter. Before reading a counter, you must obtain a valid measurement number (generally, by using MEASMONSTATUS (page 412)). In D-series and Measure pre-G08 PVUs, MEASREADACTIVE cannot read DISCOPEN, DISKFILE, or PROCESSH counters. In Measure G08 and later PVUs, the MEASREADACTIVE procedure cannot read DISCOPEN or PROCESSH counters. To access these counter values, use the MEASREAD_DIFF_ procedure to read from the currently active measurement data file.
data space), entitydesc must identify exactly one entity. If necessary, use the appropriate operating system information procedures to obtain the names and numbers that let you uniquely identify the entity. You need not specify all the fields in an entity descriptor if a subset of the fields is sufficient to identify the data to retrieve. For example, if you identify a process by its CPU and PIN, you need not also specify the process name or program file name.
(Measure G11 and later PVUs) is the appropriate template version literal from MEASDDLS for the entity type requested in entitydesc if you want ZMS style external records. If omitted or passed as 0F, legacy style records are returned. If passed as -1F, the templateversion for the current release will be used, starting with the H06.15/J06.04 RVUs. Note that returned external records, in this case, may not match the counter record definitions with which the application was compiled.
measnum input INT:value is the measurement number. Use the measnum value returned by the MEASCONFIGURE procedure or find the measnum value using MEASMONSTATUS. entitydesc input INT:ref:* is the entity type and entity specification of the desired active counter record. Specify entitydesc as one of the descriptors listed in the contab description of the MEASCONFIGURE procedure.
are then examined. If they do not match, error 3402 (ERRM^CANNOTACCESS) or error 3404 (ERRM^NOTMEASURING) is returned. templateversion input FIXED:value (Measure G11 and later PVUs) is the appropriate template version literal from MEASDDLS for the entity type requested in entitydesc if you want ZMS style external records. If omitted or passed as 0F, legacy style records are returned. If passed as -1F, the templateversion for the current release will be used, starting with the H06.15/J06.04 RVUs.
entitydesc can contain wildcards. Using a wildcard descriptor along with LIST mode is the best way to get the records of all the instances of an entity configured on a processor. This procedure does not use string-based fields, such as OSS pathname or program file name, for descriptor matching. Such strings are ignored, if provided; if you need to use them, call the MEASREADACTIVE procedure instead. NOTE: entitydesc can be a DISKFILE, DISCOPEN, or FILE entity descriptor for ANSI SQL objects or partitions.
On output • If mode is READACTIVE_MANY_EXACT, context contains information about the matched entity. Passing this value back the next time you want data for the same entity will increase the performance of the request. • If mode is READACTIVE_MANY_CHANGED or READACTIVE_MANY_LIST, a context of 0F indicates that no more entities matching entitydesc are available for retrieval. -1F indicates an error return.
• For best performance, use the ZMS entity-template-version literal from the MEASDDLS file for the current RVU. Using an earlier template version or requesting legacy style records increases performance cost. • To use this interface, you must first start a measurement containing the entities to be accessed. Measure needs that measurement and the corresponding measnum to identify counter records that have changed.
contab output INT .EXT:ref:* is an array that defines the measurement configuration. For the format of the contab array, see MEASCONFIGURE (page 383). bufsize input INT:value is the size in bytes of the destination buffer, contab. If bufsize is greater than 32,000 bytes, error 3203 (ERR^BADPARAMS) is returned. To get the contab size returned in bytesret, bufsize must be at least 4 bytes.
This table lists the entity types and their literal values and numeric identifiers.
parameter is not passed or is passed as zero, the version defaults to the currently installed Measure PVU. max^ent input INT:value is the maximum number of entities the caller allocated for the entities and ctrspace arrays. This value should normally be specified as MAX^T. (MAX^T is a literal definition from the MEASDECS file that specifies the maximum entity type code.). If specified as less than MAX^T, it limits (truncates) the amount of entities or ctrspace array information returned.
contab, bufsize, bytesret These parameters are a group. That is, one cannot be passed without the other two, or error 3201 (ERR^MISSINGPARAM) is returned. • If OSS or SQL/MX journal segment construction is requested for a measurement, the status is always reported as under construction while the measurement is active. When the measurement is stopped, the setting continues to be reported as under construction until updating of the file is complete.
MEAS_SQL_MAP_INIT_ In Measure H02 and later PVUs, MEAS_SQL_MAP_INIT_ initiates the SQL/MX mapping session. This must be done if one of the following APIs is going to be used MEAS_GETDESCINFO_ MEASLISTENAME MEASLISTGNAME and there is no SQL journal in the indicated Measure data file, so resolution comes from the SQL/MX subsystem. error := MEAS_SQL_MAP_INIT_ ( connectionInfo ); ! i,o error INT is an error code indicating the outcome of the operation.
MEAS_SQL_MAP_STOP_ In Measure H02 and later PVUs, MEAS_SQL_MAP_STOP_ stops the SQL/MX mapping session. This should be done before an application process that has called MEAS_SQL_MAP_INIT_ exits. error := MEAS_SQL_MAP_STOP_ ( connectionInfo ); ! i The only error that can be returned is 3202 (ERR^BOUNDS). connectionInfo input INT(32) .EXT:ref:25 is allocated by the caller to be at least 100 bytes.
Usage Notes • ANSI SQL objects have independent name spaces and, if identical names exist in different name spaces, each instance of the name refers to a different object. • The name spaces used by Measure are TABLE, INDEX and MODULE. TABLE is implied by using the keywords CATALOG, SCHEMA and TABLE. INDEX is implied by using the keyword INDEX. MODULE is implied for SQLSTMT (in MEASCOM there is no MODULE keyword).
name input STRING .EXT:ref:* is a buffer that contains the input name in external format to be expanded to a fully qualified name. If name does not contain the catalog and/or schema name, the default catalog (def_catalog) and/or default schema (def_schema) name is used. In that case, if the necessary def_catalog and/or def_schema name is missing, an error is returned. Wildcards are permitted. A name space keyword is required. name_length input INT:value is the size in bytes of the input name.
MEAS_SQLNAME_SCAN_ In Measure H02 and later PVUs, MEAS_SQLNAME_SCAN_ parses a fully qualified, possibly wildcarded, ANSI SQL name in external format. This routine verifies the syntax of an ANSI SQL name. error := MEAS_SQLNAME_SCAN_ ( name ,name_length ); ! i ! i error INT is an error code indicating the outcome of the operation. Possible error codes include: Error Code Description 0 Successful completion. 3239 err^badformatsqlname The specified ANSI SQL names contained syntax errors.
else ! ANSI SQL MODULE name is OK MEASSTATUS Returns information about a currently active measurement. In Measure G09 and later PVUs, MEASSTATUS has a parameter for retrieving settings that reports the configuration of the journal segment functions.
entities output INT(32):ref:MAX^T+1 is an array that contains the number of entities of each entity type currently being measured. The array elements are in entity-type order, as this table shows. A value of 0 indicates the entity type was not included in the configuration.
settings output INT:ref:1 (Measure G09 and later PVUs) is an array of flags that identifies configuration attributes of the measurement represented by the data file: settings.0:8 Reserved for future use. settings.9 1 = Counter data records are suppressed in this file. settings.10:11 Reserved for future use. settings.12 1 = An SQL/MX journal segment is under construction for this file. settings.13 1 = An SQL/MX journal segment is present in this file. settings.
error := MEASWRITE_DIFF_( dfnum ,entitydesc ,volume^subvol ,[ from^time ] ,[ to^time ] ,[ timetol ] ,[ version ] ,[ all^recs ] ,[ zero^reports ] ,[ totals ] ,[ loadid ] ,[ if^item ] ,[ templateversion ] ); ! ! ! ! ! ! ! ! ! ! ! ! ! i i i i i i i i i i i i i dfnum input INT:value is the data file access number. Use the dfnum value returned by the MEASOPEN procedure. entitydesc input INT .EXT:ref:* is the entity type and entity specification of the desired counter records.
to^time input FIXED:value is the stop time of the measurement window. Specify to^time as a Julian date based on local civil time in four-word-by-one-microsecond format as provided by the system procedure JULIANTIMESTAMP. (See the Guardian Procedure Calls Reference Manual.) A value of -1 signifies the end of the measurement. If no to^time is specified, the default is -1.
INT .
is a value in the range 1 through 4, which designates comparison type: 1 Compare for equal. 2 Compare for unequal. 3 Compare for less-than. 4 Compare for greater-than. value input is a 32-bit integer to be used for comparison with the value in the given counter. value = n * 1000. For example, 17.09 should be stored as 17090. rate input determines whether a nonrated or rated value is used in the comparison: • 0 indicates a nonrated value (similar to REPORT RATE OFF).
ip^addr^wildcard^flags determines wildcard support for the ip^addr and port fields: <0> <1> 0 Match specified address. 1 Match any address. 0 Match specified port number. 1 Match any port number. The IP address in the data record appears in hex with 1 byte per IP component. Here are some examples of various ip^addr, port, and ip^wildcard^flags field specifications: Example 1: IP address = 11.22.33.44, match any port number if^item = {....
Usage Notes • If the longer IF^ITEM^G09^DEF template is used, the G09^format must be set (see the DEFINE example). If the G09^format is not set, part of the sqlname field is interpreted as sqlindex, and any records returned are likely to be incorrect. DEFINE G09^format = relation^operator.0#; DEFINE G09^rel^operator = relation^operator.13:15#; • You can use either template format when accessing Measure G09 data files.
5 Entity Descriptors This chapter describes the Measure entity descriptors. Like the entity specifications used in the command interface, entity descriptors describe an entity or set of entities used in the programmatic interface. You use entity descriptors to define the entities to measure and to specify the counter records to maintain for each entity. Entity descriptors are part of the configuration table (contab) array that you pass to the MEASCONFIGURE procedure before starting a measurement.
Table 6 Entity Descriptors and Type Values Section Descriptor Template Structure (Systems Running D-Series RVUs) Descriptor Template Structure (Systems Running G-Series and later RVUs) Type Literal Numeric Identifier CPU CPU^DESC CPU^DESC CPU^T 1 PROCESS PROCESS^DESC PROCESS^DESC PROCESS^OSS^DESC PROCESS^T 2 PROCESSH PROCESSH^DESC PROCESSH^DESC CODE^SPACE^DESC PROCESSH^OSS^DESC CODE^SPACE^OSS^DESC PROCESSH^T 3 USERDEF USERDEF^DESC COUNTER^DESC USERDEF^DESC COUNTER^DESC USERDEF^OSS^DES
NOTE: You can use measurement applications containing D-series descriptors to measure G-series entities if the descriptors specify all entities. For example, if an application uses the DEVICE^DESC descriptor and contains -1 or an asterisk in all identifier fields, you can use that application to measure G-series DEVICE entities. You need not modify the application to use the DEVICE^SVNET^DESC or DEVICE^SVNET^DESC^G05 descriptors.
CPU^DESC is the descriptor for CPU, TMF, or OSSCPU entities. Struct CPU^desc (*) Begin Int Type; Int Len; Int CPU^number; End; ! Entity type ! Byte length of this record type INT:value:1 is one of: Entity Type Literal Numeric Identifier CPU CPU^T 1 TMF TMF^T 14 OSSCPU OSSCPU^T 20 A positive type value includes the specified entities in the measurement configuration. A negative type value excludes the specified entities from measurement.
CTRL^DESC is the descriptor for CONTROLLER entities on systems running D-series RVUs. Struct Ctrl^desc (*) BEGIN Int Type; Int Len; Int CPU^number; Int Channel; Int Ctrl; Int Ctrl^type; End; -- controller descriptor structure ------- controller entity type byte length of the descriptor CPU number channel number controller number controller type type INT:value:1 is the literal value CTRL^T or the numeric identifier 18. A positive type value includes the specified entities in the measurement configuration.
DEVICE^DESC is the descriptor for DEVICE, DISC, LINE, NETLINE, or TERMINAL entities on systems running D-series RVUs. Struct Device^desc (*) Begin Int Type; ! Entity type Int Len; ! Byte length of this record Int CPU^number; Int Ctl; ! Controller number Int Unit; ! Unit number Int Device^name[0:11]; Int Channel; ! Channel number End; NOTE: RVUs.
channel-num INT:value:1 is the channel number of the device to be measured. DEVICE^CLIM^DESC is the descriptor for DEVICE or DISC entities connected to a system running an H-series or J-series RVU. DEVICE^CLIM^DESC includes a PLPT field, which is a structure that contains fields that uniquely identify devices connected by a CLIM. You can use DEVICE^SVNET^DESC or DEVICE^SVNET^DESC^G05 to describe other non-CLIM NonStop S-series devices.
For devices connected by an FCSA or CLIM, use 0 to indicate the X fabric. By convention, dual fabric controllers are specified by X. GMS is the physical location address (group, module, slot). The GMS structure consists of these fields: group INT(32):value:1 is the group number of the device to be measured. (The group corresponds to the physical enclosure.) Use the literal -1 to indicate all groups. module INT(32):value:1 is the module number of the device to be measured.
lun INT:value:1 is the logical unit number (lun) assigned to a disk within a CLIM that is to be measured. Not used for non-CLIM devices. Use the literal -1 to indicate all luns. partition INT:value:1 (H06.23/J06.12 and later RVUs) is the partition number assigned to a partition within a partitioned disk. Not used for non-CLIM devices or non-partitioned CLIMs. Valid only with DISC entity. target^id INT:value:1 is the SCSI port identifier of a non-CLIM device to be measured. Not used by CLIM devices.
String Config^name^s[0:63] End; = Config^name; type INT:value:1 is one of: Entity Type Literal Numeric Identifier DISC DISC^T 7 DEVICE IODEV^T 8 A positive type value includes the specified entities in the measurement configuration. A negative type value excludes the specified entities from measurement. For an example, see Specifying Entity Descriptors (page 376). len INT:value:1 is the length in bytes of this descriptor.
config^name INT:value:32 is the configuration name of the disk or device, as defined by the system configuration database. config^name can also be the configuration name of either a single ServerNet addressable controller (SAC) or an adapter that contains several ServerNet addressable controllers. In this case, the specification includes all devices of the requested type supported by that SAC or adapter. config^name must be null-terminated and null-filled.
len INT:value:1 is the length in bytes of this descriptor. len is 52 if the descriptor does not include the configuration name or 116 if the descriptor includes the configuration name. cpu^number INT:value:1 is the number of the CPU on which the device is configured. Use the literal -1 to indicate all CPUs. servernet INT:value:1 is the ServerNet fabric. This field identifies a specific path or set of paths to a supported physical device.
To save space in the configuration table, omit config^name from the DEVICE^SVNET^DESC^G05 structure. You must set the descriptor length (len) to reflect whether you are using config^name because Measure uses the descriptor length to determine the start of the next descriptor in the configuration table. To not use config^name, set len to 52 and move only the first 52 bytes of the descriptor to the configuration table.
DISKFILE^DESC is the descriptor for DISKFILE entities. For OSS file pathnames, use the template DISKFILE^OSS^DESC (page 465). Struct Diskfile^desc (*); Begin Int Type; Int Len; Int Filecode; Int Filetype; Int File^name[0:11]; End; -- Diskfile descriptor declaration ------ Diskfile entity type byte length of the descriptor file code file type as it says type INT:value:1 is the literal value DISKFILE^T or the numeric identifier 19.
DISKFILE^OSS^DESC is the descriptor for DISKFILE entities that supports OSS file pathnames. Struct Diskfile^OSS^desc (*) Begin Int Type; Int Len; Int Filecode; Int Filetype; Int File^name[0:11]; Struct File^name^MID( MID^def ); End; -- Diskfile descriptor declaration ------- Diskfile entity type byte length of the descriptor file code file type as it says Internal identifier for OSS pathname Definitions of the DISKFILE^OSS^DESC fields that differ from the DISKFILE^DESC descriptor follow.
pathid INT:value:12 is the internal format representation of the OSS file pathname. If the internal format representation was used to access the disk file, initialize Pathid[0] to -1 to indicate all PATHID values. crvsn INT:value:3 is the creation version serial number that identifies a unique instance of an OSS disk file. Initialize CRVSN[0] to -1 to indicate all CRVSN values. SQLMX^obj^desc INT:value:12 is the internal format representation of the ANSI SQL object definition.
FILE^OPEN^DESC is the descriptor for DISCOPEN or FILE entities and provides fields for describing EDIT files, object files, and temporary files. Separate descriptors are used to specify OSS file pathnames and ANSI SQL names: • To specify OSS file pathnames for the DISCOPEN and FILE entities, use the template FILE^OPEN^OSS^DESC (page 469). • To specify ANSI SQL names for the DISCOPEN and FILE entities, use the template FILE^OPEN^ANSI^DESC (page 471).
opener^cpu INT:value:1 is the number of the CPU on which the opener process is running. Use the literal -1 to indicate all CPUs. opener^pin INT:value:1 is the process identification number of the opener process. To indicate all PINs, use the literal -1. Use opener^cpu and opener^pin to identify the process that is accessing the file to be measured. file^number INT:value:1 is the file identification number of the file to be measured. To indicate all files, use the literal -1.
system^number INT:value:1 is the number of the system on which the opener process is running. Use the literal -1 to indicate all systems. system^name INT:value:4 is the name of the system on which the opener process is running. The system name must be left-justified and blank-filled. It must start with a backslash (\) followed by a system name in internal format. Use a backslash and an asterisk (\*) to indicate all systems.
End; End; Definitions of the FILE^OPEN^OSS^DESC fields that differ from the FILE^OPEN^DESC descriptor follow. For definitions of all the other fields, see FILE^OPEN^DESC (page 467). File^open^desc^fields are the fields in the FILE^OPEN^DESC descriptor. FILE^NAME^MID is a structure that contains the internal-format file name for the OSS pathname. pathid INT:value:12 is the internal format representation of the OSS file pathname.
crvsn INT:value:3 is the creation version serial number that identifies a unique instance of an OSS disk file. Initialize CRVSN[0] to -1 to indicate all CRVSN values. measflags INT:value:1 is a field for future use; initialize to 0. opener^pname INT:value:4 is the name of the opener process. The process name must be left-justified and blank-filled. It must start with a dollar sign ($) followed by a process name in internal format. Use a dollar sign followed by an asterisk ($*) to indicate all processes.
FILE^NAME^MID is a structure that contains the internal-format file name for the OSS pathname. STRUCT MID^def (*) begin Int Pathid[0:11]; Int Crvsn[0:2]; end; ! High level view of Measure ID pathid INT:value:12 is the internal format representation of the OSS file pathname. If the internal format representation was used to access the disk file, initialize Pathid[0] to -1 to indicate all PATHID values.
OPDISK^DESC is the descriptor for OPDISK entities on systems running D-series RVUs. Struct Opdisk^desc (*) Begin Int Type; Int Len; Int CPU^number; Int Ctl; Int Unit; Int Device^name[0:11]; Int Volume^name[0:11]; Int Channel; End; ! Entity type ! Byte length of this record ! Controller number ! Unit number ! Channel number type INT:value:1 is the literal OPDISK^T or the numeric identifier 17. A positive type value includes the specified entities in the measurement configuration.
PROCESS^DESC is the descriptor for PROCESS entities. For OSS file pathnames, use the template PROCESS^OSS^DESC (page 475). The PROCESS^DESC descriptor is also used for OSSNS entities. Struct Process^desc (*); Begin Int Type; ! Entity type Int Len; ! Byte length of record. In case of ! process it includes ! length of all the code-space ! descriptors and for userdef ! it includes length of all the counter ! descriptors.
program^file^name INT:value:12 is the object file that the process is executing. The name must be left-justified and blank-filled and must be in local internal file-name format. You can use asterisks (*) in the volume, subvolume, and file-name fields within file^name. However, the volume name must begin with a dollar sign ($). For example, to specify all program file names, use the value "$* * *". PROCESS^OSS^DESC is the descriptor for PROCESS entities that supports OSS file pathnames.
program^file^name INT:value:12 is the object file that the process is executing. The name must be left-justified and blank-filled and must be in local internal file-name format. You can use asterisks (*) in the volume, subvolume, and file-name fields within file^name. However, the volume name must begin with a dollar sign ($). For example, to specify all program file names, use the value "$* * *".
PROCESSH^DESC is the descriptor for PROCESSH entities. PROCESSH entity descriptors are variable in length and can contain CODE^SPACE^DESC (page 478) templates following the initial PROCESSH^DESC structure. For OSS file pathnames, use the PROCESS^OSS^DESC (page 475) and CODE^SPACE^OSS^DESC (page 480) templates. Struct Processh^desc (*); Begin Int Type; ! Entity type Int Len; ! Byte length of record.
program^file^name INT:value:12 is the name of the object file that the process to be measured is executing. The name must be left-justified, blank-filled, and in local internal file-name format. You can use asterisks (*) in the volume, subvolume, and file-name fields within file^name. However, the volume name must begin with a dollar sign ($). For example, to specify all program file names, use the value "$* * *". conftime FIXED:value:1 Specify conftime as 0F. For use by the Measure subsystem only.
UC 0 through 15 UL 0 through 15 SC 0 SL 0 through 31 code^ranges^file^name INT:value:12 is the name of the file containing the code to be measured. code^ranges^file^name must be a TNS object file, EDIT file, or Executable and Linking Format (ELF) file. The name must be left-justified, blank-filled, and in local internal format. If code^ranges^file^name is an object file, MEASFH examines the procedure names and addresses and adds each procedure to the configuration.
Definitions of the PROCESSH^OSS^DESC fields that differ from the PROCESS^OSS^DESC descriptor follow. For definitions of all the other fields, see PROCESS^OSS^DESC (page 475). process^oss^desc^fields are the fields in the PROCESS^OSS^DESC descriptor. len INT:value:1 is the length in bytes of this descriptor and any CODE^SPACE^OSS^DESC descriptors associated with it. len is 80 if the PROGRAM^FILE^NAME^MID structure is included but does not include the CODE^SPACE^DESC or CODE^SPACE^OSS^DESC extensions.
SQLPROC^DESC is the descriptor for SQLPROC entities. For OSS file pathnames, use SQLPROC^OSS^DESC (page 482). Struct Sqlproc^desc (*); Begin Int Type; ! Entity type Int Len; ! Byte length of record. Int CPU^number; Int PIN; Int Process^name[0:3]; Int Program^file^name[0:11] End; type INT:value:1 is the literal SQLPROC^T or the numeric identifier 15. A positive type value includes the specified entities in the measurement configuration. A negative type value excludes the specified entities from measurement.
SQLPROC^OSS^DESC is the descriptor for SQLPROC entities that supports OSS file pathnames. Struct Sqlproc^OSS^desc (*); Begin Int Type; ! Int Len; ! Int CPU^number; Int PIN; Int Process^name[0:3]; Int Program^file^name[0:11]; Struct Program^file^name^MID; ! Begin Int Pathid[0:11]; ! Int Crvsn[0:2]; ! End; End; Entity type Byte length of record.
SQLSTMT^DESC is the descriptor for SQLSTMT entities. For OSS file pathnames, use SQLSTMT^OSS^DESC (page 484). For ANSI SQL file names, use SQLSTMT^ANSI^DESC (page 485). NOTE: When the measurement is being configured, only the process part of the descriptor is used; when performing a listing, all the fields can be used.
run^unit INT:value:32 is the name of the procedure that contains the SQL statement to measure. To include all run units, the name of the procedure should be blank-filled. Do not use the asterisk (*) wild-card character. Not used for ANSI SQL. stmt^index INT:value:1 is the statement index number within the specified run unit (procedure). Not used for ANSI SQL. run^unit^s is the string representation of the run unit name.
SQLSTMT^ANSI^DESC is the descriptor for SQLSTMT entities that supports ANSI SQL file names. NOTE: When an ANSI SQL name is copied into the descriptor’s run^unit field, it should not be enclosed in single quotes. NOTE: When the measurement is being configured, only the process part of the descriptor is used; when performing a listing, all the fields can be used.
SVNET^DESC is the descriptor for SERVERNET entities on systems running G-series and later RVUs.
module INT(32):value:1 is the module number. To indicate all modules, use the literal -1. remote^cluster INT(32):value:1 redefines module is the ServerNet node number of the system with which the entity is communicating. To indicate all ServerNet cluster systems, use the literal -1. slot INT(32):value:1 is the slot number. To indicate all slots, use the literal -1. subdevice INT(32):value:1 is the subdevice number of the SAC to be measured. Use the literal -1 to indicate all SACs.
config^name INT:value:32 is the configuration name of the SAC to be measured, as defined by the system configuration database. config^name can also be the configuration name of an adapter that contains several ServerNet addressable controllers. In this case, the specification includes all devices of the requested type that are supported by that adapter. config^name must be null-terminated and null-filled. An asterisk in the first byte indicates all configuration names.
SYSTEM^DESC is the descriptor for CLUSTER entities (on D-series RVUs only) or SYSTEM entities. Struct System^desc (*) Begin Int Type; Int Len; Int LH^CPU; Int System^number; Int System^name[0:3]; End; ! Entity type ! Byte length of this record ! Line handlers cpu number (primary or backup) type INT:value:1 is one of: Entity Type Literal Numeric Identifier SYSTEM REMSYS^T 11 CLUSTER CLUSTER^T 12 A positive type value includes the specified entities in the measurement configuration.
USERDEF^DESC is the descriptor for USERDEF entities. USERDEF^DESC is used with the COUNTER^DESC. For OSS file pathnames, use USERDEF^OSS^DESC (page 491). Struct Userdef^desc (*); Begin Int Type; ! Entity type Int Len; ! Byte length of record. Int CPU^number; Int PIN; Int Process^name[0:3]; Int Program^file^name[0:11]; Int Numcounters; ! Number of counters to be installed End; ! A subsequent counter^desc is not needed for an Entitydesc of ! a previously installed entity.
COUNTER^DESC describes a user-defined counter specification. OSS file pathnames also use COUNTER^DESC. Struct Counter^desc (*); Begin Int Int Int End; Name[0:7]; Type; Index; ! ! ! ! ! Numcounters of these structs go here when used as a Contab descriptor User assigned counter name 1=accum, 2=busy, 3=queue 0:index counters will be allocated name INT:value:8 is a left-justified, blank-filled counter name.
Definitions of the USERDEF^OSS^DESC fields that differ from the USERDEF^DESC descriptor follow. For definitions of all the other USERDEF^OSS^DESC fields, see USERDEF^DESC (page 490). PROGRAM^FILE^NAME^MID is a structure that contains the internal-format file name for the OSS pathname. pathid INT:value:12 is the internal format representation of an OSS file pathname for the specified program file name. To indicate all PATHIDs, initialize PATHID[0] to -1.
WAN^DESC is the descriptor for LINE, NETLINE, or TERMINAL entities on systems running G-series and later RVUs.
device^name INT:value:12 is the device name of the line being measured. The name must be left-justified and blank-filled. It must start with a dollar sign ($) and be in internal format. To indicate all devices, use a dollar sign followed by an asterisk ($*). device^name^s is the string representation of the device name. device^type INT:value:1 is the device type of a specific subsystem, such as 61 (X.25), as reported in the SCF LISTDEV listing. To indicate all devices, use the literal -1.
A Error Messages The Measure subsystem produces error messages and error codes. The error messages generated by MEASCOM, the Measure command language interface, are described in this appendix. For descriptions of the error codes generated by Measure callable procedures, see Appendix B: Error Codes (page 518). Each message in this appendix is preceded by MEAS and the message number. If you encounter an error while using MEASCOM, enter: HELP error-number MEAS 2000 Comment.
Recovery Enter HELP error code number to determine the appropriate corrections, and retry the operation. (For information on ALLOCATESEGMENT, see the Guardian Procedure Calls Reference Manual.) MEAS 3003 WARNING. Measurement was not stopped due to reason listed above. Cause The STOP MEASUREMENT command failed at some point during its processing. Effect The reason for failure is listed as an error returned from the callable procedures.
Effect The request is ignored. Recovery Perform a LIST entity command to obtain a list of the desired records before trying to add to the plot. MEAS 3008 WARNING. Plot became full during ADD; there can be a maximum of 26 entities; some entities were not ADDed. Cause An attempt was made to add more than 26 items to a plot. Effect Some of the requested items were not added. Recovery Delete any unneeded items so there are 26 or fewer before adding the desired counters. MEAS 3009 WARNING.
Cause The ADD MEASUREMENT command was issued for a data file that was previously added, and the MEASFH process is still active. Effect The recent ADD MEASUREMENT command changes the current data file, but the MEASFH program file name remains unchanged. Recovery To use a different MEASFH process on the same data file during a session, use the DELETE MEASUREMENT command to delete the current MEASFH process, then reissue the ADD MEASUREMENT command to specify a new MEASFH process. MEAS 3013 WARNING.
MEAS 3016 WARNING. The item item-name is obsolete for the current datafile version; results of the operation may be unexpected. Cause The item name entered is no longer maintained in the data file being processed. Effect The command continues, but the operation might produce unexpected results. Recovery Use an item name that is valid for the PVU of the data file being processed. MEAS 3017 WARNING.
MEASCOM /SWAP $SYSTEM/ START MEASSUBSYS MEAS 3020 WARNING. No current data file; select a current data file using ADD. Cause The command executed required a current measurement data file to perform its function. Effect The requested operation is ignored. Recovery Perform an ADD MEASUREMENT on the data file desired. MEAS 3021 WARNING. Measurement not active; unable to perform request. Cause This command is valid only when the measurement is currently active. Effect The requested operation is ignored.
Recovery Generally, no action is necessary. Translation can be processed from the OSS file-system name space on the current system. If you need access to the OSS journal segment at a later time, use the DELETE MEASUREMENT and ADD MEASUREMENT commands to delete and add the Measure data file again after OSS journal segment construction is complete. MEAS 3025 WARNING. Default Catalog not set. Cause The SQLSCHEMA command has been issued without specifying a default SQL catalog.
Recovery Rework the nesting of your OBEY files so no more than four files are used. MEAS 3051 Expecting a name of nn characters or less; shorten name. Cause The string entered must contain, at most, the number of characters indicated by the error message. Effect The requested operation is ignored. Recovery Shorten the string appropriately and retry the command. MEAS 3052 File is not network-accessible; shorten volume name.
Recovery Correct the syntax according to HELP or the syntax in this manual. For command syntax descriptions, see Chapter 2: MEASCOM Commands (page 35). For descriptions of syntax for identifying entities to be measured, see Chapter 3: Entities and Counters (page 136). MEAS 3056 Subsystem error; please report to Tandem Computers Incorporated. Trap %n, CS = %n, P = %nnnnn, E = %nnnnn, L = %nnnnn, S = %nnnnn. Cause A trap was encountered by MEASCOM. Effect The requested operation is ignored.
Cause A permanent disk file was not specified in the command. Effect The requested operation is ignored. Recovery Ensure the name entered is of the form $a.b.c. MEAS 3061 The USERDEF for this counter was not found; ADD the USERDEF. Cause The process specification entered in the current ADD COUNTER command was not previously added by an ADD USERDEF command. Effect The requested operation is ignored. Recovery Check that the process specification is correct, or add the USERDEF and retry the operation.
Cause The file name given or expanded was remote, and the command required a local file name. Effect The requested operation is ignored. Recovery For a discussion of file names that are used with the command, see the description of the specific command in Chapter 2: MEASCOM Commands (page 35). Then specify a local file and retry the operation. MEAS 3066 The name specified in the IF condition (item-name) was not found; check the spelling of the name.
MEAS 3070 Both FOR and TO have been specified; delete one and reissue command. Cause Both the FOR and TO option were specified in the same command. These options are mutually exclusive. Effect The requested operation is ignored. Recovery Delete one of the options and repeat the command. MEAS 3071 item-name not found; check the spelling and/or the appropriate entity report. Cause The plot counter name entered was not found. Effect The requested operation is ignored.
Cause An attempt was made to add an item to an already full plot. Effect The requested operation is ignored. Recovery Delete some counters and then retry the ADD. MEAS 3075 Expecting a single alpha character; see HELP. Cause The plot character was not a single alphabetic character. Effect The requested operation is ignored. Recovery Enter only a single alphabetic character. For more information, enter HELP. MEAS 3079 Measurement space overflow; DELETE unneeded descriptors.
Cause When Measure expanded the command, a buffer became full. Effect The requested command is ignored. Recovery Break the command into several commands and execute the commands separately, or qualify any file names present so MEASCOM performs a minimum of expansion. MEAS 3084 The DISCOPEN and DISKFILE entities require the filename to be either a permanent or temporary disk filename (e.g. $a.b.* or $system.#*).
MEAS 3089 Expecting a number with 3 or less decimal digits. Cause The number entered has more than three decimal digits. Effect The requested operation is not executed. Recovery Reduce the number of decimal digits, then reenter the command. MEAS 3090 Incompatible PROCESSH records in the filename data file. Please use a proper version of MEASCOM. Cause The PROCESSH record format in the data file is incompatible with the MEASCOM currently running. Effect The requested operation is not executed.
Recovery Try another command. To retrieve records, use the LIST command. MEAS 3095 Command option "list-option" not allowed with LISTALL command when STRUCTURED format is specified. Cause Illegal list-option was specified for the LISTALL command. Effect Command aborts; specified option is not satisfied. Recovery Retry the command, and do not include the list-option causing the error. MEAS 3096 Command option "TOLERANCE OFF" not allowed with LIST command.
Cause The line that Measure is attempting to be reexecute is not in the HISTORY buffer. Effect No command is executed. Recovery Perform a HISTORY command to make sure the command is still in the HISTORY buffer. Otherwise you must reenter the command in full. MEAS 3100 The command is too long, the maximum number of characters an effective record can contain is 2100 (nnnn were entered); try to enter multiple commands or remove embedded spaces as possible to shorten the command(s) entered.
Cause The program file indicated did not read the startup message sent due to the file-system error indicated. Effect The command is not executed. Recovery Correct the cause of the file-system error and retry the command. MEAS 3104 SCALE-FROM must be less than SCALE-TO; make sure the values are correct before retrying the LIST PLOT command. Cause The effective value for the SCALE-FROM attribute was greater than or equal to the SCALE-TO attribute of the PLOT object. Effect The command is not executed.
Recovery Use a counter name that appears within the entity report. MEAS 3108 item-name is not a unique item-name; do a HELP entity-type COUNTERS (e.g. HELP CPU COUNTERS) to determine the number of characters required to make the abbreviation unique. Cause The item name entered matched more than one item name. Effect The command is not executed. Recovery Add more characters to the name entered.
Cause You are using a G-series RVU, but the version of the data file precedes F40.00. Effect The command is not executed. Recovery To analyze a pre-F40.00 data file, use a pre-F40.00 version of MEASCOM that matches the version of the data file as well as the appropriate MEASFH. MEAS 3113 Unable to open data file. Cause MEASFH was unable to open the data file because there was too much data for the system to handle (file-system error 21). Effect The command is not executed. Recovery Retry the command.
Recovery If the request is part of an ADD data file that contains an OSS journal segment, you can DELETE the data file and attempt translation from the current system if the system is the same as that of the ADD data file. MEAS 3117 Unable to translate an ANSI SQL name. Cause The requested translation was not found in the ANSI SQL name space or in the current ANSI SQL journal segment. Effect The command is not executed.
MEAS 3123 Allocate segment error nnnn on work segment. Cause An error was encountered and returned from ALLOCATESEGMENT. The MEASCOM primary work segment has reached maximum capacity and cannot be extended. Effect The requested operation is ignored. Recovery Change the LIST command to retrieve a smaller subset of records. MEAS 3124 Partition number is not allowed for this command and entity.
Recovery See your system manager. MEAS 3176 System option number is not T2035; see your system manager. Cause An attempt was made to run MEASCOM with an option number other than T2035. Effect MEASCOM stops abruptly. Recovery See your system manager. MEAS 3177 Guardian version is not B30 or greater; see your system manager. Cause The NonStop operating system PVU must be B30 or later. Effect MEASCOM stops abruptly. Recovery See your system manager.
B Error Codes The Measure subsystem produces error messages and error codes. The error codes generated by Measure callable procedures are described in this appendix. For descriptions of the error messages generated by MEASCOM, the Measure command interface, see Appendix A: Error Messages (page 495). 3200 (ERR^BADMEASCB) Cause The MEASCB was not set to -1s on the initial procedure call or its contents were modified after being set by Measure. Effect The requested operation aborts.
Cause The buffer size was too small. Effect The requested operation aborts. Recovery Correct the buffer size. The buffer size needed is contained in the bytesret (byte size of record returned) parameter. If the MEASREADACTIVE procedure encountered this error and requires a buffer larger than 32 KB, use the MEAS_READACTIVE_ procedure. 3205 (ERR^NOTENOUGHSTACK) Cause Not enough data-stack space was available to send or receive the buffer. Effect The requested operation aborts.
Recovery Move either the MEASFH program file or the data file so they are both on the same node. Only read access is allowed for remote data files. 3209 (ERR^REMOTEDFILE) Cause The data file was not on the same system as the calling process. Effect The requested operation aborts. Recovery Because write access is not allowed for remote data files, use a data file on the same system as the calling process.
Effect The requested operation aborts. Recovery Use the MEASMONCONTROL procedure, indicating the CPU number of the missing MEASCTL. 3214 (ERR^MEMLOCKFAILURE) Cause Insufficient physical memory is available. Effect The requested operation aborts. Recovery Redistribute your processes to other processors or install more memory. 3215 (ERR^LINK) Cause A call to LINK failed. Effect The requested operation aborts. Recovery Retry the operation.
Effect The requested operation aborts. Recovery Verify that the Measure subsystem is not running prior to calling MEASMONCONTROL with a start subsystem request. 3219 (ERR^NOMEASMON) Cause The Measure subsystem was stopped. Effect The requested operation aborts. Recovery Verify that the Measure subsystem is running prior to calling this procedure. 3220 (ERR^UDCNOTPRESENT) Cause A user-defined counter was not installed and cannot be bumped. Effect The requested operation aborts.
Effect The requested operation aborts. Recovery Check that the entity type is valid and that it is supported in the RVU of the data file being processed. 3224 (ERR^INCONSISTENCY) Cause MEASCTL internal consistency check failed. Effect SQL activities are not measured. Recovery Contact your HP representative. 3225 (ERR^INCOMPLETEMSG) Cause The callable procedure was unable to complete a message to the MEASCTL process. Effect The requested operation was ignored. Recovery Try the operation again.
Effect The requested operation aborted. Recovery Check that the MEASFH object file exists and is properly secured. 3229 (ERR^MEASFH^OPEN) Cause An error was returned from File_Open_. Effect Unable to open the MEASFH process. The requested operation aborted. Recovery Check that the MEASFH object file exists and is properly secured. 3230 (ERR^MEASFH^SWAPFILE) Cause The SWAPVOL value specified was in error. Effect The MEASFH process or swap file is not created; the requested operation aborts.
Effect The command is not executed. Recovery Ensure that the OSS file and directory names you specify in an entity specification are translatable at the time of the measurement. 3234 (ERR^CRVSNNOTSPECIFIED) Cause You called MEASLISTPNAME without specifying the CRVSN qualifier. Effect Translation of the Guardian file name to an OSS file pathname is attempted, but the reported name might be incorrect if the file name specified is a transient entity. Recovery Include the CRVSN value in the MEASLISTPNAME call.
Effect The requested operation is aborted. Recovery Only fully qualified OSS file pathnames are allowed in DISKFILE entity specification. If you need transient file activity or specific sets of files, use Guardian file naming. For example: $DATA.*.*, or $*.ZYQ00000.*. 3238 (ERR^UNKNOWNSQLNAME) Cause The specified ANSI SQL name could not be translated, or the specified Guardian name could not be translated to an ANSI SQL name. Effect The requested operation is aborted.
a segment and that the MEASOPEN request did not return error 3240 (ERR^BUILDINGSQLJOURNAL). 3242 (ERR^NOT^IMPLEMENTED) Cause A request was made for Measure functionality that does not exist in this version of Measure. Effect The requested Measure function is not performed. Recovery Upgrade your version of Measure to the required level for the specified function. 3243 ERR^NOSPACEOSSJOURNAL Cause Disc space was not sufficient or segment allocation failed for building an OSS or ANSI SQL journal segment.
Cause The resource specified as the target of a callable function is not allocated. Effect The operation is not performed. Recovery Call the procedure that allocates the resource before you call any procedure that requires the resource.
Recovery Purge the file and create a new file with a file code of 175. 3253 (ERR^NOTDISCFILE) Cause The data file was not a disk file. Effect The requested operation aborts. Recovery Copy the measurement data file from tape to disk prior to read access. 3254 (ERR^WRONGVERSION) Cause The MEASFH version did not match the data file version. Effect The requested operation aborts. Recovery Specify the MEASFH program file name from the RVU used to generate the data file.
Recovery Try reducing the number of items to be LISTed by requesting specific items rather than using the asterisk (*) to list all items. For the PROCESSH entity, try specifying if code-range > 0 in the LIST command. 3258 (ERR^CSEGOVERFLOW) Cause MEASFH's external counter record segment overflowed. Effect The requested operation aborts. Recovery Try reducing the number of items to be LISTed by requesting specific items rather than using the asterisk (*) to list all items.
Effect The requested operation aborts. Recovery Inspect the measurement CONTAB parameter containing the entity descriptors to be measured. Check for an entity descriptor that includes the current descriptor and verify that this entity was alive during the measurement. 3263 (ERR^CRFILETYPE) Cause The PROCESSH code-range file was not a disk file. Effect The requested operation aborts. Recovery Move the code-range file to disk.
Cause The PROCESSH code-range file did not specify any code ranges, or the existing code ranges do not occur in the specified code space. Effect The requested operation aborts. Recovery Add at least one code range to the file or correct the code space specification and retry the operation. 3268 (ERR^TOOMANYCRS) Cause The PROCESSH code-range file specified more than 500 code ranges. Effect The requested operation aborts.
Recovery Reconfigure the PROCESSH measurement by reducing the number of code ranges (or taking out some code spaces). 3272 (ERR^INVALID^TAPE^OP) Cause An invalid request was made on a tape data file. Effect The requested operation aborts. Recovery The request can only be made on a disk file. Use FUP COPY to copy the measurement data from the tape to a disk file of code 175 and then retry the operation. 3273 (ERR^TRAP^OCCURRED) Cause The MEASFH process has trapped out.
3276 (ERR^MEAS^NEVER^STARTED) Cause Either the measurement was not started correctly on this data file (and error 3410 (ERRM^INSTALL) was returned by MEASMON), or START time was not written to the data file. Effect The data file does not contain valid data. Further operations on the data file could result in error. Recovery Try to start another measurement. Check the EMS log for error 3410 (ERRM^INSTALL). Use a data file big enough that the START time can be written to the data file.
Effect No records were written to the structured file. Recovery Remove the existing structured file and let MEASFH create it, or use a different subvolume. If using MEASCOM, use the VOLUME command. Be sure to regenerate the proper DDL dictionary. 3281 (ERR^FH^CTR^NAME^NOT^FOUND) Cause The counter name passed in the IF^ITEM structure was not found. Either the name is misspelled or it does not exist in the PVU being processed. Effect The operation was not performed. Recovery Pass a valid counter name.
Cause The MEASFH process was not able to read the COFF file further. Effect The operation was not performed. Recovery Designate a proper COFF file and retry the MEASCOM ADD PROCESSH command. 3286 (ERR^EDIT^NOT^ASCEND) Cause The MEASFH process was not able to perform the requested operation because the native address in the edit file was not in ascending order. Effect The requested operation aborts. Recovery Make changes in the edit file and retry the command.
3290 (ERR^VERSION^NOT^SUPPORTED) Cause The Measure PVU is not supported. Effect The data file is not read. Recovery Supply a valid Measure PVU and retry the operation. 3291 (ERR^NO^EGINFO) Cause MEASFH was not able to read the ELF file further. Effect The operation was not performed. Recovery The ELF file is partially linked. Pass the ELF file through the Native Mode Linker, NLD without the -r option, and then retry the operation with the output of the linker process.
Cause Two files were specified for configuring the same code space range of the same object in a measurement configuration. MEASFH could not complete the measurement configuration because of the invalid specification. Effect The operation was not performed. Recovery Remove one of the two files from the specification. 3295 (ERR^SQL^API^INTERNAL) Cause The MEASURE API internal check failed due to insufficient buffer space. Effect The ANSI SQL name operation could not be performed.
3299 (ERR^LCT^CHANGE) Cause The system time was changed during the measurement. Typically this error occurs at Daylight Savings Time transitions. Effect Post processing results can be incorrect. Recovery No recovery is possible. Run the measurement again. 3300 (ERRM^INVALIDMSG) Cause The type or length of the request was invalid. Effect The requested operation aborts. Recovery Correct the type or length and retry the request.
3304 (ERRM^INTERNAL) Cause An internal error occurred. Effect The requested operation aborts. Recovery Contact your HP representative. 3305 (ERRM^DATAFILEINUSE) Cause An attempt was made to use a data file that was already being used by another measurement. Effect The requested operation aborts. Recovery Specify another data file. 3306 (ERRM^MEASTABLEFULL) Cause An attempt was made to configure more than 64 measurements. Effect The requested operation aborts.
Cause In a control-measurement request, the stop time was less than the start time, an attempt was made to change the start time of an active measurement, or there was some other invalid start or stop time request. Effect The requested operation aborts. Recovery Correct the start and stop time values and retry the request. 3310 (ERRM^INVALIDDATAFILE) Cause An incorrect data file type was specified. Effect The requested operation aborts.
Cause The caller process has already opened MEASMON. Effect The requested operation aborts. Recovery The MEASMON process can be opened only once by the caller. It is already open. Check the MEASCB structure passed to the callable procedure. The caller should not have modified it. 3315 (ERRM^START^MEASSUBSYS) Cause The backup for MEASMON did not start, one or more executing CPUs does not contain an active MEASCTL or MEASIP process, or both. Effect The MEASURE subsystem is not fully operational.
3319 (ERRM^NOBINADDRSPACE) Cause The pool space for code-range addresses is full in the MEASCTL process. Effect The measurement does not run. Recovery Stop some measurements and then try to start your measurement again. 3320 (ERRM^BACKUP^DATAFILEOPENFAILED) Cause An attempt by backup MEASMON to open the data file failed. Effect The requested operation aborts. Recovery Check the file system error number and take appropriate action.
3403 (ERRM^ILLEGALDESC) Cause The entity descriptor in the request was invalid. Effect The reply does not contain any data. Recovery Correct the entity descriptor and retry the request. 3404 (ERRM^NOTMEASURING) Cause The measurement number (MEASNUM) parameter did not measure the entity, or the entity does not exist. Effect The reply does not contain any data. Recovery Check that the entity exists in the measurement CONTAB parameter.
3408 (ERRM^MEASNUMINUSE) Cause An attempt was made to start an already active measurement. Effect The operation is aborted. Recovery Contact your HP representative. 3409 (ERRM^NOCONTABSPACE) Cause The MEASCTL CONTAB pool was full. Effect The operation is aborted. Recovery Wait until some other measurements have stopped and then retry the operation. 3410 (ERRM^INSTALL) Cause A resource problem occurred with MEASCTL. Effect The operation is aborted.
3413 (ERRM^DFFAILED) Cause Write to the current data file failed. Effect Measurement is aborted. Recovery Restart the current measurement. 3414 (ERRM^LOCKMSGCTRLBUF) Cause The call to the LOCKMEMORY procedure failed. Effect Measurement is aborted. Recovery Retry the request. If problem persists, notify your system operator. 3415 (ERRM^SQLSPACEOVERFLOW) Cause SQL Control Block space is not available. Effect Measurement is aborted. Recovery Restart the SQL measurement.
3421 (ERRM^NOCOUNTERSPACE) Cause The MEASCTL process in the specified CPU ran out of internal counter space while starting a measurement. Effect Some entities are not measured. Recovery For all measurements in the system, reduce the amount of transient entities to be measured by configuring only needed entities. For example, avoid using wild-card (*) specifications in the configurations. Use the DELETE command to exclude entities that are not needed.
C Subsystem Files This list describes the files that make up the Measure performance monitor: MEASCOM process MEASCOM is the Measure subsystem command interpreter. For descriptions of the commands, see Chapter 2: MEASCOM Commands (page 35). Typically started from TACL, a MEASCOM process lets the user who started it carry on an interactive session with the Measure subsystem. A super-group user (member of user group 255) can use MEASCOM to start the Measure subsystem.
Record template names use legacy style names (for example, PROCESS rather than ZMSPROC). Fields that have been combined are referenced by redefines. MEASDECS file The MEASDECS file contains the structure declarations and literal value definitions used in the Measure callable procedures. MEASDECS is stored in $SYSTEM.SYSnn. See C and C++ Language Usage Notes. MEASCHMA file The MEASCHMA file contains the structure declarations for C and TAL. Its output is similar to the output of the MEASDECS file.
D Measure Data File Tool (MEASFT) The Measure Data File Tool (MEASFT) lets you split a measurement data file into multiple data files for consumption and display by MEASCOM or through Measure APIs. MEASFT is especially useful in the case of the very large data files that can result from measuring large numbers of entities or measuring at very small intervals. MEASFT is part of H06.07 and later RVUs, but you can use it to split measurement data files created by Measure G10 or later G-series PVUs.
Example • In H06.25/J06.14 and later RVUs, the INFO command display includes a split history record of the data file. The following example is a display of a data file (MDFD00) that was the result of splitting MDFALLI by TIME into 1 hour PARTS, followed by a split of MDFA01 by CPU, followed by a split of MDFB01 by ENTITY, followed by a split by TIME to extract 15 minutes of data starting at 11:30.
INFO MDATA5 $PERF.CAPAPP.MDATA5 (H01) File Size 133980160 bytes Data file usage (in bytes) by Entity Entity -----CPU DISC DISCOPEN DISKFILE FILE NETLINE OSSCPU OSSNS PROCESS SERVERNET SQLPROC SQLSTMT TMF -----Total Space ---------133056 2031486 25678814 17085880 35348390 173644 155626 23394 41238946 10585470 9956 82328 29766 ---------132576756 % -----0.10 1.52 19.17 12.75 26.38 0.13 0.12 0.02 30.78 7.90 0.01 0.06 0.02 -----98.
If you specify an asterisk (*), MEASFT creates the output data files with names of the form ZMSRFnnn, beginning with ZMSRF000. The number of output data files created depends upon split-criteria. If split-criteria is not specified, MEASFT behaves as described in Usage Notes (page 555). If you specify a string of characters ending in an asterisk, MEASFT regards the value as a file name prefix and appends a number, starting with "00" for the first file.
OPDISK OSSCPU OSSNS PROCESS PROCESSH SERVERNET SQLPROC SQLSTMT SYSTEM TERMINAL TMF USERDEF The last entity-listn can also be an asterisk ( * ), which instructs MEASFT to put data for all the remaining entities into the last output data file. If only entity-list1 is specified, a single output data file is created. BY CPU followed by an asterisk (*) causes MEASFT to create one output data file per processor number present in the data file.
If you omit date, the date that the measurement was started is used. time-of-day is hh:mm[:ss ] where hh is the hour (0 through 23). mm is minutes (0 through 59). ss is seconds (0 through 59). The syntax for duration is one of: n SECOND[S] n MINUTE[S] n HOUR[S] where n is an integer in the range 1 through 32767. BY TIME, INTO number causes MEASFT to create multiple output data files that contains entities that existed in the input data file.
• The “Time Period” column of the SPLIT display shows the time period of each created TIME-split data file. All the data records in the file should have a timestamp within that period. For interval data files, the effective time period is wider by one half of the interval on either side of the displayed time period. • The number in the “%” column of the SPLIT display is the size of the output data file relative to the input data file.
Examples • Extract the records from data file, MDFALLI, that occur during the 15 minutes after 16 Mar 2012, 11:15, and write them into a new data file, MIN15. The “Time period” line is the time period of the input data file being split. The “Time Period” column is the time period of each created time-split file. The number in the “%” column is the size of the output data file relative to the input data file. $PERF.
----------------------------------- ---------$PERF.CAPAPP.ZMSRF000 133842 $PERF.CAPAPP.ZMSRF001 41239732 $PERF.CAPAPP.ZMSRF004 35349176 $PERF.CAPAPP.ZMSRF005 25679600 $PERF.CAPAPP.ZMSRF006 2032272 $PERF.CAPAPP.ZMSRF009 174430 $PERF.CAPAPP.ZMSRF013 30552 $PERF.CAPAPP.ZMSRF014 10742 $PERF.CAPAPP.ZMSRF015 83114 $PERF.CAPAPP.ZMSRF017 10586256 $PERF.CAPAPP.ZMSRF018 17086666 $PERF.CAPAPP.ZMSRF019 156412 $PERF.CAPAPP.ZMSRF020 24180 • -----0.10 30.78 26.38 19.17 1.52 0.13 0.02 0.01 0.06 7.90 12.75 0.12 0.
HELP Command The HELP command displays help information for the published commands. Syntax HELP | empty Enter "MEASFT HELP" or just "MEASFT" to display help text for the MEASFT command. MEASFT Error and Warning Messages Any of the following messages can appear on the console when you run MEASFT.
Recovery It is too late to recover the data in the overwritten file. In the future, specify an output file name that doesn't conflict with existing files. MEASFT 2002 (WARN_NO_XRTIMES_AT_EOF) No Stop Measurement record, file possibly truncated/corrupted. Cause There is no Stop Measurement record in the input data file. The file is either a truncated copy or the file was corrupted. Effect The INFO or SPLIT command continues.
Cause None of the records in the Measure data file satisfied any of the split criteria so no output files were created. Effect The SPLIT command continues. Recovery Determine the cause of the missing records and decide whether this is a significant issue. MEASFT 2007 (WARN_ENTITY_HAS_NO_DATA) No data for this specified entity, entity Cause No data records were encountered for an entity that was specified in the split criteria. Effect The SPLIT command continues.
Cause No data records were encountered for the last output file, whose cpu-list was specified as a wildcard (*) in the SPLIT command. Effect The SPLIT command continues. The specified output file is not created. Recovery Determine the reason for the missing data and decide whether this is a significant issue. MEASFT 2011 (WARN_SPLIT_FILE_NOT_CREATED_2) An output file was not created due to lack of data.
Effect The SPLIT or INFO command is not executed. MEASFT terminates in error. Recovery Use the displayed File System error number, nn, to determine why the specified Measure data file name could not be resolved. Bad file name syntax is the most likely cause. MEASFT 3001 (ERR_FNAME_UNRESOLVE) Measfile Name UnResolution failed, File System Error = #nn Cause A Measure data file name was specified that could not be unresolved by FILENAME_UNRESOLVE_. Effect The SPLIT or INFO command is not executed.
Recovery Use the displayed File System error number, nn, to determine why the specified output file could not be opened. This error is unlikely to occur because the output file has just been created successfully. MEASFT 3007 (ERR_FILE_SETPOSITION) FILE_SETPOSITION_ failed, File System Error = #nn Cause The Measure data file could not be repositioned at the beginning of the file or near the end of the file if the data file has a journal section. Effect The SPLIT command does not complete.
MEASFT 3011 (ERR_READX_SPLIT_FILE) Read of output file, datafile, failed, File System Error = #nn Cause MEASFT encountered an error in reading the first record of the output file, datafile, so that it could update the journal pointer(s) in it. Effect The SPLIT command does not complete. MEASFT terminates in error. The created output files will not have correct journal pointers. Recovery Use the displayed File System error number, nn, to determine what caused the error.
Recovery Use the displayed File System error number, nn, to determine what caused the error. MEASFT 3017 (ERR_FILE_BAD_TYPE) Measure file not unstructured, file type encountered #n Cause The specified Measure data file is not unstructured. Effect The SPLIT or INFO command is not executed. MEASFT terminates in error. Recovery Measure data files are unstructured files, whereas the one specified has file type n. Enter the correct file name.
Cause The output file, datafile, could not be purged. MEASFT overwrites existing Measure data files if an output file with the same name is specified. Effect The SPLIT command does not complete. MEASFT terminates in error. Some output files may have been created. Recovery Use the displayed File System error number, nn, to determine why the specified output file could not be purged. MEASFT 3022 (ERR_SEG_ALLOCATE_) SEGMENT_ALLOCATE_ error #nn. Cause MEASFT could not allocate enough memory to run.
Recovery Either the Measure data file has been corrupted or MEASFT has a defect. Report this message to your HP representative. MEASFT 3027 (ERR_BAD_RECORD_STATE) Bad record state in record #number Cause MEASFT encountered a record state it did not expect in record number. Effect The command does not complete. MEASFT terminates in error. Some output files may have been created. Recovery Either the Measure data file has been corrupted or MEASFT has a defect. Report this message to your HP representative.
Cause MEASFT did not recognize the first record in the Measure data file. Effect The command is not executed. MEASFT terminates in error. Recovery Either the Measure data file has been corrupted or MEASFT has a defect. Report this message to your HP representative. MEASFT 3033 (ERR_REC_LENGTH_COMP_ERR) Record length could not be computed, record #nnnnn Cause MEASFT could not compute the length of record number nnnnn because of either a MEASFT defect or a corrupt Measure data file.
Recovery Verify that the specified Measure data file is correct. MEASFT 3038 (ERR_NO_XRTIMES_RECORD) No Start Measurement record in the input data file. Cause The SPLIT BY TIME command was issued, but there is no Start Measurement record in the input data file. Effect The SPLIT command does not complete. MEASFT terminates in error. Recovery Either the Measure data file has been corrupted or MEASFT has a defect. Report this message to your HP representative.
Effect The command is not executed. MEASFT terminates in error. Recovery Enter the Measure data file name. MEASFT 3202 (ERR_PARSE_FILENAME_TOO_LONG) filename parameter too long. Cause The Measure data file name entered in the command line is too long. Effect The command is not executed. MEASFT terminates in error. Recovery Enter the correct Measure data file name MEASFT 3230 (ERR_PARSE_SPLIT_NAME_TOO_LONG) SPLIT name parameter too long. Cause The indicated output file name is too long.
Effect The SPLIT command does not complete. MEASFT terminates in error. Recovery Use the displayed File System error number, nn, to determine what caused the error. This error is unlikely to occur because the output file name has already been resolved by FILENAME_RESOLVE_. MEASFT 3234 (ERR_PARSE_SPLIT_RPAREN_MISSING) SPLIT file specs right parenthesis missing. Cause The SPLIT file specifications entry starts with a left parenthesis and is missing a right parenthesis.
Effect The SPLIT command does not complete. MEASFT terminates in error. Recovery Correct the SPLIT files specifications and/or split criteria syntax so that they specify the same number of output files. MEASFT 3239 (ERR_PARSE_SPLIT_BAD_CPU_NUM2) SPLIT criteria CPU number out of range. Cause The CPU specified in the SPLIT command split criteria is either out of range (0 to 15) or is not numeric. Effect The SPLIT command does not complete. MEASFT terminates in error. Recovery Correct the invalid CPU number.
Cause If wildcard (*) is specified in the SPLIT command split criteria, then the file set must also be a wildcard. Effect The SPLIT command does not complete. MEASFT terminates in error. Recovery Correct the SPLIT split criteria so that both the file specifications and split criteria have wildcards or neither has a wildcard. MEASFT 3244 (ERR_PARSE_SPLIT_SC_EMPTY) SPLIT criteria empty. Cause Nothing was specified between the parentheses in the SPLIT command split criteria.
MEASFT 3248 (ERR_PARSE_SPLIT_ONE_FILE) SPLIT criteria required (no wildcard) for a single output file. Cause Only one output file name was specified but no split criterion was specified. Effect The SPLIT command is not executed. MEASFT terminates in error. Recovery Add split criteria to the specified SPLIT command. MEASFT 3249 (ERR_PARSE_FS_SYNTAX) SPLIT out-file-set syntax error. Cause Multiple output file names were specified in the file specifications with no parentheses.
MEASFT 3253 (ERR_PARSE_SPLIT_MISSING_DEL) SPLIT file specs missing a delimiter (',' or ')'). Cause A delimiter (either a comma or a right parenthesis) is missing in the SPLIT file specifications. Effect The SPLIT command is not executed. MEASFT terminates in error. Recovery Correct the SPLIT file specifications syntax. MEASFT 3254 (ERR_PARSE_SPLIT_TIME_KW_EXP) SPLIT 'FROM', 'TO', 'FOR' or INTO keyword expected. Cause The SPLIT BY TIME command must be followed by the FROM, TO, FOR or INTO keyword.
Effect The SPLIT command does not complete. MEASFT terminates in error. Recovery Correct the SPLIT BY TIME FROM and/or TO time values. MEASFT 3258 (ERR_PARSE_SPLIT_BAD_FOR_TIME) SPLIT FOR time value not valid. Cause The SPLIT BY TIME FOR duration value is not valid. The duration syntax is: number {SECOND[S] | MINUTES[S] | HOUR[S] } where number can be 1 to 32767 and it must be followed by SECOND, MINUTE, or HOUR. Effect The SPLIT command does not complete.
MEASFT 3262 (ERR_PARSE_SPLIT_TIME_WC) SPLIT, BY TIME * command not allowed. Cause The wildcard is not allowed in the SPLIT BY TIME command. Effect The SPLIT command does not complete. MEASFT terminates in error. Recovery Correct the SPLIT BY TIME syntax. MEASFT 3300 (ERR_CANNOT_SPLIT) The specified measure file cannot be split by entity, try splitting by CPU. Cause The Measure data file contains data records for only one entity, so it cannot be split by entity.
MEASFT 4002 (ERR_NOMEMORY_RECBUF) Not enough memory, bytes remaining = nnnn Cause MEASFT ran out of memory and reported nnnn bytes left. This condition is a MEASFT defect. Effect The SPLIT command is not executed. MEASFT abends. Recovery Contact your HP representative. MEASFT 4003 (ERR_NOMEMORY_SPLITBUF) Not enough memory for output buffer, bytes remaining = nnnn Cause MEASFT ran out of memory and reported nnnn bytes left. This condition is a MEASFT defect. Effect The SPLIT command is not executed.
Recovery Contact your HP representative. MEASFT 4010 (ERR_BAD_MF_CONTEXT) Bad Measfile_context #nn Cause MEASFT internal logic error. Effect The command is not executed. MEASFT abends. Recovery Contact your HP representative. MEASFT 4012 (ERR_PARSE_KW) Internal error, bad keyword index = #nn Cause MEASFT internal logic error in parsing the command line. This condition is a MEASFT defect. Effect The command is not executed. MEASFT abends. Recovery Contact your HP representative.
Recovery Contact your HP representative. MEASFT 4016 (ERR_NOMEMORY_TIME128) Not enough memory for TIME128 array, bytes remaining = #nn. Cause Internal error in memory allocation. This is a MEASFT defect. Effect The SPLIT command does not complete. MEASFT terminates in error. Recovery Report this message to your HP representative, citing the number, nn, displayed in the message.
Index A Abbreviations in commands, 38 See also HELP command , 29 ABLKS-INUSE-MAX counter (DISC entity), 202 ABLKS-INUSE-QTIME counter (DISC entity), 202 ABORT-TRANS counter (PROCESS entity), 299 ABORTING-TRANS counter (TMF entity), 363 ABS-SEGS-END counter (PROCESS entity), 299 ABS-SEGS-QLEN-MAX counter (PROCESS entity), 308 ABS-SEGS-QTIME counter (PROCESS entity), 299 ABS-SEGS-START counter (PROCESS entity), 300 ACCEL-BUSY-SAMPLES counter ADD PLOT command, 48 PROCESSH entity, 320 ACCEL-BUSY-TIME counter CP
BLOCK-SIZE counter (DISC entity), 196 BLOCK-SPLITS counter DISC entity, 204 DISCOPEN entity, 213 DISKFILE entity, 222 BLOCKS-INUSE-END counter (DISC entity), 197 BLOCKS-INUSE-START counter (DISC entity), 196 BLOCKS-SPLITS counter (DISC entity), 197 BRIEF report attribute, 110 Busy counter, defined, 139 BY clause LIST entity-type command, 69 C C counter (DISC entity) legacy, 202 ZMS, 196 C or C++ Language, usage notes, TAL redefines, 549 CACHE-BLOCK-SIZE identifier (OSSCPU entity), 274 CACHE-BLOCKS-ALLOWED
COMMENTS command, 51, 56 Common object format files (COFF files), 315 Communication lines, measuring See CLUSTER entities, LINE entities, NETLINE entities, and SYSTEM entities , 29 COMP-TRAPS counter CPU entity, 166 PROCESS entity, 300 CONFIG-NAME identifier DEVICE entity, 183 DISC entity, 194 OPDISK entity, 266 Configuration table, 377, 384 Configuration, through the command interface adding entities, 41 deleting entities , 51 listing all configured entities, 64 listing configured entities of a specified e
LINE entity, 254 NETLINE entity, 262 OPDISK entity, 269 SERVERNET entity, 331 CTRL-TYPE identifier CONTROLLER entity, 159 SERVERNET entity, 327 CTRL^DESC descriptor, 455 CURRENT-PRIORITY-END counter PROCESS entity, 301 CURRENT-PRIORITY-START counter PROCESS entity, 301 Custom measurements See User-defined counters , 29 CW counter (DISC entity), 204 D Data files, measurement See also Measurements , 29 disk space requirements for, 46 from a different Measure PVU, 46 on a remote system, 46 through the command
LINE entity, 251 NETLINE entity, 259 OPDISK entity, 266 DEVICE^CLIM^DESC descriptor, 457 DEVICE^DESC descriptor, 456 DEVICE^SVNET^DESC descriptor, 459 DEVICE^SVNET^DESC^G05 descriptor, 461 DIN4-BYTES counter (NETLINE entity), 259 DIN4-BYTES-F counter (NETLINE entity), 262 DIR-OPENS counter (OSSCPU entity), 275 DIR-OPENS-LIMIT counter (OSSCPU entity), 275 DIR-STREAMS counter (OSSCPU entity), 275 DIR-STREAMS-LIMIT counter (OSSCPU entity), 275 DISC entities configuring, for command interface, 41 configuring, f
ENDING-UCL-LOCK counter, CPU entity (not used)), 174 ENDING-UCME counter (CPU entity), 167 ENDING-UDS counter (CPU entity), 174 ENDING-UDS-LOCK counter, CPU entity (not used), 174 Enform files, report output for, 70, 88, 110 Entity descriptors, 376, 451 Entity specifications See also individual entity types by name , 29 adding, to measurement configuration, 41 excluding, from measurement configuration, 51 in procedure calls, 376, 451, 455 Entity types See also individual entity types by name , 29 summary of
FM-STOLEN-BLK-WT-REQS counter (OSSCPU entity), 276 FOR clause LIST entity-type command, 70 LIST PLOT command, 82 LISTACTIVE command entity-type, 88 RESET PLOT, 103 RESET REPORT command, 104 SET PLOT command, 106 SET REPORT command, 110 SHOW PLOT command, 117 SHOW REPORT command, 118 START MEASUREMENT command, 124 FORMAT clause LIST entity-type command, 70 LISTACTIVE command, 88 RESET REPORT command, 104, 110 SET REPORT command, 110 SHOW REPORT command, 118 FOX link, measuring See CLUSTER entities , 29 FREE-
HELP command (MEASFT), 559 High PINs, 549 HISTORY command, 61 HITS counter (DISC entity) legacy, 202 ZMS, 197 HOME-NET-TRANS counter (TMF entity), 363 HOME-NET-TRANS-QMAX counter (TMF entity), 365 HOME-NET-TRANS-QTIME counter (TMF entity), 364 HOME-TRANS counter (TMF entity), 364 HOME-TRANS-QMAX counter (TMF entity), 365 HOME-TRANS-QTIME counter (TMF entity), 364 HOMETERM-NAME identifier (PROCESS entity), 298 HOMETERM-SYSNAME identifier (PROCESS entity), 298 Hyphen, double (- -), 37 I IC-DIRTY-QTIME counte
syntax for names of D-series, 248 syntax for names of G-series, 249 LINE identifier LINE entity, 251 NETLINE entity, 259 Lines, communication, measuring See LINE entities and NETLINE entities , 29 LINK-LARGE-MSGS counter (CPU entity), 168 LINK-PREPUSH-MSGS counter (CPU entity), 168 LINK-READLINK-MSGS counter (CPU entity), 168 LINK-TIME counter (SYSTEM entity), 356 LINKS counter (SYSTEM entity), 356 LIST entity-type command, 68 LIST EXTNAMES command, 79 LIST OSSNAMES command, 80 LIST PLOT command, 82 LISTACT
description of, 548 during CPU restart, 120 getting status of, 127 reading active counters, 93 MEASDDLB file, 548 MEASDDLF file, 548 MEASDDLS file, 548 MEASDECS file, 376, 549 MEASFH process creating, 45, 414 description of, 548 disk space requirements, 46 reading measurement data, 93 stopping, 54, 380, 382 using a different version of, 46 using a remote, 46 MEASFT utility, 46, 550 MEASGETVERSION procedure, 392 MEASIMMU file, description of, 549 MEASINFO procedure, 394 MEASLISTENAME procedure, 399 MEASLISTE
PROCESS entity, 304 NATIVE-PROCESS identifier (PROCESS entity), 298 NETLINE entities configuring, for command interface, 41 configuring, for programmatic interface, 384, 456 DDL record, field definitions, 258 DDL record, for legacy format, 257 DDL record, for legacy format D-series, 257 DDL record, for legacy format G-series, 257 DDL record, for ZMS style format, 256 displaying names of, 61 displaying names of configured, 62 syntax for names of D-series, 255 syntax for names of G-series, 256 Network lines,
DDL record, for ZMS style format, 285 syntax for names of, 284 usage notes, 290 OSSNS-DD-CALLS counter (PROCESS entity), 305 OSSNS-MESSAGE-BYTES counter (PROCESS entity), 305 OSSNS-REDIRECTS counter (PROCESS entity), 305 OSSNS-REQUESTS counter (PROCESS entity), 305 OSSNS-WAIT-TIME counter (PROCESS entity), 305 OSSPATH command, 101 OSSPID identifier PROCESS entity, 298 PROCESSH entity, 319 SQLPROC entity, 340 SQLSTMT entity, 348 USERDEF entity, 370, 371 OUT clause ENV command, 55 HELP command, 59 INFO COUNTE
FILE entity, 237 SERVERNET entity, 328 Pound symbol (#), 86 PP-ALLOC-FAILURES counter (OSSCPU entity), 280 PP-BLOCK-INUSE-QMAX counter (OSSCPU entity), 282 PRES-PAGES-END counter (PROCESS entity), 305 PRES-PAGES-MAX counter (PROCESS entity), 309 PRES-PAGES-QTIME counter (PROCESS entity), 305 PRES-PAGES-START counter (PROCESS entity), 306 PRIORITY identifier PROCESS entity, 298 PROCESSH entity, 319 SQLPROC entity, 340 SQLSTMT entity, 348 Procedures, callable See also individual procedures by name , 29 data d
R RATE clause effect on counter values, 137 LIST entity-type command, 72 LISTACTIVE entity-type command, 88 RESET REPORT command, 104 SET REPORT command, 112 SHOW REPORT command, 118 READ-BUSY-TIME counter DEVICE entity, 187 DISC entity, 205 LINE entity, 252 NETLINE entity, 260 OPDISK entity, 269 READ-BYTES counter CPU entity, 171 FILE entity, 241 SERVERNET entity, 329 READ-CBYTES counter (SERVERNET entity), 329 READ-CREQUESTS counter (SERVERNET entity), 330 READ-QBUSY-TIME counter DEVICE entity, 185 DISC e
TERMINAL entity, 360 REQUESTS-BLOCKED counter DISC entity, 200 DISCOPEN entity, 215 DISKFILE entity, 224 RESET PLOT command, 103 RESET REPORT command, 104 Resetting attributes PLOT, 103 REPORT, 104 Response time counter, defined, 141 RESPONSE-TIME counter CPU entity, 170 LINE entity, 253 TERMINAL entity, 360 RETRIES counter LINE entity, 253 SERVERNET entity, 330 RETURNED-BYTES counter CLUSTER entity, 156 PROCESS entity, 307 RETURNED-BYTES-F counter CLUSTER entity, 157 PROCESS entity, 309 RETURNED-CBYTES cou
SQL statements, measuring see SQLSTMT entities SQL-DELETES counter (DISKFILE entity), 224 SQL-ENDING-ROWS counter (DISKFILE entity), 223 SQL-INSERTS counter (DISKFILE entity), 224 SQL-NEWPROCESS-TIME counter (SQLPROC entity), 341 SQL-NEWPROCESSES counter (SQLPROC entity), 341 SQL-OBJ-RECOMPILE-TIME counter (SQLPROC entity), 341 SQL-OBJ-RECOMPILES counter (SQLPROC entity), 341 SQL-OPERATION-TIME counter (DISCOPEN entity), 215 SQL-STMT-RECOMPILE-TIME counter (SQLPROC entity), 341 SQL-STMT-RECOMPILES counter (
SWAPS counter CPU entity, 172 DISC entity, 201 SWAPVOL command, 132 Symbol to use in plot, specifying, 67 Syslink counter, defined, 143 System code space, 315, 478 SYSTEM command, 132 System data space, 128 SYSTEM entities configuring, for command interface, 41 configuring, for programmatic interface, 384, 489 DDL record, field definitions, 355 DDL record, for legacy format, 355 DDL record, for ZMS style format, 354 displaying names of configured, 62 syntax for names of, 353 usage notes, 356 System library
TTY-OPENS-LIMIT counter (OSSCPU entity), 282 Type field in configuration table header record, 378 in configuration table trailer record, 378 in entity descriptors, 376 TYPE identifier USERDEF entity, 371 U U1024-BYTES counter (NETLINE entity), 261 U128-BYTES counter (NETLINE entity), 261 U2048-BYTES counter (NETLINE entity), 261 U256-BYTES counter (NETLINE entity), 261 U4096-BYTES counter (NETLINE entity), 261 U512-BYTES counter (NETLINE entity), 261 U64-BYTES counter (NETLINE entity), 261 UCL-LOCK-MAX cou
WRITE-QBUSY-TIME counter DEVICE entity, 186 DISC entity, 201 OPDISK entity, 268 SERVERNET entity, 331 WRITE-QLEN-MAX counter DEVICE entity, 187 DISC entity, 205 SERVERNET entity, 332 WRITE-QTIME counter DEVICE entity, 186 DISC entity, 201 OPDISK entity, 269 SERVERNET entity, 331 WRITE-REQUESTS counter CPU entity, 171 SERVERNET entity, 331 WRITES counter DEVICE entity, 186 DISC entity, 201 FILE entity, 242 LINE entity, 253 NETLINE entity, 262 OPDISK entity, 269 TERMINAL entity, 360 Writing measurement data t