Date: mm/dd/ccyy ******************************************************************************** Page: 1 * CA Datacom/DB * Time: hh.mm.ss * General Utility * Release: 12 * Copyright © 2009 CA. All rights reserved. * SPnn ******************************************************************************** CONTROL CARD(S) .........1.........2.........3.........4.........5.........6.........7.........8 Œ REPORT AREA=PXX,JOBSTAT=DETAIL,SYSSTAT=SUMMARY FUNCTION=REPORT AREA=PXX JOBSTAT=DETAIL SYSSTAT=SUMMARY Ž PXX STARTED DURING MUF ENABLE JOB=DBDVM0 MUFNAME=DBDVM0 5/06/2009 8.27.09 PXX MUF INFORMATION UPDATE JOB=DBDVM0 SVC=235 MUFNAME=DBDVM01 5/06/2009 8.27.09 PXX END, PXX CLOSED 5/06/2009 8.27.14 PXX RESET BY CLRPXX JOB=DBDVM0 SVC=235 MUFNAME=DBDVM01 5/06/2009 8.32.13 PXX END, PXX NOT CLOSED (MUF STILL ENABLED OR PXX WAS FULL AT EOJ) 0/00/0000 0.00.00
The command exactly as entered.
An analysis of keywords encountered and expected. Any errors found are flagged with a note in the left margin.
Any messages related to syntax processing are presented in this area.
The first message in the example just shown notes that the PXX was started when the MUF was enabled. This message has no SVC number, because the PXX is started before MUF determines the number to be used. It also has the same MUF name as the MUF job name, because the PXX is started before the MUF determines the MUF name to be used. After the MUF has determined this information, the second line occurs, providing the update to include the SVC and MUF name the MUF is using. The PXX END, PXX CLOSED indicates the MUF has closed the PXX during EOJ processing and at this time the PXX was not full. nt. In z/VSE the SVC is never printed, and the MUFNAME= field contains the SUBID=n value being used.
If the PXX is cleared with CLRPXX, the first message reflects PXX RESET BY CLRPXX with the current identity information. The message PXXEND, PXX NOT CLOSED (MUF STILL ENABLED OR PXX WAS FULL AT EOJ) reflects that the PXX has no close record. This occurs if MUF is still enabled or occurs if the MUF did EOJ but the PXX was full and MUF could not write the close record as shown in the following example:
Date: mm/dd/ccyy ******************************************************************************** Page: 2 * CA Datacom/DB * Time: hh.mm.ss * General Utility * Release: 12 * Copyright © 2009 CA. All rights reserved. * SPnn ******************************************************************************** DATA BASE STARTED AFTER COMM CLRPXX JOB=DBDVMUF SVC=209 SUBID=0 1/06/2003 7.10.15
This page can contain information such as the following:
The report continues with statistics for each job that wrote statistics.
Date: mm/dd/ccyy ******************************************************************************** Page: 3 * CA Datacom/DB * Time: hh.mm.ss * General Utility * Release: 12 * Copyright © 2009 CA. All rights reserved. * SPnn ******************************************************************************** DATA BASE JOB STATISTICS CLOSING JOB-81471-DRH402E TASK-001 1/06/2003 16.24.01 Œ Ž ‘ ’ “ ” TABLE URT SQBUF UPD SYN SUB DXC GBL REQUESTS RETRIEVALS ADDS DELETES UPDATES BAS00002 0 0 Y Y N Y N 0 0 0 0 0 ARA00002 0 0 Y Y N Y N 0 0 0 0 0 FIL00002 0 0 Y Y N Y N 0 0 0 0 0 AGR00002 0 0 Y Y N Y N 8 4 0 0 0 FLD00002 0 0 Y Y N Y N 0 0 0 0 0 KEY00002 0 0 Y Y N Y N 0 0 0 0 0 ELM00002 0 0 Y Y N Y N 0 0 0 0 0 ALS00002 0 0 Y Y N Y N 0 0 0 0 0 KWC00002 0 0 Y Y N Y N 0 0 0 0 0 REL00002 0 0 Y Y N Y N 8 2 2 0 0 TXT00002 0 0 Y Y N Y N 125 58 3 1 24 ATZ00002 0 0 Y Y N Y N 0 0 0 0 0 DVW00002 0 0 Y Y N Y N 0 0 0 0 0 JOB00002 0 0 Y Y N Y N 0 0 0 0 0 LIB00002 0 0 Y Y N Y N 0 0 0 0 0 MEM00002 0 0 Y Y N Y N 0 0 0 0 0 MOD00002 0 0 Y Y N Y N 0 0 0 0 0 NOD00002 0 0 Y Y N Y N 0 0 0 0 0 PER00002 0 0 Y Y N Y N 0 0 0 0 0 PGM00002 0 0 Y Y N Y N 0 0 0 0 0 PNL00002 0 0 Y Y N Y N 0 0 0 0 0 RPT00002 0 0 Y Y N Y N 0 0 0 0 0 STP00002 0 0 Y Y N Y N 0 0 0 0 0 SYS00002 0 0 Y Y N Y N 0 0 0 0 0 HSD00002 0 0 Y Y N Y N 4 4 0 0 0 PLN00002 0 0 Y Y N Y N 3 2 0 0 1 STM00002 0 0 Y Y N Y N 22 12 0 0 8 PRT00002 0 0 Y Y N Y N 0 0 0 0 0 DDD00015 0 0 Y N N Y N 616 549 14 14 0 CNO00015 0 0 Y N N Y N 2 0 0 0 0 A3300016 0 0 Y N N Y N 206 101 100 0 0 TTM00017 0 0 Y N N Y N 2 0 1 0 0 TOTALS 996 732 120 15 33
Database Job Statistics
Message beginning the job statistics. It includes the job name and run unit number of the job that closed a User Requirements Table that caused the statistics to be written to the Statistics and Diagnostics Area.
TABLE - The DATACOM-NAME of the table concatenated with the DATACOM-ID of the database containing the table.
URT - The number used by CA Datacom/DB to keep track of the User Requirements Table. For SQL, this number is always zero.
SQBUF - The number of sequential buffers.
UPD - Indicates the value for UPDATE= in the User Requirements Table. UPDATE=YES means the program can update the table and can hold the table's records under exclusive control. UPDATE=NO means that there is no update capability and records being read cannot be held under exclusive control.
SYN - Indicates the value for SYNONYM= in the User Requirements Table. YES indicates that the table name may be duplicated in different databases. NO specifies that there are no duplicate table names.
SUB - Indicates the value for DBFLSUB= in the User Requirements Table. YES indicates the CA Datacom/DB table load subroutine DBFLSUB is being used.
DXC - Indicates the value for AUTODXC= in the User Requirements Table. YES indicates that the GETIT, GETPS, GSETL, GSETP, RDUBR, RDUNE, RDUNK, RDUNR and RDUNX commands automatically drop exclusive control established by the previous command from the same Request Area. NO specifies the previous commands do not automatically drop exclusive control.
GBL - Indicates the User Requirements Table parameter GETBLK= is being used to specify the buffer size used for blocked GETIT and GETPS commands.
Date: mm/dd/ccyy ******************************************************************************** Page: 4 * CA Datacom/DB * Time: hh.mm.ss * General Utility * Release: 12 * Copyright © 2009 CA. All rights reserved. * SPnn ******************************************************************************** DATA BASE JOB STATISTICS CLOSING JOB-81471-DRH402E TASK-001 1/06/2003 16.24.53 Œ Ž TABLE URT SQBUF UPD SYN SUB DXC GBL REQUESTS RETRIEVALS ADDS DELETES UPDATES BAS00002 0 0 Y Y N Y N 0 0 0 0 0 ARA00002 0 0 Y Y N Y N 0 0 0 0 0 FIL00002 0 0 Y Y N Y N 0 0 0 0 0 AGR00002 0 0 Y Y N Y N 8 4 0 0 0 FLD00002 0 0 Y Y N Y N 0 0 0 0 0 KEY00002 0 0 Y Y N Y N 0 0 0 0 0 ELM00002 0 0 Y Y N Y N 0 0 0 0 0 ALS00002 0 0 Y Y N Y N 0 0 0 0 0 KWC00002 0 0 Y Y N Y N 0 0 0 0 0 REL00002 0 0 Y Y N Y N 8 2 2 0 0 TXT00002 0 0 Y Y N Y N 125 58 3 1 24 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
REQUESTS - The number of total requests per table, including all CA Datacom/DB internal requests. This number includes any requests processed during transaction backout. If no requests are recorded for a table, that table is listed with zero counts.
Since internal requests are included in the total, the number of retrieval, add, delete and update requests may not equal the number shown in this column.
RETRIEVALS - The number of retrieval requests recorded against this table. The count includes all requests that return data to the user (not locate commands). The commands included are: GETxx, RDUxx, REDxx, SELFR, SELNR, and SELSM. CA Datacom/DB commands and SQL statements are included in this count.
ADDS - The number of add requests recorded against this table. CA Datacom/DB commands and SQL statements are included in this count.
DELETES - The number of delete requests recorded against this table. CA Datacom/DB commands and SQL statements are included in this count.
UPDATES - The number of update requests recorded against this table. CA Datacom/DB commands and SQL statements are included in this count.
Date: mm/dd/ccyy ******************************************************************************** Page: 5 * CA Datacom/DB * Time: hh.mm.ss * General Utility * Release: 12 * Copyright © 2009 CA. All rights reserved. * SPnn ******************************************************************************** DATA BASE SUMMARY STATISTICS 1/06/2003 16.27.05 REQUESTS 104} SQL REQUESTS 102} Œ DATA MANAGER REQUESTS 1,138} I/O EVENTS - READS 268 I/O EVENTS - WRITES 155 EXCLUSIVE CONTROL REQUESTS 154 EXCLUSIVE CONTROL CONFLICTS 0 K/E BUFFER REORGANIZATION 0 NO IXX BUFFER AVAILABLE 0 NO DXX BUFFER AVAILABLE 0 NO DATA BUFFER AVAILABLE 0 NO EXPAND BUFFER AVAILABLE 0 LAST IXX SEQUENCE NUMBER 512 LAST DXX SEQUENCE NUMBER 670 LAST DATA SEQUENCE NUMBER 969 LAST EXPAND SEQUENCE NUMBER 735 IXX BUFFER USED - 1 16 IXX BUFFER USED - 2 11 IXX BUFFER USED - 3 8 IXX BUFFER USED - 4 8 IXX BUFFER USED - 5+ 469 DXX BUFFER USED - 1 66 DXX BUFFER USED - 2 44 DXX BUFFER USED - 3 29 DXX BUFFER USED - 4 26 DXX BUFFER USED - 5+ 505 DATA BUFFER USED - 1 95 DATA BUFFER USED - 2 74 DATA BUFFER USED - 3 30 DATA BUFFER USED - 4 24 DATA BUFFER USED - 5+ 746 SEQ. READ AHEAD I/O EVENTS 0 LOG AREA MAXIMUM PERCENT FULL 13 SECONDARY E/C CONFLICTS 0 TASK COMPLETION WAITS 0 REQUEST COMPLETION WAITS 0 SPILL COMPLETION WAITS 0
Database Summary Statistics
These fields report the number of requests processed:
Reports the total number of service requests processed. When using the GETIT or GETPS commands in blocked mode, the total request count is the number of calls to Multi-User, not the number of detail requests for the next record.
The number of requests to the Multi-User Facility for SQL work. SQL requests are also included in the previously shown REQUESTS count. The difference in the number of REQUESTS reported and the number of SQL REQUESTS is equal to the number of CA Datacom/DB commands issued to the Multi-User Facility.
This is the number of calls to the low-level CA Datacom/DB processor. Each CA Datacom/DB command issued to the Multi-User Facility generates one data manager request. Each SQL request generates as many data manager requests as necessary to process the SQL request. Data manager requests can also be generated (infrequently) by internal MUF processes such as restart.
Date: mm/dd/ccyy ******************************************************************************** Page: 6 * CA Datacom/DB * Time: hh.mm.ss * General Utility * Release: 12 * Copyright © 2009 CA. All rights reserved. * SPnn ******************************************************************************** DATA BASE SUMMARY STATISTICS 1/06/2003 16.27.05 REQUESTS 104 SQL REQUESTS 102 DATA MANAGER REQUESTS 1,138 I/O EVENTS - READS 268 } I/O EVENTS - WRITES 155 } EXCLUSIVE CONTROL REQUESTS 154 } Ž EXCLUSIVE CONTROL CONFLICTS 0 } K/E BUFFER REORGANIZATION 0 } NO IXX BUFFER AVAILABLE 0 } NO DXX BUFFER AVAILABLE 0 } NO DATA BUFFER AVAILABLE 0 } NO EXPAND BUFFER AVAILABLE 0 } LAST IXX SEQUENCE NUMBER 512 } LAST DXX SEQUENCE NUMBER 670 } LAST DATA SEQUENCE NUMBER 969 } ‘ LAST EXPAND SEQUENCE NUMBER 735 } IXX BUFFER USED - 1 16 } IXX BUFFER USED - 2 11 } IXX BUFFER USED - 3 8 } IXX BUFFER USED - 4 8 } IXX BUFFER USED - 5+ 469 } DXX BUFFER USED - 1 66 } DXX BUFFER USED - 2 44 } DXX BUFFER USED - 3 29 } ’ DXX BUFFER USED - 4 26 } DXX BUFFER USED - 5+ 505 } DATA BUFFER USED - 1 95 } DATA BUFFER USED - 2 74 } DATA BUFFER USED - 3 30 } DATA BUFFER USED - 4 24 } DATA BUFFER USED - 5+ 746 } SEQ. READ AHEAD I/O EVENTS 0 LOG AREA MAXIMUM PERCENT FULL 13 SECONDARY E/C CONFLICTS 0 TASK COMPLETION WAITS 0 REQUEST COMPLETION WAITS 0 SPILL COMPLETION WAITS 0
The I/O EVENTS READS/WRITES fields reflect the number of physical I/O requests to the operating system (execute channel programs--EXCPs), not the number of blocks read and written (in some cases, multiple blocks may be read with a single I/O event, such as during sequential processing). The ratio of I/Os to the number of requests varies greatly depending on application requirements and design.
These two fields report exclusive control activity:
The number of times primary exclusive control was granted for any data record.
The number of times a task requested primary exclusive control of a record that was currently held under primary exclusive control by another task.
Exclusive control conflicts normally occur when multiple online tasks request the same records for update. The number of exclusive control conflicts compared to the number of exclusive control requests provides no clear (good/bad) analysis. A value of 1 that required a task to wait a long time can be considered bad. A value that is a high percentage of exclusive control requests can be satisfactory when the wait time was short and the heavily used application required the control.
K/E BUFFER REORGANIZATION is the number of times a key/element definition had to be removed from memory to allow a different key/element definition to be read.
NO ... BUFFER AVAILABLE is the number of times all of the buffers of the identified type required by a task were in use. The following are the buffer types:
Index Area buffers (high-level index blocks)
Index Area buffers (low-level index blocks)
Data area primary and secondary buffers (data area blocks)
Expand buffers
Note: Because of the dynamic acquisition and freeing of buffer space by CA Datacom/DB, this value cannot be maintained at zero.
The LAST ... is the total number of references to a particular class of buffers.
The ... BUFFER USED counters tell you how much the IXX, DXX, and data buffers allocated in the Multi-User startup options are being used and any added with the FLEXPOOL function. The following is a definition of each type of buffer:
Index Area buffers (high-level index blocks)
Index Area buffers (low-level index blocks)
Data area buffers (data area blocks)
The first time a buffer is used, after an index or data block is read into a block of memory, the first counter is set to 1. A second use of the same block, if it is still in memory, increments the second counter to 2. The third and fourth uses increment the third and fourth counters to 3 and 4, respectively. Fifth and subsequent uses of the block increment the fifth counter to 5+. No counter is incremented if a block is not used, such as after an I/O error, or if a block is read-ahead sequentially but the application terminated the sequential process prior to end-of-file.
Date: mm/dd/ccyy ******************************************************************************** Page: 7 * CA Datacom/DB * Time: hh.mm.ss * General Utility * Release: 12 * Copyright © 2009 CA. All rights reserved. * SPnn ******************************************************************************** DATA BASE SUMMARY STATISTICS 1/06/2003 16.27.05 REQUESTS 104 SQL REQUESTS 102 DATA MANAGER REQUESTS 1,138 I/O EVENTS - READS 268 I/O EVENTS - WRITES 155 EXCLUSIVE CONTROL REQUESTS 154 EXCLUSIVE CONTROL CONFLICTS 0 K/E BUFFER REORGANIZATION 0 NO IXX BUFFER AVAILABLE 0 NO DXX BUFFER AVAILABLE 0 NO DATA BUFFER AVAILABLE 0 NO EXPAND BUFFER AVAILABLE 0 LAST IXX SEQUENCE NUMBER 512 LAST DXX SEQUENCE NUMBER 670 LAST DATA SEQUENCE NUMBER 969 LAST EXPAND SEQUENCE NUMBER 735 IXX BUFFER USED - 1 16 IXX BUFFER USED - 2 11 IXX BUFFER USED - 3 8 IXX BUFFER USED - 4 8 IXX BUFFER USED - 5+ 469 DXX BUFFER USED - 1 66 DXX BUFFER USED - 2 44 DXX BUFFER USED - 3 29 DXX BUFFER USED - 4 26 DXX BUFFER USED - 5+ 505 DATA BUFFER USED - 1 95 DATA BUFFER USED - 2 74 DATA BUFFER USED - 3 30 DATA BUFFER USED - 4 24 DATA BUFFER USED - 5+ 746 SEQ. READ AHEAD I/O EVENTS 0 Œ LOG AREA MAXIMUM PERCENT FULL 13 SECONDARY E/C CONFLICTS 0 Ž TASK COMPLETION WAITS 0 } REQUEST COMPLETION WAITS 0 } SPILL COMPLETION WAITS 0 }
SEQ. READ AHEAD I/O EVENTS is the number of read-ahead I/O events which occurred during the execution of GETIT and GETPS commands.
This number indicates how effective these commands are in batch jobs. The number of blocks per I/O is generally equal to one-half the number of SEQBUFS= provided in the User Requirements Table. The next low-level index block is read ahead (when using the GETIT command).
LOG AREA MAXIMUM PERCENT FULL is the highest percent full the Log Area reached.
SECONDARY E/C CONFLICTS is the number of times there was a conflict of exclusive control at the secondary level caused by jobs with transaction backout specified. Once such jobs update a record, CA Datacom/DB holds the record under secondary exclusive control until a checkpoint is issued. This allows access to the record, if CA Datacom/DB needs to invoke transaction backout.
... COMPLETION WAITS reports the number of times requests had to wait for service. Continues on the next page of the report.
Indicates the number of times one task had to wait on the completion of another task because of transaction backout contention. This happens only when there are secondary exclusive control conflicts. It is therefore related to the previously shown secondary exclusive control statistic and may show high contention for individual rows by applications.
The number of times a request had to wait either to log a request because the Log Area was full and the oldest log block was not spillable (due to a long-running maintenance request) or for the index queue processing to complete prior to closing the index data set.
The number of times one task had to wait on the completion of a Log Area spill before being able to log a record and continue. This statistic could signify a Log Area which is too small.
Date: mm/dd/ccyy ******************************************************************************** Page: 8 * CA Datacom/DB * Time: hh.mm.ss * General Utility * Release: 12 * Copyright © 2009 CA. All rights reserved. * SPnn ******************************************************************************** DATA BASE (CONTINUED) ACCOUNTING COMPLETION WAITS 0 } Œ SECURITY COMPLETION WAITS 0 } INDEX QUEUE PROCESSING 0 } INDEX QUEUE OVERFLOWS 0 } INDEX BLOCK SPLITS - IXX 0 } INDEX BLOCK SPLITS - DXX 0 } Ž INDEX BLOCK DELETES 0 } CBSBFR SPILL COUNT 0 } CBSBFR ENTRIES AVAILABLE 53 } MOST CBS ENTRIES USED 0 } SETS PROCESSED 1 } SETS REQUIRING TEMPORARY INDEX 0 } TEMPORARY INDEX ENTRIES BUILT 0 } CBS MEMORY ALLOCATIONS 0 } WRITEPEND WRITE IO 0 } DATASECURE KEY DELETES 0 } ‘ DATASECURE LOOP DELETES 0 } RE-EXPAND COMPRESSED RECORDS 1 LOGPOOL BUFFER WAITS 0 INDIRECT I/O WAITS 2,551 LOG CONTROL BLOCK WRITE I/O 139 LOG BUFFER FULL WRITE I/O 3,784 LOG COMMAND WRITE I/O 932 LOG OTHER WRITE I/O 1,203 LOGPOOL WRITE I/O 141 LOG 2-PHASE COMMIT WRITE I/0 3 DATA2 BUFFER USED - 1 95 DATA2 BUFFER USED - 2 74 DATA2 BUFFER USED - 3 30 DATA2 BUFFER USED - 4 24 DATA2 BUFFER USED - 5+ 746 REQUEST PENDING - 01 2,220 REQUEST PENDING - 02 331 REQUEST PENDING - 03 27
... COMPLETION WAITS reports the number of times requests had to wait for service.
The number of times that the CA Datacom/DB Accounting Facility tried to add information to the accounting tables but the buffer was full or closed and the task had to wait. If this number is large, review the accounting tables defined and the accounting parameters specified in the Multi-User startup options.
The number of times that security had to wait for system information.
These two fields report index queue processing:
The number of index queue entries processed. Queue processing includes:
This statistic is useful primarily to Technical Support for debugging purposes.
The number of times the index queue overflowed. If an entry does not fit in the queue, it is ignored and attempted later. This statistic is useful primarily to Technical Support for debugging purposes.
INDEX BLOCK SPLITS - reports the number of times a high-level (IXX) or low-level (DXX) index block was split into two blocks and the number of times an index block, both high and low levels, was logically deleted. These statistics are useful primarily to Technical Support for debugging purposes.
The following fields report Compound Boolean Selection processing:
The number of times the Compound Boolean Selection buffer became full, causing CA Datacom/DB to store active set definitions as index entries with a special key ID. The buffer in which active set definitions are stored is defined by the Multi-User startup option CBS.
The number of 185-byte entries available in the Compound Boolean Selection buffer (defined by the Multi-User startup option CBS value) for storing active set definitions. A set definition typically requires one to three entries.
The maximum number of Compound Boolean Selection buffer entries used to store active set definitions. If this number is equal to the number of available entries, the previously shown CBSBFR SPILL COUNT value indicates how often a set definition index entry was accessed.
The total number of sets processed.
The number of sets processed that built a temporary index.
The number of temporary index entries built to satisfy SELFR command processing.
The number of times that the Compound Boolean Selection Facility had to dynamically allocate the RWTSA due to the complexity of the requests and the number of keys being used.
WRITEPEND WRITE IO is the number of logging (physical) writes (controlled by the Multi-User startup option LOGPEND) that occurred during this cycle.
DATASECURE ... these fields are reserved for future use.
Date: mm/dd/ccyy ******************************************************************************** Page: 9 * CA Datacom/DB * Time: hh.mm.ss * General Utility * Release: 12 * Copyright © 2009 CA. All rights reserved. * SPnn ******************************************************************************** DATA BASE (CONTINUED) ACCOUNTING COMPLETION WAITS 0 SECURITY COMPLETION WAITS 0 INDEX QUEUE PROCESSING 0 INDEX QUEUE OVERFLOWS 0 INDEX BLOCK SPLITS - IXX 0 INDEX BLOCK SPLITS - DXX 0 INDEX BLOCK DELETES 0 CBSBFR SPILL COUNT 0 CBSBFR ENTRIES AVAILABLE 53 MOST CBS ENTRIES USED 0 SETS PROCESSED 1 SETS REQUIRING TEMPORARY INDEX 0 TEMPORARY INDEX ENTRIES BUILT 0 CBS MEMORY ALLOCATIONS 0 WRITEPEND WRITE IO 0 DATASECURE KEY DELETES 0 DATASECURE LOOP DELETES 0 RE-EXPAND COMPRESSED RECORDS 1 } LOGPOOL BUFFER WAITS 0 } INDIRECT I/O WAITS 2,551 } LOG CONTROL BLOCK WRITE I/O 139 } LOG BUFFER FULL WRITE I/O 3,784 } LOG COMMAND WRITE I/O 932 } LOG OTHER WRITE I/O 1,203 } LOGPOOL WRITE I/O 141 } ’ LOG 2-PHASE COMMIT WRITE I/0 3 } DATA2 BUFFER USED - 1 95 } DATA2 BUFFER USED - 2 74 } DATA2 BUFFER USED - 3 30 } DATA2 BUFFER USED - 4 24 } DATA2 BUFFER USED - 5+ 746 } } REQUEST PENDING - 01 2,220 } REQUEST PENDING - 02 331 } REQUEST PENDING - 03 27 }
These fields report the following:
As compressed records are read, they are expanded into an expand buffer. If a record is updated or deleted, the same buffer is found or the record is re-expanded. Priority of expand buffers is to preserve records under exclusive control. This parameter is provided to allow tuning of the number of expand buffers.
This is the count of times that transaction backout was running, attempting to read a log block, but no LOGPOOL buffer was available. The number of log buffers for transaction backout is provided during Multi-User Facility startup. If multiple transaction backouts occur concurrently, contention can occur if fewer LOGPOOL buffers exist than tasks. Because multiple transaction backouts only rarely occur together, trying to keep this stat at zero would, in general, misuse memory, but be aware that when it does occur transaction backout is slowed.
An indirect I/O wait is the condition where a task needs a block which is being read/written. The current task did not start the I/O with wait, but must wait on its completion. The current task could have started the I/O without wait such as reading ahead. These counts are also included in REQUEST PENDING.
This is the number of times the Log control block was written. The control block information is required for RESTART. This counter is part of the information to provide details on the LXX I/O.
This is the number of times that a log buffer was written because it was full.
This is the number of times the Log buffer containing logged data was written because of a command. Commands which cause transaction boundaries (COMIT, LOGCP, LOGCR, ROLBK, CLOSE) are always written. Maintenance commands ADDIT, DELET, and UPDAT are only written if the table definition indicated writes, the URT indicates writes, or the Multi-User startup option LOGPEND indicated no pipelining. Commands which are optional are LOGIT and LOGDW. Some internal system commands may also write the log block prior to full.
Comparing the ratio of COMMAND WRITE I/O to BUFFER FULL WRITE I/O can be used to attempt tuning by reducing the number of LXX writes.
Indicates the times a non-full log block had to be written to support RESTART when an index or data block had to be written.
This represents the number of write I/O to the LOGPOOL buffers used by transaction backout.
This represents the number of write I/O to the LOGPN buffers used by two-phase commit to manage system failure between a coordinating site and one or more remote sites.
These counters tell you how much the secondary data area buffers allocated in the Multi-User startup options are being used and any added with the FLEXPOOL function. See the IXX, DXX, and DATA BUFFER USED fields for an explanation of how these counters are used.
The number of non-I/O waits that occurred during this cycle. The number following the dash indicates the number of tasks that had to wait. A non-I/O wait count means:
Date: mm/dd/ccyy ******************************************************************************** Page: 10 * CA Datacom/DB * Time: hh.mm.ss * General Utility * Release: 12 * Copyright © 2009 CA. All rights reserved. * SPnn ******************************************************************************** DATA BASE SUMMARY STATISTICS 1/06/2003 16.27.05 Œ Ž TABLE/AREA REQUESTS I/O EVENTS - READS I/O EVENTS - WRITES LOGICAL READS LOGICAL WRITES CXX 0 88 36 753 129 LXX 0 3 37 0 201 AGR00002 8 3 0 6 0 ALS00002 0 1 0 2 0 ARA00002 0 1 0 2 0 ATZ00002 0 1 0 2 0 BAS00002 0 1 0 2 0 DVW00002 0 1 0 2 0 ELM00002 0 1 0 2 0 FIL00002 0 1 0 2 0 FLD00002 0 1 0 2 0 HSD00002 4 5 0 6 0 IXX00002 0 41 24 304 53 JOB00002 0 1 0 2 0 KEY00002 0 1 0 2 0 KWC00002 0 1 0 2 0 . . . . . . . . . . . . . . . . . . STP00002 0 1 0 2 0 SYS00002 0 1 0 2 0 TXT00002 125 11 9 86 36 IXX00006 0 2 1 3 1 CNO00015 2 0 0 0 0 DDD00015 616 31 11 579 37 IXX00015 0 30 20 247 103 SIT00015 0 1 0 2 0 A3300016 206 0 0 0 0 IXX00016 0 5 3 112 102 SQ100016 0 10 3 211 102 IXX00017 0 4 1 8 1 TTM00017 2 1 0 2 0 TOTAL 996 268 151 2,396 781
These columns provide a breakdown of the totals given for REQUESTS, I/O EVENTS - READS, and I/O EVENTS - WRITES printed at the beginning of the report. Use this information in conjunction with the Accounting Facility to pinpoint inefficiencies by determining what kind of requests are occurring.
You can use the ratio of LOGICAL READS to I/O EVENTS - READS and LOGICAL WRITES to I/O EVENTS - WRITES as an indicator of the efficiency of your buffer size. The greater the number of logical reads or writes to I/O events, the better your buffer use is and the more effective is your pipeline.
REQUESTS - The number of requests per table which have occurred and been recorded in the Statistics and Diagnostics Area.
I/O EVENTS - READS - The number of I/O reads which have occurred and been recorded in the Statistics and Diagnostics Area.
I/O EVENTS - WRITES - The number of I/O writes which have occurred and been recorded in the Statistics and Diagnostics Area.
LOGICAL READS - The number of internal read requests which have occurred and been recorded in the Statistics and Diagnostics Area.
LOGICAL WRITES - The number of internal write requests which have occurred and been recorded in the Statistics and Diagnostics Area.
Date: mm/dd/ccyy ******************************************************************************** Page: 11 * CA Datacom/DB * Time: hh.mm.ss * General Utility * Release: 12 * Copyright © 2009 CA. All rights reserved. * SPnn ******************************************************************************** DATA BASE SUMMARY STATISTICS 1/06/2003 16.27.05 TABLE RETRIEVALS ADDS DELETES UPDATES AGR00002 4 0 0 0 HSD00002 4 0 0 0 PLN00002 2 0 0 1 REL00002 2 2 0 0 STM00002 12 0 0 8 TXT00002 58 3 1 24 CNO00015 0 0 0 0 DDD00015 549 14 14 0 A3300016 101 100 0 0 TTM00017 0 1 0 0 TOTALS 732 120 15 33
Database Summary Requests Totals
The column headings refer to the following:
The total number of retrieval requests since Multi-User enable. This count includes all requests that return data to the user. If no requests are recorded for a table, that table is not listed.
The total number of add requests since Multi-User enable.
The total number of delete requests since Multi-User enable.
The total number of update requests since Multi-User enable.
|
Copyright © 2015 CA Technologies.
All rights reserved.
|
|