Summary analysis groups the essential data of given areas, e.g. for catalog management or PCS monitoring purposes. It provides a quick overview of system performance. The monitoring data is output in the form of tables.
For some monitoring data, only the average for the whole analysis period is provided (mean values), while in other cases the minimum and maximum values are provided as well.
Format
PRINT-SUMMARY | ||||||||||||||||
| ||||||||||||||||
Operands
PARTITION =
Defines the number of output operations. Outputs are made at regular intervals as soon as the specified period has elapsed, or more specifically, at the end of the last SM2 monitoring cycle to be started.
PARTITION = *EVALUATION-PERIOD
Only one output is made at the end of the analysis period (or at the end of the file if the file does not cover the whole analysis period).
PARTITION = <integer 1..1000>
Defines the period between the individual outputs.
DIMENSION = *MINUTES / *SECONDS / *HOURS / *DAYS
Defines the time units which determine the period between the individual outputs. If *HOURS or *DAYS is specified, this also stipulates full hours/days for the beginning of the first output cycle.
SUMMARY-REPORTS = *ACTIVITY / *ALL / list-poss(6): *ACTIVITY / *CMS / *DAB / *PCS / *POSIX / *UTM
Defines the SUMMARY reports to be output. These are the ACTIVITY, CMS, DAB, PCS, POSIX and UTM reports. If the user specifies the operand value *ALL, all the existing SUMMARY reports are output.
Example
//PRINT-SUMMARY PARTITION=60,SUMMARY-REPORTS=*PCS
Within the analysis period, this statement supplies a PCS summary report every 60 minutes.
SUMMARY ACTIVITY report
The summary report on activity contains an overview of important data monitored during the analysis period and combined in various groups. For most of these values, the minimum/maximum values and the standard deviation are output in addition to the average over the observed period.
REPORTING PERIOD
SYSTEM DATA
#TASKS
QUEUE STATISTICS
RESPONSE TIME STATISTICS
CPU STATISTICS
IO STATISTICS
MEMORY STATISTICS
PAGINGAREA STATISTICS
CHANNEL STATISTICS
DEVICE STATISTICS
All monitoring data refers to the specified analysis period.
REPORTING PERIOD
The first and last time stamp found within the analysis period are output. The BY entry corresponds to the SM2 gathering cycle. FOCUS and EXCEPT specifications valid for the analysis period must be taken from the preceding cover page.
If EXCEPT has been specified for the analysis, the statistics contain the message *** EXCEPTION PERIOD EXISTS ***.
Next, the system data applying to the appropriate partition is output:
SYSTEM DATA
For a data description see section "Monitoring environment output".
Then the following monitoring data is output:
#TASKS
This lists the average number of tasks within the analysis period. Assignment to the respective categories occurs at task creation.
Monitored variable | Meaning |
SYS | See report 5 |
BATCH | See report 5 |
DIALOG | See report 5 |
TP | See report 5 |
QUEUE STATISTICS
The number of queued tasks is output for the entire system, i.e. all categories are consolidated.
Monitored variable | Meaning |
IN CPU QUEUE | See report 31 |
IN PAGING QUEUE | See report 31 |
IN IO QUEUE | See report 31 |
ACTIVE | See report 30 |
INACT READY | See report 30 |
NOT ADMITTED | See report 30 |
INACT NOT READY | See report 30 |
TOTAL | See report 28 |
RESPONSE TIME STATISTICS
The monitoring data is not output unless the RESPONSETIME monitoring program was active during the analysis period. If only one definition was active, blanks are output for the other definitions. The output comprises system-global values, i.e. the think, response and transaction times over all connections.
Monitored variable | Meaning |
MEAN THINK TIME | See report 46 |
MEAN RESPONSE TIME (1) | See report 47 |
MEAN RESPONSE TIME (2) | See report 48 |
MEAN TRANSACTION TIME | See report 49 |
TRANSACTION RATE | See report 50 |
RESPONSE RATE (1) | See report 50 |
RESPONSE RATE (2) | See report 50 |
CPU STATISTICS
For multiprocessor serverss, the average value of all processors is output.
Monitored variable | Meaning |
TU TIME | See report 1 |
TPR TIME | See report 1 |
SIH TIME | See report 1 |
IDLE TIME | See report 1 |
STOP TIME | See report 1 |
ACTIVE LOGICAL MACHINES | See report 6 |
IO STATISTICS
For multiprocessor servers, the total of all processors is output.
Monitored variable | Meaning |
NON PAGING DISK IO' S | See report 3 |
PAGING IO' S | See report 3 |
TAPE IO' S | See report 3 |
PRINTER IO' S | See report 3 |
OTHER | See report 3 |
MEMORY STATISTICS
For the page fault statistics and the quotient from working set and available main memory, only the values for AVERAGE are output.
Monitored variable | Meaning |
| #CLASS 3 PAGES | See report 53 |
#CLASS 4 PAGES | See report 53 |
AVAILABLE PAGES (NPP) | See report 15 |
WSET ACT TASKS (PPC) | See report 15 |
WSET INACT READY TASKS (PPC) | See report 15 |
USED PAGES ACT TASKS | See report 58 |
USED PAGES INACT TASKS | See report 58 |
(WSET ACT + INACT READY)/NPP | Quotient from the planned working set of active and inactive ready tasks and the available main memory |
TOTAL # PAGE FAULTS | See report 54 |
# PAGE RECLAIMS | See report 54 |
# FIRST PAGE ACCESS | See report 54 |
# PAGE WRITES TO DISK | See report 55 |
# PAGE READS FROM DISK | See report 56 |
PAGINGAREA STATISTICS
Monitored variable | Meaning |
PAGES ON PAGING DEVICE(S) | See report 9 |
USED PAGES ON PAGING DEVICE(S) | See report 9 |
CHANNEL STATISTICS
The values for the ten most heavily loaded channels are output in the channel statistics. The values for CHANNEL IO RATE and CHANNEL PAM PAGES are only output if the CHANNEL-IO monitoring program was active during the monitoring period.
Monitored variable | Meaning |
CHANNEL BUSY STATE | See report 10 |
CHANNEL IO RATE | See report 102 |
CHANNEL PAM PAGES | See report 101 |
DEVICE STATISTICS
The values for the ten most heavily loaded devices are output in the device statistics. To this end, the value for DEVICE BUSY STATE (NON PAGING) is used.
Monitored variable | Meaning |
DEVICE BUSY STATE (NON PAGING) | See report 11 |
DEVICE BUSY STATE (PAGING) | See report 11 |
DEVICE IO RATE | See report 35 |
DEVICE PAM PAGES | See report 100 |
CMS summary report
All monitored catalog management data is analyzed if the user specifies the operand value *CMS.
Up to five catalog identifiers are output on one page. The catalog identifiers are sorted in alphabetical order.
The text POOL OF PRIVATE DISKS is output instead of the $ character (see reports 66 to 72) to denote the complete range of private disks.
If available, monitoring data for private disks is always output first. The average as well as the minimum and the maximum values are supplied for all monitored variables.
Monitored variable | Meaning |
LENGTH OF QUEUES | |
SERIAL | See report 66 |
REQUEST | See report 66 |
USERID | See report 66 |
BLOCK | See report 66 |
CATALOG ENTRY | See report 66 |
SPEEDCAT REQUEST | See report 185 |
CAT INDEX REQ | See report 185 |
ACCESSES [1/S] | |
PHYSICAL READ | See report 67 |
PSEUDO READ | See report 67 |
PHYSICAL WRITE | See report 67 |
IO-ERROR | See report 67 |
RESP. TIME [MS] | |
READ NO LBN | See report 72 |
READ LBN | See report 72 |
SCAN | See report 72 |
UPDATE / RENAME | See report 72 |
WRITE / CLEAR | See report 72 |
LOC FILE/JV [1/S] | |
READ NO LBN | See report 68 and 70 |
READ LBN | See report 68 and 70 |
SCAN | See report 68 and 70 |
UPDATE / RENAME | See report 68 and 70 |
WRITE / CLEAR | See report 68 and 70 |
REM FILE/JV [1/S] | |
READ NO LBN | See report 69 and 71 |
READ LBN | See report 69 and 71 |
SCAN | See report 69 and 71 |
UPDATE / RENAME | See report 69 and 71 |
WRITE / CLEAR | See report 69 and 71 |
SH FILE/JV [1/S] | |
READ NO LBN | See report 103 and 104 |
READ LBN | See report 103 and 104 |
SCAN | See report 103 and 104 |
UPDATE / RENAME | See report 103 and 104 |
WRITE / CLEAR | See report 103 and 104 |
DAB summary report
If the user specifies the operand value *DAB, the summary statistics for the various DAB cache areas are output.
As with all SM2R1 summaries, the average as well as the minimum and the maximum values are supplied for all monitoring data.
The following information is also output for each DAB cache area:
MEDIUM | Cache medium (MM = main memory) |
SIZE | Size of the DAB cache area in MB |
Monitored variable | Meaning |
DAB READ | |
REQUESTS | See report 79 |
HITS | See report 79 |
DAB WRITE | |
REQUESTS | See report 80 |
HITS | See report 80 |
PCS summary report
The operand value *PCS serves to select the summary statistics for the Performance Control Subsystem.
This report supplies more comprehensive data than the report group PCS. The categories are sorted in alphabetical order except for the standard categories SUM, SYS, DIALOG, BATCH and TP, which appear (in this order) at the beginning of the list.
Data not available for the pseudo-category SUM is replaced by blanks. The average as well as the minimum and the maximum values are supplied for all monitored variables.
As an additional option the name of the PCS option is output.
The following overview lists all the monitored data. The SUM column provides information about monitored variables that are output for the pseudo-category SUM.
Monitored variables | SUM | Meaning |
SERVICE-QUOTA- | ||
ACTUAL | See report 73 | |
PLANNED | See report 73 | |
MAX | See report 73 | |
MIN | See report 73 | |
REQUEST-DELAY- | ||
ACTUAL | x | See report 74 |
MAX | x | See report 74 |
MIN | See report 74 | |
THROUGHPUT QUOTA | x | Percentage specifying the throughput optimization share: 0% = no optimization (full response time optimization) 100% = full local optimization (no response time opt.) For SUM, this variable corresponds to the global PCS parameter with the same name |
DURATION | ||
# SERVICE UNITS | Number of used SERVICE UNITs after which an automatic category switch occurs. | |
RUNOUTS | x | See report 75 |
RUNOUTS PREEMT | x | See report 75 |
SERVICE-RATE- | ||
ACTUAL | x | See report 77 |
CPU | x | |
MEMORY | x | |
IO | x | |
PLANNED | x | See report 77 |
POSIX summary report
If the user specifies the operand value *POSIX, the summary statistics for POSIX are output.
As in the case of all other SM2R1 summaries, the mean, maximum and minimum values are supplied for all monitored variable.
Monitored variables | Meaning |
OPTION A | |
IGET | See report 141 |
NAMEI | See report 142 |
OPTION B | |
LREAD | See report 143 |
BREAD | See report 143 |
LWRITE | See report 144 |
BWRITE | See report 144 |
PREAD | See report 146 |
PWRITE | See report 146 |
OPTION C | |
SCALL | See report 147 |
SREAD | See report 148 |
SWRITE | See report 148 |
FORK | See report 148 |
EXEC | See report 148 |
RCHAR | See report 149 |
WCHAR | See report 149 |
OPTION M | |
MSG | See report 150 |
SEMA | See report 151 |
UTM summary report
If the user specifies the operand value *UTM, the summary statistics for the universal transaction monitor are output.
As with all SM2R1 summaries, the average as well as the minimum and the maximum values are supplied for all monitoring data.
The following information is also output for every application:
MODE | Mode of the UTM application: |
#DB | Number of database systems with which the application is coordinated. |
Monitored variables | Meaning |
# USERS | See report 129 |
# TASKS | See report 130 |
MAX # ASYN TASKS | See report 130 |
# DIAL STEPS [1/S] | See report 128 |
# DIAL TA | See report 128 |
# DIAL WITH DB | Number of dialog steps per second with database calls |
# DIAL WITH DDP | Number of dialog steps per second with distributed processing calls |
# ASYN CONV | Number of completed asynchronous conversations per second |
# ASYN TA | See report 128 |
# ASYN WITH DB | Number of asynchronous conversations with database calls per second |
# ASYN WITH DDP | Number of asynchronous conversations with distributed processing calls per second |
DIALOG [S/STEP] | |
TOTAL TIME | In the case of dialog steps: total time in seconds per dialog step from acceptance of the input by UTM to the sending of the dialog message by UTM. |
TOTAL T WITH DB | Value as for TOTAL TIME, but only for dialog steps with database calls |
TOTAL T WITH DDP | Value as for TOTAL TIME, but only for dialog steps with distributed processing calls |
TIME IN DB | Time UTM waits for execution of database calls per dialog step |
TIME IN DDP | Time UTM waits for a message from a remote application per dialog step |
TACCL WAIT TIME | Average time UTM waits as a result of TAC class bottlenecks per dialog step |
ASYNCHR. [S/CONV] | |
TOTAL TIME | Total time in seconds for asynchronous conversation from the beginning to the end of processing (excluding wait time before the start) |
TOTAL T WITH DB | Value as for TOTAL TIME, but only for asynchronous conversations with database calls |
TOTAL T WITH DDP | Value as for TOTAL TIME, but only for asynchronous conversations with distributed processing calls |
TIME IN DB | Time UTM waits for execution of database calls per asynchronous conversation |
TIME IN DDP | Time UTM waits for a message from a remote application per asynchronous conversation |
DIALOG [1/STEP] | |
# IO | Number of IOs of UTM tasks during processing of the dialog step, including subroutines of the user |
# IO IN DB | Number of IOs in the called DB systems per dialog step |
# DB CALLS | Number of DB calls per dialog step |
ASYNCHR. [1/CONV] | |
# IO | Number of IOs of UTM tasks during processing of the asynchronous operation, including subroutines of the user |
# IO IN DB | Number of IOs in the called DB systems per asynchronous operation |
# DB CALLS | Number of DB calls per asynchronous operation |
DIALOG [MS/STEP] | |
CPU TIME | CPU time used by UTM tasks in milliseconds per dialog step, including subroutines of the user |
CPU TIME IN DB | CPU time used in DB systems in milliseconds per dialog step |
ASYNCHR. [MS/CONV] | |
CPU TIME | CPU time used by UTM tasks in milliseconds per asynchronous operation, including subroutines of the user |
CPU TIME IN DB | CPU time used in DB systems in milliseconds per asynchronous operation |