Using system statistics

From m204wiki
Jump to navigation Jump to search

Overview

The tables on this page are provided for your use in identifying and tracking statistics that are reported by Model 204 in various situations. The offsets, given in decimal and hexadecimal, are to help you locate particular statistics within statistical output. The length is in bytes.

The statistics are collected in the CCAJRNL data set, or if allocated, in CCAJLOG. The offsets in the tables below locate each statistic in the corresponding record. To accommodate sites that need to run Model 204 around the clock for many days at a time, space for most Model 204 statistics is a double word.

Description of statistics

The "Statistics with descriptions" table lists the statistics that are collected and that you can evaluate. It also describes the purpose of each statistic followed by the circumstances when you can collect the information and an offset in the journal record.

To collect several statistics you must preset a parameter to a nonzero value. Statistics that require a parameter are also noted in the "Statistics with descriptions" table. The table in Setting parameters to collect certain statistics lists those statistics with the required parameter.

Tables, identified by type, later in this topic, describe the layouts of the various journal records.

Statistics with descriptions
Statistic Records...
APSYLD Number of APSY loads
APSYLDD Number of APSY loads from a dataspace
APSYLDT Number of tiny APSY loads
AUDIT Number of lines written to the journal and/or the audit trail
BACKOUTS Number of BACKOUTs
BADD Number of fields added to Table B
BCHG Number of fields changed in Table B
BDEL Number of fields deleted from Table B
BLKCFRE Number of times user held a critical file resource (CFR) exclusively and forced another user to wait

Parameter setting required: see Parameters to set to collect certain statistics.

BLKI Percentage of performance samples at which the user was in a server but blocked from running and waiting for an external event to occur (reported only if the performance subtask is active)

Parameter setting required: see Parameters to set to collect certain statistics.

BLKO Percentage of performance samples at which the user was swapped out to the server data set but blocked from running and waiting for an external event to occur (reported only if the performance subtask is active)

Parameter setting required: see Parameters to set to collect certain statistics.

BLKRLK Number of times user was the first user identified as blocking another from obtaining a record lock

Parameter setting required: see Parameters to set to collect certain statistics.

BXCHNG Number of value entries changed in the ordered index. This occurs when a record number is added to or removed from an existing value entry.
BXDELE Number of value entries deleted from the ordered index
BXFIND Number of ordered index searches to locate field name=value pairs in the ordered index
BXFREE Number of ordered index pages (nodes) emptied and released from the ordered index
BXINSE Number of new value entries inserted into the ordered index
BXNEXT Number of value entries touched in the ordered index during a range retrieval
BXRFND Number of times the cursor was repositioned in the ordered index during a search
BXSPLI Number of Table D pages (nodes) split in the ordered index
CCATDIFF Difference between CCATEMP pages allocated and freed by a thread. This is an imperfect indicator of the number of CCATEMP pages a thread has allocated. CCATDIFF is imperfect because it is difficult to fully account for pages allocated or freed on behalf of the system or other users. However, if there is an acute CCATEMP page shortage, threads with high CCATDIFF values are likely culprits.

CCATDIFF is available in Model 204 7.7 and higher.

CCATDIFH High water mark of CCATDIFF in a request. This can be useful if a request allocated a lot of CCATEMP pages, causing an acute shortage but then freed those pages.

CCATDIFH is available in Model 204 7.7 and later.

CDLWAIT Number of constraint dependency lock waits
CNCT Elapsed time in seconds
COMMITS Number of COMMITs
CPU Total CPU time consumed in milliseconds. The system CPU statistics include both maintask and subtask CPU usage.
CPUONZIP Total CPU time consumed in milliseconds on zIIP engines
CPUTOTZE Total CPU time consumed in milliseconds by workloads eligible for offload to zIIP engines
DEQ Total number of units of work taken by the MAINTASK and each MP and zIIP subtask, regardless of the queue that the work was taken from. Reported at Model 204 termination or in MONITOR TASKS output.

(type Subtask)

DEV# used the MXI interface to hold the IODEV type ( e.g. 7 for IODEV7 )
DEV5 Output lines to procedures defined during the run
DEV6 Input lines from included procedures
DEV7 Output lines to SNA Communications Server (formerly VTAM) 3270s
DEV8 Input lines from SNA Communications Server 3270s
DEV9 Page headers or trailers defined in requests
DEV10 Output lines written to directed output (USE) data sets
DEV11 Output lines to remote User Language full screen
DEV12 Input lines from a remote User Language full screen
DEV13 Output lines to remote SQL
DEV14 Input lines from remote SQL
DEV17 Output lines to SQL CRAM IUCU
DEV18 Input lines from SQL CRAM
DEV19 Output lines to remote SQL LU62
DEV20 Input lines from remote SQL LU62
DEV23 Input lines to CRAM IFAM
DEV24 Input lines from CRAM IUCU thread
DEV27 Terminal error messages issued for inbound Distributed Application Facility conversations
DEV28 Reserved
DEV31 IFDISP output lines to IFAM1 programs
DEV32 Input arguments supplied by certain IFAM1 calls
DEV37 Output lines to SNA Communications Server 2741s and Teletypes
DEV38 Input lines from SNA Communications Server 2741s and Teletypes
DEV49 output lines to Remote Command line applications( rate or total)
DEV50 Lines sent to the printer when ROUTER=CICS
DEV53 Output lines from an IFAM1 application
DEV54 Input lines from an IFAM1 application
DEV55 Reserved
DEV56 Reserved
DEV57 Reserved
DEV58 Reserved
DEV59 Reserved
DEV60 Reserved
DEV61 Reserved
DEV62 Reserved
DEV63 Misused for UTI stats if X'08' bit set in the MONPARM parameter|MONPARM]] system parameter.
DEV64 Misused for UTI stats if X'08' bit set in the MONPARM system parameter.
DEV65 Misused for UTI stats if X'08' bit set in the MONPARM system parameter.
DEV66 Misused for UTI stats if X'08' bit set in the MONPARM system parameter.
DEV67 Misused for UTI stats if X'08' bit set in the MONPARM system parameter.
DEV68 Misused for UTI stats if X'08' bit set in the MONPARM system parameter.
DEV69 Misused for UTI stats if X'08' bit set in the MONPARM system parameter.
DEV7 Output lines to SNA Communications Server (formerly VTAM) 3270s
DEV70 Misused for UTI stats if X'08' bit set in the MONPARM system parameter.
DEV71 Misused for UTI stats if X'08' bit set in the MONPARM system parameter.
DEV72 Misused for UTI stats if X'08' bit set in the MONPARM system parameter.
DEV73 Misused for UTI stats if X'08' bit set in the MONPARM system parameter.
DEV74 Misused for UTI stats if X'08' bit set in the MONPARM system parameter.
DIRRCD Number of records searched in direct searches of Table B
DKAR Number of buffers allocated without page read requests
DKPR Number of requests for a page (may not require real I/O)
DKPRF Number of fast logical page reads. Number of DKPRs that were satisfied by pending or deferred close buffers
DKRD Number of physical page reads from Model 204 files
DKRDL Number of physical page reads into the below the bar buffer pool
DKRR Number of real disk reads for recently requested pages
DKSAWB Anticipatory writes from the bottom of the LRU queue. The page in the buffer is then deleted and no longer available without a re-read. Usually a small statistic and typically only incremented by a few rare events.

Offset: see the Disk buffer monitor statistics table.

DKSAWBL Same as DKSAWB but only incremented for anticipatory writes from below the bar

Offset: the Disk buffer monitor statistics table.

DKSAWW Anticipatory writes from the anticipatory write windows set by LDKBMWND and/or LDKBMWNG

Offset: the Disk buffer monitor statistics table.

DKSAWWL Same as DKSAWW but only incremented for anticipatory writes from below the bar
DKSDIR High-water mark of modified (dirty) buffers found within the DKBM statistics window which is created by the LDKBMW parameter. This parameter is not recommended due to the extremely heavy performance penalty imposed by a non-zero value.

Offset: see the Disk buffer monitor statistics table.

Parameter setting required: see Parameters to set to collect certain statistics.

DKSDIRT The total of all modified buffers found within the DKBM statistics window

Offset: see the Disk buffer monitor statistics table.

Parameter setting required: see Parameters to set to collect certain statistics.

DKSFBS Times a buffer was needed, and a scan was necessary, because the oldest free buffer was not immediately available

Offset: see the Disk buffer monitor statistics table.

DKSKIP The highest number of buffers skipped during any search for a free buffer. A buffer is skipped because it is not immediately available.

Offset: see the Disk buffer monitor statistics table.

DKSKIPT The total of all buffers skipped when looking for a free buffer

Offset: see the Disk buffer monitor statistics table.

DKSRR Number of times a page that was expected to be in the buffer pool could not be located, which necessitates a physical I/O to disk

Offset: see the Disk buffer monitor statistics table.

DKSRRFND Number of times that a page that was expected to be in the buffer pool was located there, which eliminates the need for a physical I/O
DKSTBLA Table A reads causing real I/O

Offset: see the Disk buffer monitor statistics table.

DKSTBLB Table B reads causing real I/O

Offset: see the Disk buffer monitor statistics table.

DKSTBLC Table C reads causing real I/O

Offset: see the Disk buffer monitor statistics table.

DKSTBLD Table D reads causing real I/O

Offset: see the Disk buffer monitor statistics table.

DSKTBLE Table E reads causing real I/O

Offset: see the Disk buffer monitor statistics table.

DKSTBLF FCT reads causing real I/O

Offset: see the Disk buffer monitor statistics table.

DKSTKQC Times the current Table B page had to be closed to allow a fourth buffer to be opened

Offset: see the Disk buffer monitor statistics table.

DKSWRP The highwater mark of buffers with writes outstanding within the DKBM statistics window, created by the parameter LDKBMW. Rocket Software advises against setting this parameter to a non-zero value due to the extreme performance penalty imposed.

Offset: see the Disk buffer monitor statistics table.

Parameter setting required: see Parameters to set to collect certain statistics.

DKSWRPT The total number of buffers with outstanding writes within the DKBM statistics window

Offset: see the Disk buffer monitor statistics table.

Parameter setting required: see Parameters to set to collect certain statistics.

DKUPTIME Milliseconds of DKUPDT time
DKWR Number of physical page writes to Model 204 files
DKWRL Number of physical page writes from the below the bar pool
DKWTMICS Total wait time in microseconds for Disk I/O waits counted by DKWTS.

As of V7.7, with 77z339 applied

DKWTS Number Disk I/O waits

As of V7.7, with 77z339 applied

DUMP Writes to Model 204 file backups (DUMP command output)
DUPDTS Records written to a deferred update data set for the file
ECCALL Number of External Call Facility calls
ECCNCT External Call Facility - Elapsed time for external program
ECCTOUT Number of External Call Facility calls that timed out (program or subtask unavailable)
ECCWAITM ECF – Number of calls that waited for module to become available
ECCWAITS Number of External Call Facility calls that waited for a subtask to become available
ECDELETE Number of EXTERNAL DELETE calls to External Call Facility
ECLOAD Number of EXTERNAL LOAD calls to External Call Facility
ECMODMAX ECF – Highwater mark of modules loaded
ECNAMMAX External Call Facility – High-water mark of call names defined
ECTSKMAX External Call Facility – High-water mark of subtasks active
ECTWAITM External Call Facility – Elapsed time spent waiting for a module to become available
ECTWAITS External Call Facility – Elapsed time spent waiting for a subtask to become available
ERRPDL Highwater mark of the system pushdown list
FBMX disk buffer hwm
FBWT Number of waits for a disk buffer
FINDS Number of FIND statements evaluated
FSCB High-water mark of bytes used in the FSCB
FSCBSW Number of full-screen buffer swaps
FTBL Highwater mark of FTBL appearing in CMPL and EVAL lines for requests containing a field name reference in group context. If one or more groups are opened but no field name references are made, the portion of FTBL used by the open groups is not reported.
GTBL Highwater mark of GTBL
GTBLRS Number of GTBL rearrangements required to add a global object
GTBLRU Number of GTBL rearrangements required to add a global string variable
HEAP Dynamic memory highwater mark
HRDRSTRT number of hard restarts
IN Number of terminal input lines
INCMFS Input lines from CMS full screen
INCMIO Input lines from CMS non-full screen
INCRAM Input lines from Remote User Language threads (IODEV 29)
INVMFS Input lines from IUCV or VMCF full-screen thread
INVMIF Input arguments supplied by certain IFAM2 or IFAM4 calls from CMS programs
INVMIO Input lines from IUCV or VMCF non-full screen thread
INXX Input lines from QSAM users (IODEV 3)
ITBL Highwater mark of ITBL
IXADD Number of index entries added to Tables C and D, including attempts to add duplicates
IXDEL Number of entries deleted from Tables C and D
LKPOST Times an MP subtask posted another task to indicate that an MP lock had become available
LKWAIT Times MP subtasks invoked z/OS WAIT macros to wait for an available MP lock
LONGUPDTIME(MS) Total milliseconds of 'too long' update units

Parameter setting required: see Parameters to set to collect certain statistics.

LONGUPDTS Number of 'too long' update units

Parameter setting required: see Parameters to set to collect certain statistics.

LWTIM Elapsed time (ms), spent waiting for a multiprocessing lock. Reported for each MP and zIIP subtask at Model 204 termination or in MONITOR TASKS output (only on Model 5 terminals).
MAXIOX Number of times Model 204 waited for a free database buffer
MBO Highwater mark of transaction backout pages

As of V7.7, with zap 77z257 applied.

MCP Highwater mark of transaction checkpoint spans

As of V7.7, with zap 77z257 applied.

MOVE Number of times the scheduler switched from one user to another
MPHASHD MP duplicate hash entries found
MPLKPREM Total elapsed time in milliseconds, across the maintask and all subtasks, the Online spent waiting due to operating system preemption.

This is the elapsed time between when an MP lock becomes available (lock post) making a task ready to run, and when the task actually gets the CPU. That preemption delay is caused by the operating system dispatching other tasks ahead of this task.

Only collected in MP/204 systems.

MPLKWTIM Total elapsed time in milliseconds, across the maintask and all subtasks, the Online spent waiting for MP locks

Only collected in MP/204 systems.

MPR Highwater mark of transaction pre-images

As of V7.7, with zap 77z257 applied.

MQAPICNT MQ/204 – Count of MQ/Series API calls
MQAPITIM MQ/204 – Elapsed time of MQSeries API calls
MQBYTEIN MQ/204 – Total bytes retrieved (MQGET)
MQBYTEOU MQ/204 – Total bytes sent (MQPUT/MQPUT1)
MQGETS MQ/204 – Number of MQGETs performed
MQGWTCNT MQ/204 – Number of MQGETs with nonzero, non-unlimited wait
MQGWTSUC MQ/204 – Number of MQGETs with nonzero, non-unlimited wait that succeeded
MQGWTTIM MQ/204 – Elapsed time for MQGETs with nonzero, non-unlimited wait
MQGWTTSP MQ/204 – Total wait time specified on MQGETs with nonzero, non-unlimited wait
MQHWQU MQ/204 – High-water mark of queues concurrently in use
MQHWTASK MQ/204 – High-water mark of MQ/204 subtasks in use
MQNUMQM MQ/204 – Number of distinct queue managers connected
MQNUMQU MQ/204 – Number of distinct queues accessed
MQPUTS MQ/204 – Number of MQPUTs performed
MQWTM Mean queue wait time, the average time a subtask spent waiting for work (waiting for PCBs to come into the offload queue)
MTDEQ The number of times that the MAINTASK took a unit of work from the MAINTASK queue and processed it
MTSDEQ The number of times that the MAINTASK took a unit of work from an MP subtask queue and processed it because the MAINTASK was idle
MTZDEQ The number of times that the MAINTASK took a unit of work from a zIIP subtask queue and processed it because the MAINTASK was idle
NTBL Highwater mark of NTBL
OBJSWAP number of object swaps for Sirius objects
OFFIN Count of records read in by the OFFLOAD subtask
OFFOU Count of records written to the OFFLOAD stream
OUT Number of terminal output lines
OUTCMFS Output lines to CMS full-screen thread
OUTCMIO Output lines to CMS line-by-line thread
OUTCRAM Output lines from remote User Language threads (IODEV 29)
OUTPB Highwater mark of OUTPB
OUTVMFS Output lines to IUCV or VMCF full-screen thread
OUTVMIF Output lines to IFAM2 or IFAM4 calls from CMS programs
OUTVMIO Output lines to IUCV or VMCF line-by-line threads
OUTXX Output lines to QSAM users (IODEV 3)
PBRSFLT Number of private buffer reservation faults
PCPU Percentage of time Model 204 was given the CPU when it wanted CPU time. In an MP environment, system PCPU is the sum of CPU time in all tasks divided by the sum of elapsed time in all tasks.
PDL Highwater mark of the pushdown list
PETIM Elapsed time (ms), spent waiting due to OS preemption. Reported for the MAINTASK and each MP and zIIP subtask at Model 204 termination or in MONITOR TASKS output (only on Model 5 terminals)
PNDGTIME Milliseconds of unnecessary broken time
PR Real time consumed, in milliseconds. PCPU = CPU / PR.
QTBL Highwater mark of QTBL
RCV Number of blocks needed for subtransaction recovery

As of V7.7, with zap 77z257 applied.

RECADD Number of records started in Table B, not including extension records
RECDEL Number of records deleted from Table B, not including extension records or records deleted by DELETE RECORDS or IFDSET
RECDS Number of records processed by FOR statements, SORT statements, IFGET calls, or IFPOINT calls, or similar SQL record processing
REDY Percentage of performance samples at which the user was ready to run (reported only if the performance subtask is active)

Parameter setting required: see Parameters to set to collect certain statistics.

REQ Number of SOUL requests evaluated
REST Number of reads from a Model 204 file backup (RESTORE command)
RETRYA Page retries in Table A
RETRYC Page retries in Table C
RQTM Elapsed time for the activity being reported, exclusive of terminal I/O time
RSXCOMP Number of compactions of the record locking table. If greater than 0, increase LRETBL.
RUNG Percentage of performance samples at which the user was running (reported only if the performance subtask is active)

Parameter setting required: see Parameters to set to collect certain statistics.

SCHDCPU In MP configuration, total scheduler CPU for the user or request
SCREENS Total number of full-screen reads (READ SCREEN statements) evaluated by a SOUL request.
SFTRSTRT number of soft restarts.
SGMTI Number of input lines from included procedures
SGMTO Number of output lines to procedures defined by the user
SLIC Number of times that the user was time-sliced by the Model 204 scheduler
SMPLS Number of performance sample points at which the user was active or total number of performance samples taken at the system level (reported only if the performance subtask is active)

Parameter setting required: see Parameters to set to collect certain statistics.

SORTS Number of SOUL SORT statements or SQL sorts evaluated
SQLI Remote SQL input highwater mark
SQLO Remote SQL output highwater mark
SQRD Number of terminal input lines
SQWR Number of terminal output lines
STBL Highwater mark of STBL
STCPU Total amount of time, in milliseconds, that the user has run in an MP offload subtask
STDEQ Number of times that an MP subtask took a unit of work from an MP subtask queue and processed it
STIMERS Total number of STIMER or STIMERM calls
STPOST Times MP subtasks performed real operating system POSTs to send work to the maintask
STRECDS Number of records processed by SORT statements or SQL sorts
STWAIT Number of times MP subtasks invoked z/OS WAIT macros to wait for work from the maintask
STZDEQ Number of times that an MP subtask took a unit of work from a zIIP subtask queue and processed it because the MP subtask was idle
SVAC Active servers (an average, included on a performance line)
SVMX Highwater mark for servers (0 if no server swapping)
SVPAGES SVPAGES * 4096 is the number of bytes transferred as a result of server reads and writes
SVRD Number of server reads
SVWR Number of server writes
SWPG Percentage of performance samples at which the user was swapping in or out (reported only if the performance subtask is active)
SWT Times Model 204 issued a real WAIT while not in user-switching mode
TEMX Highwater mark of CCATEMP pages used in the expansion area
TFMX Highwater mark of CCATEMP pages used
TSMX Highwater mark of CCATEMP pages used in the small model page pool
TTBL Highwater mark of TTBL
UBUFHWS Highwater mark, in bytes, of user's Universal Buffer
UDD Number of lines written to a directed output (USE) data set
UPDTTIME(MS) Number of milliseconds that a file was being updated

UPT(MS) appears as the column header for this statistic on a report

UPDTTIME Number of milliseconds that the user was actively updating at least one file
USMX Highwater mark for simultaneous active users (reported only if the performance subtask is active)
USRS Average number of active users (reported only if the performance subtask is active)

Parameter setting required: see Parameters to set to collect certain statistics.

VTBL Number of VTBL entries at the end of compilation (CMPL) processing or the end of evaluation (EVAL) processing
WAIT Number of real operating system waits
WTCFR Number of times that the user waited to obtain a critical file resource, either share or exclusive

Parameter setting required: see Parameters to set to collect certain statistics.

WTRLK Number of times that the user waited to obtain a record lock

Parameter setting required: see Parameters to set to collect certain statistics.

WTSV Percentage of performance samples at which the user was waiting for a server (reported only if the performance subtask is active)

Parameter setting required: see Parameters to set to collect certain statistics.

XTBL Highwater mark of XTBL
ZTCPU The CPU (ms) consumed on a zIIP processor by the MAINTASK, an MP subtask or a zIIP subtask. Reported for each task at Model 204 termination and in MONITOR TASKS output (for Model 5 terminals only)
ZTDEQ The number of times that a zIIP subtask took a unit of work from a zIIP subtask queue and processed it

Header and trailer entries (Type 0)

Each journal block begins with a header record and ends with a trailer record.

Header entries (Type 0)

Header entries have the format shown in the following table:

Header entry formats
Offset dec(hex) Length Data type Description
0(0) 2 Binary Length of entire journal record, including header and trailer
2(2) 1 Binary Type indicator = X'00'
3(3) 1 Binary Flags used by recovery
4(4) 4 Unsigned packed data Julian date = 0CYYDDDF
8(8) 4 Unsigned packed data Time = HHMMSSTH
12(0C) 4 Binary Sequence number
16(10) 10 Binary System Store Clock Extended (STCKE) value to provide picosecond accuracy on the journal date/time stamp

Bytes 1-9 || byte 16

26(1A) 2 Binary Length of header, type 00
28(1C) 2 Binary Version in hex
30(1E) 8 EBCDIC Version in EBCDIC
38(26) 8 EBCDIC Job name
46(2E) 8 EBCDIC JES Job ID
54(36) 4 EBCDIC SMF System ID
  • If the flag byte at offset 3 of type 00 has the X'02' bit on, then it is a header record from Model 204 version 7.4.0 or later with its length at offset X'1A'.
  • If that flag byte has the X'01' bit on, then the header is from version 7.1.0 and is 26 bytes in length.
  • If neither of those bits is on, then the header is a from a pre-7.1.0 version and is 16 bytes in length.

Recovery flags

Flags used by recovery can have the value of X'00' or a sum of any of the following:

Value Block contains these records...
X'80' Type 1
X'40' Type 2
X'20' Type 3
X'10' Type 4
X'08' Type 5
X'04' Type 6

Sequence numbers

The sequence number represents the number of each journal record, starting at 0 and increasing by 1. A merged journal (output of the MERGEJ utility) sequence number consists of a 2-byte hexadecimal number at offset 12(0C), which indicates the input file this journal record came from, and a 2-byte hexadecimal field at offset 14(0E), which indicates the sequence within the INPUT journal.

Trailer entries (Type 0)

Trailer entries have the format shown in the following table:

Trailer entry formats
Offset dec(hex) Length Data type Description
0(0) 2 Binary Length of entry = X'000C'
2(2) 1 Binary Type indicator = X'00'
3(3) 1 Unused
4(4) 4 Unsigned packed data Julian date = 0CYYDDDF
8(8) 4 Unsigned packed data Time = HHMMSSTH

Recovery entries (Types 1-6)

Entry types X'01' through X'06' contain information used for recovery. They follow the same general format as all other journal entries. There is no type X'07' journal entry.

The format of the recovery entries is meaningful only in the context of the Model 204 ROLL FORWARD algorithm. However, entry types might be of interest in some statistics gathering applications, as follows:

Type Contains recovery information from...
1 Discontinuity
2 Begin update unit.

More than one type 2 entry can be written for each update unit. The update unit number appears as a 4-byte binary number at offset 16(10).

3 End update unit.

The update unit number appears in the same format and location as the type 2 entry.

4 File state information: opened, closed, dumped, file parameter list (FPL) page preimage logged or open at the time of a checkpoint
5 Checkpoint
6 File update

System statistics entries (Type 8)

Cumulative statistics, listed by offset in the Layout of system subtype X'02' entries table, are kept for the entire system. Final, partial, and performance system statistics are written to the journal during Model 204 termination in a line that begins with:

ST $$$ SYSTEM='version#'

For additional information about:

Final statistics

The system final statistics line, similar to the following, is produced as part of system termination:

ST $$$ SYSTEM='5A 'AUDIT=43 OUT=38 IN=14 WAIT=19 ...

Partial statistics

Statistics in a system partial statistics line, similar to the following, accumulate from the beginning of the run, if the User 0 ACCTIM parameter is nonzero:

ST $$$ SYSTEM='6.1.0'PARTIAL=3 AUDIT=19 OUT=10 IN=5 WAIT=3 ...

Performance statistics

System performance statistic lines, similar to the following, are written if the User 0 RPTCNT and SMPLTIM parameters are nonzero:

ST $$$ SYSTEM='6.1.0' PERFORMANCE=n

The numbers determined at the performance sample points are averaged for the entire run.

Entry formats

System statistics entries have the format shown in the following table:

System statistics entry formats
Offset dec(hex) Length Data type Description
0(0) 2 Binary Length of entry
2(2) 1 Binary Type indicator = X'08'
3(3) 1 Binary Subtype indicator:

X'00' = final statistics

X'01' = partial statistics

X'02' = performance statistics

X'04' = additional disk buffer monitor

X'08' = subtask statistics

4(4) 4 Unsigned packed data Julian date = 0CYYDDDF
8(8 4 Unsigned packed data Time = HHMMSSTH
12(0C) 2 Binary Subtask number if subtype = X'08'; otherwise unused
14(0E) 2 Binary Unused
16(10) to end of journal

Contents depend on subtypes, shown in tables in the following sections:

System final and partial statistics

For subtypes X'00' (final statistics) and X'01' (partial statistics), the layout of the remainder of the entry is shown in the following table. For a description of each statistic, see the Statistics with descriptions table. The following statistics are data type Binary. However, the length varies — either 4 or 8, to accommodate 64-bit system architecture.

System final and partial statistics
Offset dec(hex) Length Description
16(10) 8 AUDIT
24(18) 8 OUT
32(20) 8 IN
40(28) 8 OUTXX
48(30) 8 INXX
56(38) 8 DEV5
64(40) 8 DEV6
72(48) 8 DEV7
80(50) 8 DEV8
88(58) 8 DEV9
96(60) 8 DEV10
104(68) 8 DEV11
112(70) 8 DEV12
120(78) 8 DEV13
128(80) 8 DEV14
136(88) 8 OUTTTY
144(90) 8 INTTY
152(98) 8 DEV17
160(A0) 8 DEV18
168(A8) 8 DEV19
176(B0) 8 DEV20
184(B8) 8 Spare
192(C0) 8 Spare
200(C8) 8 DEV23
208(D0) 8 DEV24
216(D8) 8 DEV25
224(E0) 8 DEV26
232(E8) 8 DEV27
240(F0) 8 DEV28
248(F8) 8 OUTCRAM
256(100) 8 INCRAM
264(108) 8 DEV31
272(110) 8 DEV32
280(118) 8 Spare
288(120) 8 Spare
296(128) 8 DEV35
304(130) 8 DEV36
312(138) 8 DEV37
320(140) 8 DEV38
328(148) 8 OUTVMIO
336(150) 8 INVMIO
344(158) 8 OUTVMFS
352(160) 8 INVMFS
360(168) 8 OUTVMIF
368(170) 8 INVMIF
376(178) 8 OUTCMIO
384(180) 8 INCMIO
392(188) 8 OUTCMFS
400(190) 8 INCMFS
408(198) 8 DEV49
416(1A0) 8 DEV50
424(1A8) 8 OFFIN
432(1B0) 8 OFFOU
440(1B8) 8 DEV53
448(1C0) 8 DEV54
456(1C8) 8 DEV55
464(1D0) 8 DEV56
472(1D8) 8 DEV57
480(1E0) 8 DEV58
488(1E8) 8 DEV59
496(1F0) 8 DEV60
504(1F8) 8 DEV61
512(200) 8 DEV62
520(208) 8 DEV63
528(210) 8 DEV64
536(218) 8 DEV65
544(220) 8 DEV66
552(228) 8 DEV67
560(230) 8 DEV68
568(238) 8 DEV69
576(240) 8 DEV70
584(248) 8 DEV71
592(250) 8 DEV72
600(258) 8 DEV73
608(260) 8 DEV74
616(268) 4 WAIT
620(26C) 4 MPLKWTIM
624(270) 8 DKRD
632(278) 8 DKWR
640(280) 4 SVRD
644(284) 4 SVWR
648(288) 8 CPU
656(290) 8 REQ
664(298) 8 MOVE
672(2A0) 4 DUMP
676(2A4) 4 REST
680(2A8) 4 SLIC
684(2AC) 4 CNCT
688(2B0) 4 FBWT
692(2B4) 4 SWT
696(2B8) 4 ERRPDL
700(2BC) 4 MPLKPREM
704(2C0) 8 RECADD
712(2C8) 8 RECDEL
720(2D0) 8 BADD
728(2D8) 8 BDEL
736(2E0) 8 BCHG
744(2E8) 8 IXADD
752(2F0) 8 IXDEL
760(2F8) 8 FINDS
768(300) 4 SORTS
772(304) 4 Spare
776(308) 8 RECDS
784(310) 8 STRECDS
792(318) 8 DKAR
800(320) 8 DKPR
808(328) 4 DKRR
812(32C) 4 FBMX
816(330) 4 TFMX
820(334) 4 USMX
824(338) 4 SVMX
828(33C) 4 SFTRSTRT (number of soft restarts)
832(340) 8 APSYLD
840(348) 8 APSYLDD
848(350) 4 APSYLDT
852(354) 4 Spare
856(358) 8 DKPRF
864(360) 4 SMPLS
868(364) 4 USRS
872(368) 4 SVAC
876(36C) 4 RUNG
880(370) 4 REDY
884(374) 4 BLKI
888(378) 4 WTSV
892(37C) 4 BLKO
896(380) 4 SWPG
900(384) 4 PCPU
904(388) 4 DIRRCD
908(38C) 4 Spare
912(390) 8 STCPU
920(398) 8 STDEQ
928(3A0) 8 STWAIT
936(3A8) 4 STPOST
940(3AC) 4 LKWAIT
944(3B0) 4 LKPOST
948(3B4) 4 RSXCOMP
952(3B8) 4 SCHDCPU
956(3BC) 4 SCREENS
960(3C0) 4 STIMERS
964(3C4) 4 HRDRSTRT (number of hard restarts)
968(3C8) 8 SVPAGES
976(3D0) 4 PBRSFLT
980(3D4) 4 COMMITS
984(3D8) 4 BACKOUTS
988(3DC) 4 LONGUPDTS
992(3E0) 4 LONGUPDTIME(MS)
996(3E4) 4 MPHASHD
1000(3E8) 4 CDLWAIT
1004(3EC) 4 MQGETS
1008(3F0) 4 MQPUTS
1012(3F4) 4 UBUFHWS
1016(3F8) 4 MQHWTASK
1020(3FC) 4 Spare
1024(400) 8 MQBYTEIN
1032(408) 8 MQBYTEOU
1040(410) 4 MQHWQU
1044(414) 4 MQNUMQU
1048(418) 4 MQNUMQM
1052(41C) 4 Spare
1056(420) 8 MQAPITIM
1064(428) 8 MQAPICNT
1072(430) 8 MQGWTTIM
1080(438) 4 MQGWTCNT
1084(43C) 4 Spare
1088(440) 8 MQGWTTSP
1096(448) 4 MQGWTSUC
1100(44C) 4 ECLOAD
1104(450) 4 ECDELETE
1108(454) 4 Spare
1112(458) 8 ECCALL
1120(460) 4 ECCWAITM
1124(464) 4 ECCWAITS
1128(468) 4 ECTWAITM
1132(46C) 4 ECTWAITS
1136(470) 4 ECCTOUT
1140(474) 4 ECCNCT
1144(478) 4 ECMODMAX
1148(47C) 4 ECNAMMAX
1152(480) 4 ECTSKMAX
1156(484) 4 GTBLRU
1160(488) 4 GTBLRS
1164(48C) 4 TSMX
1168(490) 4 TEMX
1172(494) 4 MAXIOX
1176(498) 8 DKRDL
1184(4A0) 8 DKWRL
1192(4A8) 8 CPUTOTZE
1200(4B0) 8 CPUONZIP
1208(4B8) 4 MPR

as of V7.7 with 77z257 applied

1212(4BC) 4 MBO

as of V7.7 with 77z257 applied

1216(4C0) 4 MCP

as of V7.7 with 77z257 applied

1220(4C4) 4 RCV

as of V7.7 with 77z257 applied

1224(4C8) 8 Spare
1232(4D0) 8 DKWTMICS

as of V7.7, with 77z339 applied

1240(4D8) 8 DKWTS

as of V7.7, with 77z339 applied

1248(4E0) 8 MTDEQ
1256(4E8) 8 MTSDEQ
1264(4F0) 8 MTZDEQ
1272(4F8) 8 STZDEQ
1280(500) 8 ZTDEQ

System performance statistics

Entry format

For subtype X'02'(system performance statistics), the layout of the remainder of the entry is shown in the following table. Each statistic in the table is: Length, 4; Data type, Binary. See Statistics with descriptions table for the meaning of each statistic.

Layout of system subtype X'02' entries
Offset dec(hex) Statistic
16(10) SMPLS
20(14) USRS
24(18) SVAC
28(1C) RUNG
32(20) REDY
36(24) BLKI
40(28) WTSV
44(2C) BLKO
48(30) SWPG

Additional disk buffer monitor statistics

For subtype X'04' (additional disk buffer monitor statistics), the layout of the entry is shown in the following table. Each statistic in the table is: Length, 8; Data type, Binary. For a description of each statistic, see the Statistics with descriptions table.

Disk buffer monitor statistics
Offset dec(hex) Disk buffer monitor statistic
16(10) Spare
24(18) DKSTKQC
32(20) DKSWRP
40(28) DKSWRPT
48(30) DKSDIR
56(38) DKSDIRT
64(40) DKSKIP
72(48) Spare
80(50) Spare
88(58) DKSRR
96(60) DKSFBS
104(68) DKSKIPT
112(70) DKSAWT
120(78) DKSAWW
128(80) DKSAWWL
136(88) DKSAWB
144(90) DKSAWBL
152(98) DKSRRFND
160(A0) DKSTBLF
168(A8) DKSTBLA
176(B0) DKSTBLB
184(B8) DKSTBLC
192(C0) DKSTBLD
200(C8) Spare
208(D0) Spare
216(D8) Spare
224(E0) Spare
232(E8) Spare
240(F0) Spare
248(F8) Spare
256(100) DKSTBLX
264(108) DKSTBLE
272(110) Spare

Multiprocessing (MP) subtask statistics

The following statistics apply only to installations configured for multiprocessing using MP/204. For subtype X'08' (subtask specific statistics), the layout of the entry is shown in the following table. For the meaning of each statistic, see the Statistics with descriptions table.

MP subtask statistics and offsets
MP statistic Offset dec(hex) Length
CPU 16(10) 8
PR 24(18) 8
PCPU 32(20) 4
STWAIT 36(24) 4
STPOST 40(28) 4
LKWAIT 44(2C) 4
LKPOST 48(30) 4
SPARE 52(34) 4
MQWTM 56(38) 8
DEQ 64(40) 8
ZTDEQ 72(48) 8
STDEQ 80(50) 8
LWTIM 88(58) 4
PETIM 92(5C) 4
ZCPU 96(60) 8

User statistics entries (Type 9)

User lines report on an individual user's activity. These statistics include:

  • Final (logout) and partial statistics
  • Since-last and ordered index statistics
  • Performance statistics

Each user line starts with the following statistics:

ST $$$ USERID='userid' ACCOUNT='account'

The subtype, in this case PERFORMANCE, follows USERID and ACCOUNT statistics:

ST $$$ USERID='userid' ACCOUNT='account' PERFORMANCE=n

The format of user statistics entries is shown in the following table:

User statistics entry formats
Offset dec(hex) Length Data type Description
0(0) 2 Binary Length of entry
2(2) 1 Binary Type indicator = X'09'
3(3) 1 Binary Subtype indicator:
  • X'00' = logout statistics
  • X'01' = since-last statistics
  • X'02' = partial statistics
  • X'04' = performance statistics
  • X'81' = since-last statistics including conflict statistics
  • X'41' = user restart forced out a since-last statistics record.
  • X'C1' = user restart forced out a since last statistics record which included conflict statistics
4(4) 4 Unsigned packed data Julian date = 0CYYDDDF
8(8) 4 Unsigned packed data Time = HHMMSSTH
12(0C) 2 Binary Server number
14(0E) 2 Binary User number
16(10) 10(0A) EBCDIC

Account:

  • Second argument of LOGON/LOGIN command
  • Default is user ID
26(1A) 10(0A) EBCDIC User ID
36(24) 4 EBCDIC Since-last ID = See the Activity types table where the possible activity types are listed.
40(28) 1 Binary IODEV number (added for MXI)
41(29) 3 N/A Unused
44(2C) to end of journal   Binary

Contents depend on subtypes, shown in:

User final and partial statistics

Final (logout) statistics

Final (logout) statistics, listed in the Layout of user subtype X'00' and X'02' entries table, summarize a user's activities since the last login. A LOGOUT statistics line, such as the following, is written after each logout if the parameter following ACCOUNT is not PARTIAL, LAST, or PERFORMANCE:

yydddhhmmssnnnsssuuuuu ST $$$ ACCOUNT='NO ACCOUNT' SQRD=5 ...

  • yyddd is the year and Julian day.
  • hhmmss is the time of day in hours, minutes, and seconds.
  • nnn is a counter to distinguish lines produced in the same second.
  • sss is the number of the server currently handling the user. Leading zeros are suppressed.
  • uuuuu is a 5-digit user number associated with the audit trail line. Leading zeros are suppressed.

In some cases, the account name appears as "NO USERID NO ACCOUNT". Statistics in this category are not chargeable to any one user and should be considered system overhead.

Final lines are recognized by the absence of an indicator for another type of line.

Partial statistics

Optional partial statistics, listed in the "Layout of user subtype X'00' and X'02' entries" table below, can be provided for a run up to a few minutes before a system failure that prevents normal termination and production of final user and system statistics.

User partial statistics lines, such as the following, are made available for active users:

ST $$$ ACCOUNT='NO ACCOUNT' PARTIAL=21 ...

Statistics accumulate either:

  • As long as the user is logged in

or

  • From the first terminal input to disconnection when a login is not required

Partial statistics are requested by setting the time interval, in minutes, between successive partial lines on User 0's parameter line with the ACCTIM parameter. ACCTIM must be nonzero.

The contents of the statistics portion of the entry, starting at offset 40(28), depends upon the subtype of the entry.

Partial statistics lines are indicated on the journal by the parameter PARTIAL=n, where n is a counter from the beginning of the run. PARTIAL=n immediately follows the type parameter.

For subtypes X'00' (logout statistics), and X'02' (partial statistics), the layout of the remainder of the entry is shown in the following table. Each statistic in the table is: Length, 4; Data type, Binary. For the meaning of each statistic, see the Statistics with descriptions table.

Layout of user subtype X'00' and X'02' entries
Offset dec(hex) Statistic
40(28) CNCT
44(2C) DKRD
48(30) DKWR
52(34) SQRD
56(38) SQWR
60(3C) SGMTI
64(40) SGMTO
68(44) SVRD
72(48) SVWR
76(4C) CPU
80(50) REQ
84(54) MOVE
88(58) DUMP
92(5C) REST
96(60) SLIC
100(64) AUDIT
104(68) WAIT
108(6C) FBWT
112(70) UDD
116(74) RECADD
120(78) RECDEL
124(7C) BADD
128(80) BDEL
132(84) BCHG
136(88) IXADD
140(8C) IXDEL
144(90) FINDS
148(94) SORTS
152(98) RECDS
156(9C) STRECDS
160(A0) DKAR
164(A4) DKPR
168(A8) DKRR
172(AC) COMMITS
176(B0) BACKOUTS
180(B4) UPDTTIME(MS)
184(B8) LONGUPDTS
188(BC) LONGUPDTIME(MS)
192(C0) SMPLS
196(C4) RUNG
200(C8) REDY
204(CC) BLKI
208(D0) WTSV
212(D4) BLKO
216(D8) SWPG
220(DC) PCPU
224(E0) DIRRCD
228(E4) BXCHNG
232(E8) BXDELE
236(EC) BXNEXT
240(F0) BXFIND
244(F4) BXINSE
248(F8) BXSPLI
252(FC) BXRFND
256(100) BXFREE
260(104) STCPU
264(108) STDEQ
268(10C) SCHDCPU
272(110) SCREENS
276(114) SVPAGES
280(118) PBRSFLT
284(11C) MQGETS
288(120) MQPUTS
292(124) UBUFHWS
296(128) MQHWTASK
300(12C) MQBYTEIN
304(130) MQBYTOU
308((134) MQHWQU
312(138) MQNUMQU
316(13C) MQNUMQM
320(140) MQAPITIM
324(144) MQAPICNT
328(148) MQGWTTIM
332(14C) MQGWTCNT
336(150) MQGWTTSP
340(154) MQGWTSUC
344(158) ECLOAD
348(15C) ECDELETE
352(160) ECCALL
356(164) ECCWAITM
360(168) ECCWAITS
364(16C) ECTWAITM
368(170) ECTWAITS
372(174) ECCTOUT
376(178) ECCNCT
380(17C) GTBLRU
384(180) GTBLRS
388(184) FSCBSW
392(188) Spare
396(18C) MPR (V7.7 onwards)
400(190) MBO (V7.7 onwards)
404(194) DEV#
408(198) ZTCPU
412(19C) ZTDEQ
416(1A0) CCATDIFF

User since-last statistics

Since-last statistics

Statistics are kept for each SOUL request, host language interface call, and a number of Model 204 system commands. When each request or command completes, a since-last statistics line is written. The line includes the ACCOUNT parameter at the beginning, followed by the IODEV parameter and the LAST parameter.

Several MONITOR commands produce displays containing the user's last or current activity in the FUNC column. When this activity completes, statistics are also written to CCAAUDIT and the activity is described in ST lines as LAST='activity'.

Syntax

ST $$$ USERID='userid' ACCOUNT='accountname' IODEV='devicetype' LAST='acty'

Where:

acty can be one of the following:

Activity types
Option Meaning
BLDR BLDREUSE command
BLDX Z command
CMPL Compilation of a SOUL request
CPTB COMPACTB command
CPTE COMPACTE command
CPYP COPY PROCEDURE command
CREA CREATE FILE filename command
DISP DISPLAY PROCEDURE command
DUMP DUMP file command
EDIT Editing a request
EVAL Evaluation of a SOUL request
EXEC Execution of an SQL request
FLDC Compilation of a FLOD procedure
FLDE Evaluation of a FLOD procedure
IMPC Implied commit
INCR IN FILE filename INCREASE DATASETS command
INIT IN FILE filename INITIALIZE command
LOAD Loading a precompiled (ASPY) procedure
PREP Compilation of an SQL request
REST RESTORE file command

Since-last highwater marks

The since-last statistics include highwater marks of various work tables used by SOUL requests. Values are reported in terms of the units in which each table is allocated. For example, QTBL is allocated in units of 16 bytes.

Since-last in relation to final and partial statistics

Many since-last statistics, such as CNCT, CPU, and DKRD, are also reported on user final and partial statistics lines. (See the Statistics with descriptions table.) The statistics on these lines are roughly equal to the sum of all of the preceding since-last statistics for the user. Because since-last statistics are not maintained for every Model 204 command, final statistics are generally larger.

For subtype X'01' (since-last statistics), the layout of the remainder of the entry begins as shown in the following table. For the meaning of each statistic in this table and in the Layout of user subtype X'01' entries (CFRJRNL=0) table, see the Statistics with descriptions table. The layout of the end of the entry depends on whether you chose to collect conflict statistics:

Layout of user subtype X'01' and X'81' entries
Offset dec(hex) Statistic
44(2C) NTBL
48(30) GTBL
52(34) QTBL
56(38) STBL
60(3C) TTBL
64(40) VTBL
68(44) PDL
72(48) FTBL
76(4C) XTBL
80(50) ITBL
84(54) FSCB
88(58) OUTPB
92(5C) HEAP
96(60) SQLI
100(64) SQLO
104(68) CNCT
108(6C) CPU
112(70) DKRD
116(74) DKWR
120(78) UDD
124(7C) OUT
128(80) SLIC
132(84) IN
136(88) RECADD
140(8C) RECDEL
144(90) BADD
148(94) BDEL
152(98) BCHG
156(9C) IXADD
160(A0) IXDEL
164(A4) FINDS
168(A8) SORTS
172(AC) RECDS
176(B0) STRECDS
180(B4) PCPU
184(B8) RQTM
188(BC) DIRRCD
192(CO) BXCHNG
196(C4) BXDELE
200(C8) BXNEXT
204(CC) BXFIND
208(DO) BXINSE
212(D4) BXSPLI
216(D8) BXRFND
220(DC) BXFREE
224(E0) STCPU
228(E4) STDEQ
232(E8) SCHDCPU
236(EC) SCREENS
240(F0) SVRD
244(F4) SVWR
248(F8) DKPR
252(FC) SVPAGES
256(100) COMMITS
260(104) BACKOUTS
264(108) UPDTTIME(MS)
268(10C) LONGUPDTS
272(110) LONGUPDTIME(MS)
276(114) MQGETS
280(118) MQPUTS
284(11C) UBUFHWS
288(120) MQHWTASK
292(124) MQBYTEIN
296(128) MQBYTEOU
300(12C) MQHWQU
304(130) MQNUMQU
308(134) MQNUMQM
312(138) MQAPITIM
316(13C) MQAPICNT
320(140) MQGWTTIM
324(144) MQGWTCNT
328(148) MQGWTTSP
332(14C) MQGWTSUC
336(150) ECLOAD
340(154) ECDELETE
344(158) ECCALL
348(15C) ECCWAITM
352(160) ECCWAITS
356(164) ECTWAITM
360(168) ECTWAITS
364(16C) ECCTOUT
368(170) ECCNCT
372(174) GTBLRU
376(178) GTBLRS
380(17C) FSCBSW
384(180) OBJSWAP
388()184 SIRIUS02
392(188) SIRIUS03
396(18C) ZTCPU
400(190) ZTDEQ
404(194) CCATDIFF
408(198) CCATDIFH

Conflict statistics

Whether you are collecting conflict statistics determines the layout at the end of the entry. Two of the conflict statistics keep track of critical file-resource conflicts, and two keep track of record-locking conflicts. Whether these statistics are collected and journaled depends on settings of the parameters CFRLOOK and CFRJRNL.

These statistics are not displayed with the output of the MONITOR SL command. Instead, you can view them (with all the other since-last statistics) by issuing the MONITOR command with the keyword CONFLICT.

The remainder of the entry depends on the setting of CFRJRNL.

The following table shows the layout when CFRJRNL is set to 0:

Layout of user subtype X'01' entries (CFRJRNL=0)
Offset dec(hex) Length Data type Description
412(19C) 10 EBCDIC Subsystem name
422(1A6) 8 EBCDIC Procedure file name
430(1AE) 1 Binary Length of following procedure name
431(1AF) LAUDPROC EBCDIC Procedure name

The following table shows the layout when CFRJRNL is set to 1. It reflects the change in offset and that the user subtypes are now X'81' entries.

Layout of user subtype X'81' entries (CFRJRNL=1)
Offset dec(hex) Length Data type Entry type
412(19C) 4 Binary BLKCFRE stastic
416(1A0) 4 Binary BLKRLK statistic
420(1A4) 4 Binary WTCFR statistic
424(1A8) 4 Binary WTRLK statistic
428(1AC) 10 EBCDIC Subsystem name
438(1B6) 8 EBCDIC Procedure file name
446(1BE) 1 Binary Length of following procedure name
447(1BF) LAUDPROC EBCDIC Procedure name

Note: The conflict statistics, BLKCFRE, BLKRLK, WTCFR, and WTRLK, are inexact. Although a user might at times conflict with multiple other users, only the first conflict encountered is counted.

User performance statistics

Optional performance statistics lines provide information about the operation of the Model 204 scheduler during a short period of time. Information about the length of the scheduler queues, the number of active users and servers, and the queue distribution for each user, is obtained by sampling.

A performance line starts with:

ST $$$ USERID='userid' ACCOUNT='accountname' PERFORMANCE=n

Parameters to activate performance statistics

To request performance statistics, specify the number of samples to take between performance statistics lines (RPTCNT parameter) and the time interval between samples (SMPLTIM parameter) on User 0's parameter line. When you request scheduler and server performance data (SMPLTIM parameter), an asynchronous operating system task, called the performance subtask, is created. Both RPTCNT and SMPLTIM must be nonzero.

Sampling

The performance subtask spends most of its time in a wait state. Every SMPLTIM milliseconds, it records the current state of the system. Data from the most recent sample is available at an Online terminal through the MONITOR command, described in ONLINE monitoring.

The number of samples taken during a run (RPTCNT), times the number of users in the run (NUSERS), must be less than 4.2 million. In large systems that run for a very long period of time, the time interval between samples cannot be extremely short. If the limit is exceeded, an OC1 interrupt occurs when the SMPLS statistic value is calculated during the printing of the system statistics, either a performance or final line.

Layout of performance statistics

For subtype X'04' (performance statistics), the layout of the remainder of the entry is shown in the following table. Each statistic in the table is: Length, 4; Data type, Binary. For the description of each statistic, see the Statistics with descriptions table.

Layout of user subtype X'04' entries
Offset dec(hex) Statistic
40(28) SMPLS
44(2C) RUNG
48(30) REDY
52(34) BLKI
56(38) WTSV
60(3C) BLKO
64(40) SWPG

File statistics entries (Type 10)

File statistics, summarized in the File statistics entry formats table, are kept for every file opened during a run, including CCAGRP, CCASYS, and CCATEMP. The activity of a file for all users is summarized from the time the file is first opened until the last user issues a CLOSE command. The statistics appear after the last CLOSE command in lines that follow the date and time information (yyyydddhhmmssnnnsssuuuuu) with:

ST $$$ FILE='filename'

System termination writes file statistics for each file still open at the end of the run.

Final statistics

A final statistics line similar to the following is produced when the last user of a file closes the file:

ST $$$ FILE='PAYABLES'DKRD=25 DKRD=25 DKWR=15 REQ=3 RECADD=6 ...

Partial statistics

File partial statistics lines similar to the following are furnished for each open file:

ST $$$ FILE='PAYABLES'PARTIAL=3 ...

File statistics journal records layout

File statistics entries have the format shown in the following table. For a description of each statistic, see the Statistics with descriptions table.

File statistics entry formats
Offset dec(hex) Length Data type Type of entry
0(0) 2 Binary Length of entry
2(2) 1 Binary Type indicator = X'0A'
3(3) 1 Binary Subtype indicator:
  X'00' = file close
X'01' = partial
4(4) 4 Unsigned packed data Julian date = 0CYYDDDF
8(8) 4 Unsigned packed data Time = HHMMSSTH
12(0C) 4 Unused
16(10) 8 EBCDIC File name
24(18) 2 Unused

The following table lists the file statistics and the decimal(hexidecimal) offset. These statistics are data type Binary and eight bytes long.

File statistics
File statistic Offset dec(hex)
DKRD 26(1A)
DKWR 34(22)
REQ 42(2A)
RETRYA 50(32)
RETRYC 58(3A)
DUPDTS 66(42)
RECADD 74(4A)
RECDEL 82(52)
BADD 90(5A)
BDEL 98(62)
BCHG 106(6A)
IXADD 114(72)
IXDEL 122(7A)
DIRRCD 130(82)
BXCHNG 138(8A)
BXDELE 146(92)
BXNEXT 154(9A)
BXFIND 162(A2)
BXINSE 170(AA)
BXSPLI 178(B2)
BXRFND 186(BA)
BXFREE 194(C2)
UPDTTIME 202(CA)
PNDGTIME 210(D2)
DKUPTIME 218(DA)
COMMITS 226(E2)
BACKOUTS 234(EA)

Text entries (Type 11, Type 13)

Type 11 (X'0B') and Type 13 (X'0D') entries record most of the actual text of the audit trail.

Noncontinued text entries, Type 11 (X'0B')

Each Type 11 journal entry corresponds to one logical entry in the audit trail, even though it might generate more than one line of physical output in the audit trail.

The format of a Type 11 entry is shown in the following table:

Noncontinued text entry formats
Offset dec(hex) Length Data type Description
0(0) 2 Binary Length of entry
2(2) 1 Binary Type indicator = X'0B'
3(3) 1 Binary Subtype indicator:

X'01' = LI line

X'02' = LP line

X'03' = LR line

X'04' = CI line

X'05' = QT line

X'08' = CP line

X'10' = OI line

X'20' = OO line

X'40' = LS line

X'80' = CS line

4(4) 4 Unsigned packed data Julian date = 0CYYDDDF
8(8) 4 Unsigned packed data Time = HHMMSSTH
12(0C) 2 Binary Server number
14(0E) 2 Binary User number
16(10) Variable EBCDIC Text of audit trail entry

Possibly continued text entries, Type 13 (X'0D')

Each Type 13 journal entry corresponds to one, or part of one, logical entry in the audit trail. See the Types of audit trail lines table for a definition of the line types.

The format of a Type 13 entry is shown in the following table:

Possibly continued text entry formats
Offset dec(hex) Length Data type Description
0(0) 2 Binary Length of entry
2(2) 1 Binary Type indicator = X'0D'
3(3) 1 Binary

Subtype indicator:

X'01' = US line

X'02' = XX line

X'04' = RK line

X'08' = AD line

X'10' = MS line

X'20' = ER line

4(4) 4 Unsigned packed data Julian date = 0CYYDDDF
8(8) 4 Unsigned packed data Time = HHMMSSTH
12(0C) 2 Binary Server number
14(0E) 2 Binary User number
16(10) Variable EBCDIC Text of audit trail entry

Note: Subtype indicator X'80' summed with any subtype means continuation into the next entry.

Initialization entries, Type 12 (X'0C')

One Type 12 journal entry is written as the first nonheader entry in the journal during initialization. It contains basic information about the run being initialized.

The format of a Type 12 entry is shown in the following table:

Initialization entry formats
Offset dec(hex) Length Data type Description
0(0) 2 Binary Length of entry = X'001A'
2(2) 1 Binary Type indicator = X'0C'
3(3) 1 Unused
4(4) 4 Unsigned packed data Julian date = 0CYYDDDF
8(8) 4 Unsigned packed data Time = HHMMSSTH
12(0C) 8 EBCDIC Version ID (for example, 4.2.0G)
20(14) 2 Binary Spare for the run
22(16) 4 EBCDIC SMF ID (SYSID parameter)

Time-stamp entries, Type 14 (X'0E')

Type 14 entries are used to force time-stamps to the journal.

The format of a Type 14 entry is shown in the following table:

Time-stamp entry formats
Offset dec(hex) Length Data type Description
0(0) 2 Binary Length of entry = X'000C'
2(2) 1 Binary Type indicator = X'0E'
3(3) 1  — Unused
4(4) 4 Unsigned packed data Julian date = 0CYYDDDF
8(8) 4 Unsigned packed data Time = HHMMSSTH

Merged journal bracketing entries (Type 15)

Type 15 entries appear only in the output of the MERGEJ utility. See Using the MERGEJ utility.

The format of a Type 15 entry is shown in the following table:

Merged journal bracketing entry formats
Offset dec(hex) Length Data type Description
0(0) 2 Binary Length of entry = X'002E'
2(2) 1 Binary Type indicator = X'0F'
3(3) 1 Binary

Subtype indicator:

X'01' = begin bracket

X'02' = end bracket

4(4) 12(0C)  — Unused
16(10) 2(2) Binary Number of journals merged together
18(12) 28(1C) EBCDIC Text:
  • For subtype X'01' = BEGINNING OF MERGED JOURNAL
  • For subtype X'02' = END OF MERGED JOURNAL

System Management Facilities

You can write certain statistics to the System Management Facilities (SMF) data set by:

  • Setting the SMFSLRN and SMFLORN parameters on the User 0 command line to activate the SMF feature. (z/OS and z/VM)
  • Modifying the program that processes the SMF data set to handle the records written by Model 204. (z/VM only)

See the z/VM requirements and the the z/OS requirements sections for details.

Since-last statistics or logout statistics can be written concurrently with statistics that are written to the journal. Logout statistics are more complete than since-last statistics. Many more since-last records than logout records are usually written.

If the operating system crashes before a user logs out, the SMF data set has a since-last record for every request run, but no logout record.

File and system statistics cannot be written to the SMF data set.

During system initialization, Model 204 issues a message reporting the SMF system ID, the run's job name, the step name, and the JES ID. For operating systems other than z/OS, the JES ID is blank.

z/VM requirements

In the CMS interface to Model 204, you can incorporate a user-written accounting exit that invokes the ACCTEXIT routine to process each statistics record as it is written. The following considerations apply:

  • Take care to preserve the CMS environment in which Model 204 is running. Do not change the contents of general registers 13 and 14.
  • You must specify the SMFLORN and/or SMFSLRN parameters on User 0's parameter line to activate accounting record production.
  • TEXT file for the ACCTEXIT module must be available on an accessed disk when generating M204CMS using M204GEN. For complete installation procedures, see the Rocket Model 204 z/VM Installation Guide.

z/OS requirements

To write since-last SMF records in a z/OS environment, activate the SMF feature.

To activate SMF, set the following parameters on User 0's parameter line:

  • SMFSLRN — SMF record number for since-last records
  • SMFLORN — SMF record number for logout records

The following considerations apply:

  • Since-last SMF records are written only if the SMFSLRN parameter is set to a non-zero value at the beginning of the run.
  • Logout SMF records are written only if the SMFLORN parameter is set to a non-zero value at the beginning of the run.
  • If both parameters are set, each type of record is written.

SMF record numbers, which identify the record type to the programs that process the SMF data set, are written as part of the header of the SMF records. The record numbers must be between 128 and 255.

See the "System Management Facility record layout and statistics" section, which follows.

System Management Facility record layout and statistics

SMF logout record layout

The following table shows the basic format of the Model 204 SMF logout record:

Basic format of Model 204 SMF logout record
Offset(hex) Length Description
0(0) 2 Length of record
2(2) 2 Zeros
4(4) 1 System configuration
5(5) 1 Record number (SMFSLRN or SMFLORN)
6(6) 8 Time and date of this record

4 bytes, time in hundredths of a second from midnight

4 bytes, Julian date, packed decimal

14(E) 4 System identification
18(12) 8 Job name
26(1A) 8 Time and date of login

4 bytes, time in hundredths of a second from midnight

4 bytes, Julian date, packed decimal

34(22) 10 Login user ID
44(2C) 10 Login account
54(36) 2 Binary user number
56(38) 1 Unused (reserved)
57(39) 1 Model 204 version number in hex:

For V7.4: x'4A' (requires zap 74Z4196)

For V7.5: x'4B' (requires zap 75Z326)

For V7.6: x'4C'

For V7.7: x'4D'

For V7.8: x'4E'

58(3A) 2 Unused (reserved)
60(3C)   Binary fullwords of statistics
  • Statistics depend upon the record number.
  • The SMF system identification field (offset 16) is the site-definable 4-character field retrieved from the SMF control area.

    The SMF control area field contains the CPU identification string that is stored at system IPL time, and it can be used to determine on which CPU Model 204 is running in a multi-CPU environment.

Offsets for SMF logout record statistics

The following table shows the layout of SMF logout record statistics. Each statistic in the table is: Length, 4 (bytes); data type, Binary. This record has the same statistics in the same order as the CCAJRNL logout record, only the offsets are different. For a description of each statistic, see the Description of statistics table.

Offset location of SMF logout record statistics
Offset dec(hex) Statistic
60(3C) CNCT
64(40) DKRD
68(44) DKWR
72(48) SQRD
76(4C) SQWR
80(50) SGMTI
84(54) SGMTO
88(58) SVRD
92(5C) SVWR
96(60) CPU
100(64) REQ
104(68) MOVE
108(6C) DUMP
112(70) REST
116(74) SLIC
120(78) AUDIT
124(7C) WAIT
128(80) FBWT
132(84) UDD
136(88) RECADD
140(8C) RECDEL
144(90) BADD
148(94) BDEL
152(98) BCHG
156(9C) IXADD
160(A0) IXDEL
164(A4) FINDS
168(A8) SORTS
172(AC) RECDS
176(B0) STRECDS
180(B4) DKAR
184(B8) DKPR
188(BC) DKRR
192(C0) COMMITS
196(C4) BACKOUTS
200(C8) UPDTTIME(MS)
204(CC) LONGUPDTS
208(D0) LONGUPDTIME(MS)
212(D4) SMPLS
216(D8) RUNG
220(DC) REDY
224(E0) BLKI
228(E4) WTSV
232(E8) BLKO
236(EC) SWPG
240(F0) PCPU
244(F4) DIRRCD
248(F8) BXCHNG
252(FC) BXDELE
256(100) BXNEXT
260(104) BXFIND
264(108) BXINSE
268(10C) BXSPLI
272(110) BXFND
276(114) BXFREE
280(118) STCPU
284(11C) STDEQ
288(120) SCHDCPU
292(124) SCREENS
296(128) SVPAGES
300(12C) PBRSFLT
304(130) MQGETS
308(134) MQPUTS
312(138) UBUFHWS
316(13C) MQHWTASK
320(140) MQBYTEIN
324(144) MQBYTEOU
328(148) MQHWQU
332(14C) MQNUMQU
336(150) MQNUMQM
340(154) MQAPITIM
344(158) MQAPICNT
348(15C) MQGWTTIM
352(160) MQGWTCNT
356(164) MQGWTTSP
360(168) MQGWTSUC
364(16C) ECLOAD
368(170) ECDELETE
372(174) ECCALL
376(178) ECCWAITM
380(17C) ECCWAITS
384(180) ECTWAITM
388(184) ECTWAITS
392(188) ECCTOUT
396(18C) ECCNCT
400(190) GTBLRU
404(194) GTBLRS
408(198) FSCBSW
412(19C) Spare
416(1A0) Spare
420(1A4) Spare
424(1A8) ZTCPU
428(1AC) ZTDEQ
432(1B0) Spare
436(1B4) Spare
440(1B8) Spare

SMF since-last record layout

The following table shows the basic format of since-last Model 204 SMF records.

Since-last record basic format
Offset(hex) Length (bytes) Description
 0(0) 2 Length of record
 2(2) 2 Zeros
 4(4) 1 System configuration
 5(5) 1 Record number (SMFSLRN)
 6(6) 8 Time and date of this record

4 bytes, time in hundredths of a second from midnight

4 bytes, Julian date, packed decimal

14(E) 4 System identification
18(12) 8 Job name
26(1A) 8 Time and date of login

4 bytes, time in hundredths of a second from midnight

4 bytes, Julian date, packed decimal

34(22) 10 Login user ID
44(2C) 10 Login account
54(36) 2 Binary user number
56(38) 1 SMFSLTYP - X'01' if CRFJRNL=1 (conflict counters present)

SMFSLTYP - X'00' if CRFJRNL=0 (conflict counters not present)

57(39) 1 Model 204 version number in hex:

For V7.4: x'4A' (requires zap 74Z4196)

For V7.5: x'4B' (requires zap 75Z326)

For V7.6: x'4C'

For V7.7: x'4D'

For V7.8: x'4E'

58(3A) 2 Unused (reserved)
60(3C) 4 Since-last type (EVAL CMPL DUMP REST...)
64(40) 8 File name
72(48) 10 Since-last subsystem name
82(52) 40 Since-last procedure name
122(7A) 2 Reserved two bytes
124(7C)   Binary fullwords of statistics
  • Statistics depend upon the record number.
  • The SMF system identification field (offset 16) is the site-definable 4-character field retrieved from the SMF control area.
  • The SMF control area field contains the CPU identification string that is stored at system IPL time, and it can be used to determine on which CPU Model 204 is running in a multi-CPU environment.

Offsets for SMF since-last record statistics

The following table shows the layout of SMF since-last record statistics. Each statistic in the table is: Length, 4 (bytes); Data type, Binary. This record has the same statistics in the same order as the CCAJRNL user since-last records; only the offsets are different. For a description of each statistic, see the Description of statistics table.

SMF since-last record statistics
Offset dec(hex) Description
124(7C) NTBL
128(80) GTBL
132(84) QTBL
136(88) STBL
140(8C) TTBL
144(90) VTBL
148(94) PDL
152(98) FTBL
156(9C) XTBL
160(A0) ITBL
164(A4) FSCB
168(A8) OUTPB
172(AC) HEAP
176(B0) SQLI
180(B4) SQLO
184(B8) CNCT
188(BC) CPU
192(C0) DKRD
196(C4) DKWR
200(C8) UDD
204(CC) OUT
208(D0) SLIC
212(D4) IN
216(D8) RECADD
220(DC) RECDEL
224(E0) BADD
228(E4) BDEL
232(E8) BCHG
236(EC) IXADD
240(F0) IXDEL
244(F4) FINDS
248(F8) SORT
252(FC) RECDS
256(100) STRECDS
260(104) PCPU
264(108) RQTM
268(10C) DIRRCD
272(110) BXCHNG
276(114) BXDELE
280(118) BXNEXT
284(11C) BXFIND
288(120) BXINSE
292(124) BXSPLI
296(128) BXRFND
300(12C) BXFREE
304(130) STCPU
308(134) STDEQ
312(138) SCHDCPU
316(13C) SCREENS
320(140) SVRD
324(144) SVWR
328(148) DKPR
332(14C) SVPAGES
336(150) COMMITS
340(154) BACKOUTS
344(158) UPDTTIME(MS)
348(15C) LONGUPDTS
352(160) LONGUPDTIME(MS)
356(164) MQGETS
360(168) MQPUTS
364(16C) UBUFHWS
368(170) MQHWTASK
372(174) MQBYTEIN
376(178) MQBYTEOU
380(17C) MQHWQU
384(180) MQNUMQU
388(184) MQNUMQM
392(188) MQAPITIM
396(18C) MAPICNT
400(190) MQGWTTIM
404(194) MQGWTCNT
408(198) MQGWTTSP
412(19C) MQGWTSUC
416(1A0) ECLOAD
420(1A4) ECDELETE
424(1A8) ECCALL
428(1AC) ECCWAITM
432(1B0) ECCWAITS
436(1B4) ECTWAITM
440(1B8) ECTWAITS
444(1BC) ECCTOUT
448(1C0) ECCNCT
452(1C4) GTBLRU
456(1C8) GTBLRS
460(1CC) FSCBSW
464(1D0) Sirius
468(1D4) Sirius
472(1D8) Sirius
476(1DC) ZTCPU
480(1E0) ZTDEQ
484(1E4) Spare
488(1E8) Spare

The following table lists the conflict statistics, if CFRJRNL=1:

Conflict statistics, if CFRJRNL=1
Offset dec(hex) Description
492(1EC) BLKCFRE
496(1F0) BLKRLK
500(1F4) WTCFR
504(1F8) WTRLK