Processing IMS System Log Information
The system log data sets are a basic source for statistics about the processing performed by the online system. Individual log record types contain data that can be
in many ways. For example, you can select and format all activity pertaining to a specified
ID or about IMS pools.
Using IMS System Log Utilities
IMS provides several utilities to assist with extracting log records from the system log. These utilities also assist with reducing and merging data that is spread across several log data sets. The sections that follow describe several of these utilities:
"File Select and Formatting Print Utility"
"Fast Path Log Analysis Utility" on page 407
"Log Transaction Analysis Utility" on page 407
"The Statistical Analysis Utility" on page 408
"Knowledge-Based Log Analysis" on page 408
File Select and Formatting Print Utility
You can use the File Select and Formatting Print utility (DFSERA10) if you want to examine message segments or database change activity in detail. This utility prints the contents of log records contained in the OLDS, SLDS, or the CQS log stream. Each log record is presented as one or more segments of 32 bytes. The printed output gives each segment in both character and hexadecimal formats.
You can specify selection criteria for a subset of the records rather than printing all records. You can also specify a starting record number and the number of records to process. You can use an exit routine to customize the selection and formatting of the log records.
Although you can use the File Select and Formatting Print utility to copy entire input logs, you can more conveniently use the Log Archive utility to do so. You use one or more SLDSs as input and specify a user data set as output. Also, you need to specify DBRC=NO in the
statement to prevent DBRC from making entries in the RECON data set about your backup log. Making backup copies of the system log data sets can be useful to obtain an alternative input source for statistics and other monitoring activities that occur in parallel with production use of the system log.
Fast Path Log Analysis Utility
Use the Fast Path Log Analysis utility (DBFULTA0) to prepare statistical reports for Fast Path, based on data recorded on the IMS system log. This utility is an offline utility and produces three data sets, one of which contains the following formatted
Detail Listing of Exception Transactions
Summary of Exception Detail by Transaction Code for IFP Regions
Overall Summary of Transit Times by Transaction Code for IFP Regions
Overall Summary of Resource Usage and Contentions for All Transaction Codes and PSBs
Summary of Region Occupancy for IFP Regions by PST
Summary of VSO Activity
Recapitulation of the Analysis
These reports are useful for system installation, tuning, and troubleshooting. The Fast Path Log Analysis utility is not
to the IMS Monitor or the Log Transaction Analysis utility. For more information about the IMS Monitor, see "IMS Monitor" on page 414.
Log Transaction Analysis Utility
In an IMS DB/DC or DCCTL environment, you can use the Log Transaction Analysis utility (DFSILTA0) to collect information about individual transactions, based on records in the system log. Many events are tabulated in the Log Analysis report that is produced by this utility, including total response time, time on the input queue, processing time, and time on the output queue.
You can select a start time for the report tabulation; analysis begins at the first checkpoint after the start time. To control how much transaction activity is tabulated, you can specify an interval (in minutes) of elapsed time from the start time before the utility ends the tabulation, or you can relate the activity
to a number of IMS checkpoints.
You can retitle a Log Analysis report or change the sequence in which the detailed transaction lines are printed. You can
by transaction code or by any of the fields in the report. You can also suppress printing so that the output is stored on a DASD data set.
Using the Log Transaction Analysis utility, you can create an output data set (in system log format) that is a copy of all or part of the input system logs. By having a copy of the system log, you can monitor system activity without
the use of the OLDS for recovery.
The Statistical Analysis Utility
In an IMS DB/DC or DCCTL environment, you can produce several summary reports using the IMS Statistical Analysis utility (DFSISTS0). You can use these reports to obtain actual transaction loads and response times for the system. The statistics produced are dependent on the input system log data sets. The following set of reports is produced by the IMS Statistical Analysis utility:
Telecommunication line and terminal (distributed traffic over a 24-
Transaction (distributed activity over a 24-hour day)
Messages queued but not sent (listing by destination and by transaction code)
Program-to-program messages (listing by destination and by transaction code)
Knowledge-Based Log Analysis
Knowledge-Based Log Analysis (KBLA) is a collection of IMS utilities that select, format, and analyze log records. KBLA also provides an ISPF interface to create and run the jobs for various log-related utilities, and to access other ISPF applications. KBLA uses an ISPF panel-driven user interface to simplify JCL job creation and to prevent JCL errors. KBLA generates the JCL and control statements necessary to run the supported utilities. This JCL preparation allows you to focus on the output of the utility used rather than on how to code JCL to extract information.
The following KBLA log-formatting routines help you examine and display data from IMS log data sets:
The Knowledge-Based Formatting Print routine (DFSKBLAK) provides a clear, simple description of each event represented by a log record, including the meaning of the various fields and flags. It does not display null fields.
The Knowledge-Based Summary Formatting Print routine (DFSKBLAS) prints and displays the header and description of a log record, but not the record itself. This information is useful when you need a general understanding of the log records associated with a particular resource.
The Knowledge-Based Basic Formatting Print routine (DFSKBLA3) is similar in function to the Record Format and Print Module (DFSERA30), but its output also provides a brief description of the log record identifier (or log record type). This information can be useful if you are unfamiliar with the record types present in an IMS log. The Knowledge-Based Basic Formatting Print routine also interprets the prefix fields for the IMS type X'01' and type X'03' log records.
The following KBLA routines help with log analysis and diagnosis:
DBCTL Transaction Analysis Utility (DFSKDBC0)
The DBCTL Transaction Analysis utility (DFSKDBC0) performs the functions of both the Log Transaction Analysis utility (DFSILTA0) and the Fast Path Log Analysis utility (DBFULTA0). The DBCTL Transaction Analysis utility also
a sort of the data. The DBCTL Transaction Analysis utility uses the X'07' and X'5937' log records or the X'5938' log record to gather statistics.
Deadlock Trace Record Analysis Utility (DFSKTDL0)
The Deadlock Trace Record Analysis utility is used to format and summarize data extracted from X'67FF' pseudoabend records for database-related deadlocks. These records are identified by a four-character requester identification that indicates a pseudoabend condition, and an eight-character element identification that indicates deadlock.
The Deadlock Trace Record Analysis utility processes the logical deadlock events and the individual deadlocks to produce reports that:
Provide details about the deadlock
Show the hierarchy of the
in the deadlocks (relative to the victim in each deadlock)
Summarize the deadlock activity, including deadlocks by:
- Lock type
IMS Records User Data Scrub Utility (DFSKSCR0)
The IMS Records User Data Scrub utility (DFSKSCR0) performs a scan of all of the IMS logs. The utility deletes those record
that might contain sensitive or confidential customer business transaction information, in particular, X'01', X'03', X'50', X'5901', X'5903', X'5950', and X'67'. The deletion does not compromise the integrity or the content of the
IMS system data. This scrub is useful when you must send IMS log data to an outside organization for analysis.
IRLM Lock Trace Analysis Utilities (DFSKLTA0, DFSKLTB0, DFSKLTC0)
IRLM Lock Trace Analysis consists of three programs (DFSKLTA0, DFSKLTB0, and DFSKLTC0) that run
to perform IRLM Lock Trace Analysis. DFSKLTA0 is run first to create the control file of global data management block (DMB)
and their respective database
. DFSKLTB0 is then used to create the lock wait detail and summary records. DFSKLTC0 formats and prints the information and creates the optional output data set.
The IRLM Lock Trace Analysis utilities can be useful in finding database or application issues that impact transaction response times by
frequent and long lock waits.
You can save the report to a data set, which you can then sort using either the Knowledge-Based Log Analysis ISPF tools or an editor's
command. The IRLM Lock Trace Analysis panel provides some pre-defined sort options, such as Sorted by Database Name and Sorted by RBA.
Log Record Processing Rate Analysis Utility (DFSKRSR0)
The Log Record Processing Rate Analysis utility (DFSKRSR0) is used to generate reports that summarize the volume of log data that is being generated by an IMS subsystem. The volume of log data is
in number of records per second and number of bytes per second. The detailed log record processing rate data is broken down by log record type, and by subtype within record type, if
The Log Record Processing Rate Analysis utility can be used to:
Selection criteria for the DFSKRSR0 trace table entries can include:
Log record type
Subtypes within log record types
Time range data
Log sequence number ranges
Number of records to be
Log Summary Utility (DFSKSUM0)
The Log Summary utility (DFSKSUM0) can create a summary of the content of the log data sets, characteristics of the IMS system that produced the log data (when statistical log records are present), and some statistical information related to transactions, programs, and databases. The Log Summary utility also includes the following information:
Input IMS logs used for the utility execution
First and last Log Sequence Number (LSN) in the log
Time stamp (UTC) and local time of the first and last log record
Difference between UTC and local time (HHMM format)
Elapsed time on selected logs
Total number of log records in the log data set
Presence of internal trace record, system restarts, dump log record, and system checkpoint
Number of log records present for each record ID
Statistics related to individual transactions, programs, and databases
Gaps in the log
MSC Link Performance Formatting Utility (DFSKMSC0)
The MSC Link Performance Formatting utility (DFSKMSC0) uses the IMS MSC link trace log records to provide information about link response times, which can be used to help isolate performance problems with MSC links. You can specify either of the following types of analysis:
The Detail Selection routine (DFSKBST0) produces a report that contains the individual response times in
of every send data and receive data for each MSC link that is traced.
The Summary Selection routine (DFSKMSCD) produces a report that contains the average response time in milliseconds of the total number of send data and receive data values for each link trace.
The enhanced log formatting routines produce formatted output only if the IMS input logs contain X'6701' records generated using the
/trACE SET ON LINK
RECON Query of Log Data Set Names Utility (DFSKARC0)
The RECON Query of Log Data Set Names utility (DFSKARC0) analyzes the RECON data sets to find appropriate log data set names. Based on the control statements that you provide, it can determine the following information:
Data set names for OLDS, SLDS, or PRILOG data sets
The names for the primary or secondary log data sets
The volume serial numbers for these data sets
The logs can be selected by any or all of the following parameters:
Starting date and time
Ending date and time
Information from the RECON data sets can be extracted from either of the following places:
The RECON data sets
, when available on the system that executed DFSKARC0
A pre-existing report generated by the DBRC
Statistic Log Record Analysis Utility (DFSKDVS0)
The Statistic Log Record Analysis utility (DFSKDVS0) processes the X'45' log records that are generated at each IMS checkpoint and formats a report that is a verbal description and interpretation for all of the fields contained in the X'45' log record, including its subcodes. This information can be used to look for bottlenecks within the IMS system or to detect trends in internal resource usage that can help to determine if tuning is necessary. You must use the
command to ensure that the X'45' log record contains the information required for this enhanced formatting.
Trace Record Extract Utility (DFSKXTR0)
The need for information for problem diagnostics or accumulation of statistics
trace table logging, which often generates a large number of records. The Trace Record Extract utility (DFSKXTR0) reads an IMS log data set (OLDS or SLDS) or trace data set (DFSTRAxx) to produce a subset of information that meets specific selection criteria.
Unlike DFSERA10, which searches entire log records for specific strings, DFSKXTR0 searches individual trace table entries using selection criteria. Trace table entries are relatively small, 32- or 64-byte entries. Although the data in the entries can vary depending on what is
being traced, the structure is a constant eight words of data.
Search criteria are entered as a paired specification identifying the word, halfword, or byte to be examined, and the associated data value. Optionally, the entire log record containing the matching trace table entries can be extracted as well.
After searching the trace table entries, DFSKTXR0 stores the entries that match selection criteria into new log records. These new log records can subsequently be passed to the DFSERA60 exit routine for formatting.
Selection criteria for the trace table entries can include:
Trace table IDs
Character string searches at the trace entry level
Specific words, halfwords or bytes in the trace entries
Time range data
Entire log records to be matched
Number of records during processing
Mode of SCAN to read the log records without actual processing
For detailed information about the KBLA routines and utilities, see
IMS Version 9: Utilities Reference: System
Using the IMS Performance Analyzer for z/OS
The IMS Performance Analyzer for z/OS, program number 5697-B89, gives you the information you need to increase your IMS system performance. The IMS Performance Analyzer for z/OS provides more than 30 reports that can help optimize IMS system and application performance.
Specifically, the IMS Performance Analyzer for z/OS:
Offers an ISPF CUA-compliant user interface to build, maintain, and submit report
Provides revised, enhanced, and new reports
Supports IMS Versions 7, 8, and 9 from a single load library
Allows the optional use of GDDM for selected graphical reports
Saves selected report data for reporting using PC tools
Requires no dependency on GPAR
For more information about the IMS Performance Analyzer for z/OS, see "IBM IMS Performance Analyzer for z/OS, V3.3" on page 451.
IBM offers a number of other IMS database productivity tools. IMS Tools is a set of database performance enhancements for your IMS environment. These tools can help you automate and speed up your IMS utility operations. They can also assist you in analyzing, managing, recovering, and repairing your IMS databases. To learn more about these tools, see Chapter 26, "IBM IMS Tools," on page 443 or go to the DB2 and IMS Tool Web site at http://www.ibm.com/software/data/db2imstools/.