Investigate CAF2417W / COND CODE 0004 messages in SMF Director
search cancel

Investigate CAF2417W / COND CODE 0004 messages in SMF Director

book

Article ID: 43987

calendar_today

Updated On:

Products

JARS JARS Resource Accounting JARS SMF Director

Issue/Introduction

How to identify the cause of a CAF2417W message in SMF Director.

 

Environment

SMF Director 12.7 

Cause

CAF24172 SMF DATA FOR xxxx 05/21/2023 0:00:50 OVERLAPS DATA ALREADY IN INDEX

caused by overlapping records.

Resolution

Here's the easiest way to determine what record(s) are overlapping badly and causing this issue in SMF Director:


1) Run the LISTH report on the system(s) and time ranges where the overlap has occurred. Locate a file that seems out of place in the report. The report is sorted based on the oldest timestamp in the history file listed. Any file that has a wild overlap should be apparent in the report, as most entries in the report would have a start time that is close (or even the same) as the end time of the previous entry. 


2) Run a PRINT or EXTRACT using just the one history file with the overlap by using the SOURCE HISTORY statements. 

As an example, suppose we have a system with three history files with the following start and end time stamps (I’ll use seconds, but SMF Director tracks to the .01 second): 

File Name Start End 
DUMP.FILE01 00:00:00 01:00:00 
DUMP.FILE02 00:30:00 02:00:00 
DUMP.FILE03 02:00:00 03:00:00 

It is pretty obvious that the file with the rogue record is DUMP.FILE02. We know the file should cover the range from 1 AM to 2 AM, but there is a record (or more than one) that’s causing an issue. So what we can do is set up a PRINT or EXTRACT like this: 

// JOB 
//FINDIT EXEC PGM=SMFD,PARM=READ 
//STEPLIB DD DISP=SHR,DSN=CAI.CASFLOAD 
//SCDS DD DISP=SHR,DSN=CAI.SCDS.FILE 
//SYSPRINT DD SYSOUT=* 
//SMFIN DD DISP=SHR,DSN=DUMP.FILE02 
//SYSIN DD * 
SOURCE HISTORY(SMFIN). 
PRINT SID(sid) FROM(date 002500) TO(date 005900). 
/* 
// 

This will use only the history file with the bad entry as input, and will produce a printed report of the records outside the expected range. 

You can use UPDTX to change the start/end times on the history file entry to correct the problem.