1304 Supporting On Demand Business Applications with the IBM Problem Determination Tools
NoDup Syntax
Table A-3 Parameter NoDup syntax
There are two distinctly different types of duplicate fault detection, depending on the
execution environment:
Normal duplicate detection is applicable to all execution environments and is controlled by
the NORMAL suboption of the NoDup option.
The purpose of normal duplicate detection is to prevent multiple identical faults from being
recorded separately in a fault history file, thus assisting with DASD space preservation as
well as maintaining a fault history file containing only entries that represent unique
Fast duplicate detection is applicable to CICS only, and is controlled by the FAST
suboption of the NoDup option.
The purpose of fast duplicate detection is to prevent multiple identical CICS transaction
abends, all occurring within a short time frame, from unnecessarily slowing down the CICS
recovery by only permitting one fault analysis of each unique problem within the specified
NoDup parameter syntax
Note: Fault Analyzer analysis resulting from EXEC CICS DUMP calls is not subject to
NoDup(FAST) processing unless an active TACB is available.
Appendix A. ENV - common exit environment information 1305
period. Since the detection of duplicate faults must occur prior to fault analysis, the criteria
does not include any specific information related to the point of failure.
CICS transaction faults that have not been deemed duplicates based on the
fast duplicate
detection rules are subsequently subject to the
normal duplicate detection rules.
Regardless of the method used to designate a fault as a duplicate, the fault entry of which it is
a duplicate will have its duplicate count incremented accordingly.
The following sections explain each of the suboptions in detail.
This suboption is used to specify the number of minutes elapsed since the completion of Fault
Analyzer analysis of the last fault, during which subsequent invocations of Fault Analyzer for
other faults
in the same job step will be deemed duplicates of the last fault if they satisfy the
appropriate fault characteristics criteria.
This option is currently used under CICS to prevent multiple identical transaction abends that
occur within a short period of time from all being analyzed by Fault Analyzer with subsequent
risk of exhausting system resources. In the CICS environment, a fault is considered a
duplicate of another if the faults occurred within the specified elapse time (minutes), the |
transaction IDs are the same, and the following details in each transaction abend control
block (TACB) are identical:
򐂰 CICS abend codes
򐂰 Failing program names
򐂰 Request IDs
򐂰 System and user sense codes
򐂰 Operating system abend codes
򐂰 Offsets to the point of error
򐂰 PSWs on entry to abend
Only the TACB fields that are valid for both TACBs are included in the duplicate comparison.
The default elapse time is 5 minutes.
The valid range of minutes is from 0 to 10080 (10080 is equivalent to one week). A
specification of 0 minutes implies that duplicate fault detection is not to be performed.
When a fault is deemed a duplicate of another based on the fault characteristics and the
NoDup(FAST(minutes)) option in effect, the writing of the history file entry will be suppressed,
the duplicate count associated with the last fault is incremented by one, and message
IDI0066I is issued. This type of fault suppression is referred to as
CICS fast-path duplicate
fault suppression
Changing the NoDup(FAST) option
Typically, an IDIOPTS DD statement is used in CICS procedures to facilitate changes to
options without the need to cycle CICS. However, as far as the NoDup(FAST) option is
concerned, a limitation exists. The actual processing of the IDIOPTS data set is performed by
the main Fault Analyzer module, IDIDA. This is the main program that runs in the MVS
attached subtask. The NoDup(FAST) processing is done in the IDIXCX52/IDIXCX53 exit
code before the subtask is attached. The value used for NoDup(FAST) is whatever the value
was set to on the previous full run of Fault Analyzer in that CICS region (IDIDA attach). When
a CICS fast-path duplicate fault suppression occurs (message IDI0066I), the attach of IDIDA
and the reading of the options data set does
not occur. Another IDIDA attach must happen
before the NoDup(FAST) option change will be reflected down to the exit for NoDup(FAST)
action. This means either waiting for another unique fault (this can be forced by creating a

Get Supporting On Demand Business Applications with the IBM Problem Determination Tools (APA, DT, DT with Advanced Facilities, FA, File Export, FM, WS) now with O’Reilly online learning.

O’Reilly members experience live online training, plus books, videos, and digital content from 200+ publishers.