Contents
SREPORT 003, the IDMS DC System Statistics report, summarizes all systemwide statistics and thus provides an overview of system performance.
Sample SREPORT 003:
REPORT NO. 03 IDMS-DC SYSTEM STATISTICS Rnn.n mm/dd/yy PAGE 1
SELECTED FROM: yyddd hh:mm TO: yyddd hh:mm
CV SYSTEM START: yyddd hh:mm TO: yyddd hh:mm
CV JOB NAME: jobname
CV NUMBER: nnn
SYSTEM STATISTICS
116 TOTAL TASKS 0 STD PGMPOOL LOADS
72 TOTAL SYSTEM TASKS 0 STD PGMPOOL WAITS
0 TASKS ABENDED 0 STD PGM PAGES LOADED
0 RUNAWAY TASKS ABORTED 29 RENTPOOL LOADS
0 TIMES AT MAX TASK 0 RENTPOOL WAITS
0 SHORT ON STORAGE 842 RENT PGM PGS LOADED
0 OVER RLE THRESH 0 XA PGMPOOL LOADS
0 OVER RCE THRESH 0 XA PGMPOOL WAITS
0 OVER DPE THRESH 0 XA PGM PGS LOADED
0 OVER ILE THRESH 169 XA RENTPOOL LOADS
0 STORAGE POOL WAITS 0 XA RENTPOOL WAITS
1,624 STG REQS - PASS 1 24,987 XA RENT PGS LOADED
1,103 STG REQS - PASS 2 0 PAGE RELEASE RQSTS
0 PUT JOURNALS 0 PAGES RELEASED
0 SET TIME WAITS 0 PAGE FIX RQSTS
236 SET TIME POSTS 0 PAGES PFIXED
1 SET TIME STRTTASKS 0 PAGE FREE RQSTS
234 SET TIME CANCELS 0 PAGES PGFREED
0 AUTOSTART TASKS
DC STATISTICS
19.8805 USER MODE CPU TIME 3.8358 SYSTEM MODE CPU TIME
659,356 DC SERVICE REQUESTS 64 GET SCRATCHES
659,928 DB SERVICE REQUESTS 106 PUT SCRATCHES
245 PROGRAMS CALLED 64 DELETE SCRATCHES
375 MAX # RLE'S USED 10 GET QUEUES
329 MAX # RCE'S USED 0 PUT QUEUES
308 MAX # DPE'S USED 0 DELETE QUEUES
1,337 STACK HI WATERMARK 8 GET TIMES
2,727 GET STORAGES 0 SET TIMES
2,494 FREE STORAGES
DC EXTENDED STATISTICS
40.806072 SYSTEM MODE CPU .000000 ZIIP ON CP CPU
.000000 ZIIP ON ZIIP CPU 22.013726 USER MODE CPU
62.819799 TOTAL TCB CPU .000000 ENCLAVE CPU
DB STATISTICS
2,336,618 PAGES REQUESTED 0 CALC RECS NO OFLOW
1,041,703 PAGES READ 0 CALC RECS OFLOW
0 PAGES WRITTEN 0 VIA RECS NO OFLOW
1,319,871 CALLS TO DBMS 0 VIA RECS OFLOW
1,296,108 RECORDS REQUESTED 0 FRAGMENTS STORED
1,294,797 RECORDS CURRENT OF RU 0 RECORDS RELOCATED
330 TOTAL LOCKS 0 RECORDS UPDATED
0 PAGES FOUND IN CACHE 0 PAGES IN PREFETCH BUFF
INDEX STATISTICS
0 SR8 SPLITS 0 SR8 STORES
0 SR8 SPAWNS 0 SR8 ERASES
0 ORPHANS ADOPTED 0 SR7 STORES
0 BTREE SEARCHES 0 SR7 ERASES
0 MIN LEVELS SEARCHED 0 TOTAL LEVELS SEARCHED
0 MAX LEVELS SEARCHED
SQL STATISTICS
0 SQL COMMANDS 0 TUPLES FETCHED
0 SORTS 0 ROWS INSERTED
0 TUPLES SORTED 0 ROWS UPDATED
0 MIN SORT 0 ROWS DELETED
0 MAX SORT 0 AM RECOMPILES
Field Descriptions
A description of the fields in the IDMS DC System Statistics report follows:
Specifies the time period specified on the USE 'SREPORT 000' parameter where SELECTED FROM specifies the beginning date and time and SELECTED TO specifies the ending date and time. The date is in Julian form, yyddd, where yy is the last two digits of the year and ddd is the day. The time is in hh:mm form, where hh is hours based on a 24-hour clock and mm is minutes.
Specifies the actual time period for the report generated. If the USE 'SREPORT 000' parameter does not specify a time range, CV SYSTEM START/INTERVAL START and TO represent the time range for the entire input archive file or input file created by SREPORT 099. When the input file contains cumulative statistics, the CV SYSTEM START date and time will be displayed. When the input file contains interval based statistics, the INTERVAL START date and time will be displayed.
Specifies the starting job name of Central Version which collected the statistics.
Specifies the number of Central Version which collected the statistics.
Indicates the total number of system and user tasks, including external request units, executed for this session.
Indicates the number of system tasks active at the conclusion of this session. System-initiated tasks include FACTOTUM, MASTER, DBRC, print, and line drivers.
Indicates the number of tasks that abended during this session.
Indicates the number of tasks that terminated abnormally because their execution time exceeded the maximum amount of time defined in the RUNAWAY INTERVAL clause of the SYSTEM statement.
Indicates the number of times a new task could not start because the system was processing the maximum number of tasks specified at system generation in the MAXIMUM ERUS and MAXIMUM TASK clauses of the SYSTEM statement.
Interpretation: Generally, this value should be close to 0. Values equal to 0 may indicate that the overall system size is larger than necessary; values greater than 0 may represent peaks in system loads or chronic system overload.
To reduce the number of times at maximum tasks, make one or more of the following system adjustments:
Indicates the number of times program storage requests were not satisfied. Generally, the value should be close to zero; values greater than zero should represent peaks in storage utilization, rather than chronic shortages.
Interpretation: Values consistently greater than zero indicate the size of the storage pool or storage cushion is too small. The size of the storage pool is defined at system generation with the STORAGE POOL and XA STORAGE POOL clauses of the SYSTEM system generation statement; the size of the cushion is defined with the CUSHION clause of the same statement.
Indicates how many times the number of resource link elements (RLEs) exceeded the value allocated in the SYSTEM statement at system generation. When the threshold is reached, the task that is executing abends; if the task is a system task, the system abends. Ideally, this value should be 0.
Indicates how many times the number of resource control elements (RCEs) exceeded the value allocated in the SYSTEM statement at system generation. When the threshold is reached, the task that is executing abends; if the task is a system task, the system abends. Ideally, this value should be 0.
Indicates how many times the number of deadlock prevention elements (DPEs) exceeded the value allocated in the SYSTEM statement at system generation. When the threshold is reached, the task that is executing abends; if the task is a system task, the system abends. Ideally, this value should be 0.
Indicates the number of time the internal lock elements (ILEs) exceeded their allocated value. When the threshold is reached, the task that is executing abends. If the task is a system task, the system abends. Ideally, this value should be zero (0).
Indicates the number of times tasks had to wait for a sufficient amount of contiguous storage. This value should be as low as possible; if the value is large or increasing, increase the size of the storage cushion in the CUSHION clause of the STORAGE POOL statement at system generation.
Indicates the number of storage requests that allocated space on a previously allocated page.
Indicates the number of storage requests that allocated space on a previously allocated page and on a contiguous new page.
Indicates the number of program requests to place a task in a wait state. The program issues the request with a SET TIMER WAIT DML statement.
Indicates the number of program requests to post a user-specified event control block (ECB) after the specified time interval elapses. The program issues the request with a SET TIMER POST DML statement.
Indicates the number of program requests to initiate a user-specified task after the specified time interval elapses. The program issues the request with a SET TIMER START DML statement.
Indicates the number of program requests to cancel the effect of a previously issued SET TIMER request. The program issues the request with a SET TIMER CANCEL DML statement.
(DC/UCF only) Indicates the number of times the associated task of each queue was invoked to process queue records. The task is invoked each time the queue threshold is exceeded. The threshold is defined with the THRESHOLD clause of the QUEUE system generation statement; the task is identified by the INVOKES TASK clause of the same statement.
Indicates the number of nonresident modules loaded into the 24-bit program pool; nonresident modules include programs, CA ADS and CA ADS batch dialogs, maps (DC/UCF only), and tables.
Indicates the number of times program loads were delayed due to insufficient space in the 24-bit program pool. Ideally, this value should be zero; a value greater than zero indicates insufficient space in the program pool for the volume of program load activity in the system.
Interpretation: If this value is high, the size of the 24-bit program pool should be increased or nonresident programs should be redefined as either resident or reentrant. Program pool usage can be observed dynamically by means of the OPER WATCH PR command. For more information about program pools, see the CA IDMS System Operations Guide.
Indicates the number of pages used by programs loaded into the 24-bit program pool. The size of a page equals 4K. Only one program can occupy a program pool page; that is, a 4.5K program uses two pages of program pool.
Indicates the number of reentrant programs loaded into the 24-bit reentrant pool; reentrant programs include CA ADS dialogs and subschemas.
Indicates the number of times reentrant program loads were delayed due to insufficient space in the 24-bit reentrant pool. Ideally, this value should be zero; a value greater than zero indicates the size of the reentrant pool should be increased.
Indicates the number of pages used by reentrant programs loaded into the 24-bit reentrant pool. A page is 512 bytes (0.5K).
(Extended addressing only) Indicates the number of nonresident modules loaded into the 31-bit XA program pool; nonresident modules include programs, subschemas, maps, database procedures, and tables that have been assigned an RMODE of ANY.
(Extended addressing only) Indicates the number of times program loads were delayed due to insufficient space in the 31-bit XA program pool. Ideally, this value should be zero; a value greater than zero indicates insufficient space in the XA program pool for the volume of program load activity in the system.
(Extended addressing only) Indicates the number of pages used by nonresident modules loaded into the 31-bit XA program pool. The size of a page equals 4K. Only one program can occupy a program pool page; that is, a 4.5K program uses two pages of program pool.
(Extended addressing only) Indicates the number of reentrant modules loaded into the 31-bit XA reentrant pool; reentrant modules include reentrant programs, subschemas, CA ADS dialogs, and DC/UCF maps that have been assigned an RMODE of ANY.
(Extended addressing only) Indicates the number of times program loads were delayed due to insufficient space in the 31-bit XA reentrant pool. Ideally, this value should be zero; a value greater than zero indicates insufficient space in the XA reentrant pool for the volume of program load activity in the system.
(Extended addressing only) Indicates the number of pages used by reentrant modules loaded into the 31-bit XA reentrant pool. The size of a page equals 512 bytes (0.5K).
Indicates the number of requests to release 4K virtual pages.
Indicates the number of 4K byte pages actually released.
(VS systems only) Indicates the number of requests to fix 4K virtual pages allocated to storage pools defined to the system.
(VS systems only) Indicates the number of 4K virtual pages fixed in storage pools defined to the system.
Indicates the number of requests to release 4K fixed virtual pages allocated to storage pools defined to the system, thereby making storage eligible for paging out.
Indicates the number released of 4K fixed virtual pages allocated to system storage pools.
Indicates the number of program requests to write statistics to the journal file with the WRITE JOURNAL DML statement.
Indicates the amount of CPU time (in ten-thousandths seconds) spent in executing user code.
Indicates the number of times user programs requested DC/UCF services (for example, GET STORAGE requests). This value also includes both explicit and implicit requests for database services. For example, OBTAIN record-name is an explicit program request. PUT QUEUE is an implicit request for database services because the DC/UCF system must store the queue record.
Indicates the number of times a user or system program requests database services (for example, OBTAIN record-name).
For LRF and SQL programs, DB SERVICE REQUESTS should be less than or equal to the number of CALLS TO DBMS. You can use these values to evaluate how efficiently the LRF or SQL path extracts data. For example, a program OBTAIN logical-record command increments DB SERVICE REQUESTS by one, but may greatly increment the CALLS TO DBMS value, especially if an area sweep occurs due to the NULL SELECT clause.
Indicates the number of programs called (for example, the number of #LOAD PGM requests issued).
Indicates the highest number of resource link elements (RLEs) used during this session. If this value approximates the threshold established at system generation, increase the threshold.
Indicates the highest number of resource control elements (RCEs) used during this session. If this value approximates the threshold established at system generation, increase the threshold.
Indicates the highest number of deadlock prevention elements (DPEs) used during this session. If this value approximates the threshold established at system generation, increase the threshold.
Indicates the largest amount of the task control element (TCE) stack area used by any task. The stack size is defined at system generation with the STACKSIZE clause of the SYSTEM statement.
Indicates the number of program requests to acquire variable storage dynamically from a DC/UCF storage pool or obtain the address of a previously acquired storage area. The program issues the request with the GET STORAGE DML statement.
Indicates the number of program requests to free all or part of a DC/UCF storage area. The program issues the request with the FREE STORAGE DML statement.
Indicates the amount of CPU time (in ten-thousandths seconds) spent in executing system code.
Indicates the number of program requests for scratch records from the DDLDCSCR area. The program issues the request with the GET SCRATCH DML statement.
Indicates the number of program requests to store or replace a scratch record in the DDLDCSCR area. The program issues the request with the PUT SCRATCH DML statement.
Indicates the number of program requests to delete scratch records from the DDLDCSCR area. The program issues the request with the DELETE SCRATCH DML statement.
Indicates the number of program requests to retrieve a queue record from the DDLDCRUN area and place it in a storage area associated with the issuing program. The program issues the request with the GET QUEUE DML statement.
Indicates the number of program requests to store a queue record in the DDLDCRUN area. The program issues the request with the PUT QUEUE DML statement.
Indicates the number of program requests to delete queue records from the DDLDCRUN area. The program issues the request with the DELETE QUEUE DML statement.
Indicates the number of program requests for the system date and time. The program issues the request with the GET TIME DML statement.
Indicates the number of SETTIME requests to define an event that is to occur after a specified time interval.
Indicates the amount of CPU time (in microseconds) spent in executing system code.
Indicates the amount of CPU time (in microseconds) spent in executing user code.
Indicates the amount of CPU time (in microseconds) spent on zIIP.
Indicates the amount of CPU time (in microseconds) spent on CP, while qualified for zIIP.
Indicates the amount of CPU time (in microseconds) spent on CP in either system mode or user mode.
Indicates the amount of CPU time (in microseconds) spent on CP or zIIP, while qualified for zIIP.
Indicates the number of pages requested by IDMSDBMS (including pages found in a buffer). A page request does not result in a page read if the page is in the buffer pool.
Interpretation: The ratio of PAGES REQUESTED/PAGES READ is the buffer utilization ratio. It measures the effectiveness of the buffer-pool size and design of the database (for example, CALC and VIA clustering). The higher the ratio the better. Ratios consistently below 2.0 indicate that processing is random or that the buffer-pool size is too small.
The buffer utilization ratio may be artificially high for transactions that keep locks, due to the nature of the internal locking mechanism. IDMSDBMS cannot hold a buffer while requesting a lock; therefore, when locks are kept, IDMSDBMS must free and request a page each time a record is requested.
Indicates the number of database pages read from disk.
Indicates the number of database pages physically written to disk. A page can be updated several times before it is actually written back to the database.
Indicates the number of calls to the database management system.
Note: Execution of each navigational DML request involves one call; execution of each logical record facility (LRF) and SQL request typically involves multiple calls.
Indicates the number of database records requested by IDMSDBMS.
Interpretation: The ratio of RECORDS REQUESTED to PAGES READ is the space management ratio. The space management ratio measures how well space is allocated (for example, VIA options, CALC distribution, and buffering). The higher the ratio the better. Ratios less than 4 or less than the norm indicate that the size of the buffer should be increased and database tuning should be performed.
The space management ratio may be artificially high for transactions that keep locks, due to the nature of the internal locking mechanism. IDMSDBMS cannot hold a buffer while requesting a lock; therefore, when locks are kept, IDMSDBMS must free and request a page each time a record is requested.
Indicates the number of records made current of transaction.
Interpretation: The ratio of RECORDS REQUESTED to RECORDS CURRENT is the effectiveness ratio. The effectiveness ratio measures the amount of work CA IDMS/DB is doing for the programmer (that is, how many records the DBMS has to examine to find the one requested). The lower the ratio the better. If the ratio is high, examine set options (for example, sorted order or next pointers only) for appropriateness. If the options are correct, examine the program logic for accurate use of currency.
Indicates the number of all locks acquired and released by all completed transactions. This is NOT a count of locks currently held.
Indicates the number of requested database pages that have been found in a shared cache or an ESA dataspace.
Indicates the number of CALC records stored on the target page.
Indicates the number of CALC records not stored on the target page.
Interpretation: The ratio of CALC records stored on their target page to the total number stored (that is, hits plus overflows) is the CALC cluster ratio. The ratio reflects the efficiency of the CALC algorithm.
The CALC cluster ratio is especially important when the database is loaded or restructured. Ideally, the ratio should be 1, which indicates no overflow. Ratios consistently less than 1 or less than the norm indicate that space utilization is getting high and the database should be tuned.
Indicates the number of VIA and/or DIRECT records stored on the target page.
Indicates the number of VIA and/or DIRECT records not stored on the target page.
Interpretation: The ratio of VIA records stored on their target page to the total number of VIA records stored (that is, hits plus overflows) is the VIA cluster ratio. The ratio reflects how well VIA records cluster around their owner.
Ideally, the ratio should be 1, which indicates no overflow. Ratios less than 1 or less than the norm indicate very large data clusters, high utilization of space, or small page size.
Indicates the number of noncontiguous segments (fragments) stored for variable-length records.
Indicates the number of records relocated from their home page.
Indicates the number of database pages that have been directly found in a prefetch buffer.
Indicates the number of SR8 splits.
Indicates the number of SR8 spawns.
Indicates count of Index members or SR8s whose up-level pointers were corrected to point to the actual SR8 in which they appear.
Indicates number of Btree index probes.
Indicates least number of levels descended.
Indicates highest number of levels descended.
Indicates count of SR8s created.
Indicates count of SR8s erased.
Indicates count of SR7s created.
Indicates count of SR7s erased.
Indicates total number of levels descended.
Indicates the number of SQL commands executed.
Indicates the number of SQL sorts performed.
Indicates the number of rows participating in all sorts.
Indicates the least number of rows sorted.
Indicates the largest number of rows sorted.
Indicates number of tuples FETCHed.
Indicates the number of rows INSERTed.
Indicates the number of rows UPDATed.
Indicates the number of rows DELETEd.
Indicates the number of automatic access module recompilations.
For a detailed explanation of DML commands, see the CA IDMS Navigational DML Programming Guide. For more information about system generation statements, see the CA IDMS System Generation Guide.
|
Copyright © 2013 CA.
All rights reserved.
|
|