4. Operation › 4.10 Integration with CA SMF Director › 4.10.4 Checklists › 4.10.4.3 Recovering a Failed CA SMF Director Split Data Set
4.10.4.3 Recovering a Failed CA SMF Director Split Data Set
Should the CA MICS update job fail on the existence of the
split file state code of 'X', overt action will need to be
taken to recover the split data set. You will need to
execute a CA SMF Director EXTRACT job to recover the missing
data. You must include the INDEXDSN and INDEXDUP parameters
to ensure the CA SMF Director primary and duplicate split
index are updated when the job completes. This allows
CA MICS to be restarted without any intervention.
+ - - - - - - - - IMPORTANT CONSIDERATIONS - - - - - - - - +
|CA SMF Director will not create any new split data sets for|
|this orgsysid (SID) until the failed split data set has |
|been recovered. |
| |
|You must have sufficient security authorization to create |
|a data set with the same high level qualifier as the failed|
|split data set. |
+ - - - - - - - - IMPORTANT CONSIDERATIONS - - - - - - - - +
1) Review the MICSLOG of the CA MICS step that failed.
Locate the BAS64026E error message which indicates a 'X'
record has been detected and note the data set name of the
failed split data set. The data set name of the duplicate
split index can be found in the preceding BAS64022I message.
2) Using ISPF, browse the data set identified in the
BAS64022I message. Locate the 'X' record using the data set
name from the BAS64026E message. The file state code 'X' is
located in column 55.
Make note of the ORGSYSID (SID) which begins in column 147.
Also note the abend code which begins in column 152. The
abend code is not needed by the extract job but should be
used to determine the cause of the split data set failure.
This should assist you in taking corrective action to prevent
a recurrence. For example, if you encounter a SB37
(insufficient space) you should modify the TIMESPACE
parameter for the failed split data set accordingly.
3) Immediately following the 'X' record will be a 'D' record
as identified by the 'D' in column 55. The 'D' record is a
diagnostic record which only exists when a 'X' record
is generated.
Make note of the following additional fields from the 'D'
record:
- the name of the CA SMF Director control data set,
referred to as the SCDS, which begins in column 10
- the data set name of the primary split index which
begins in column 57
- the start date of the history file, which is located in
column 118 in mm/dd/yyyy format
- the start time of the history file, which is located in
column 129 in hh:mm:ss format
You will need all these values to complete and execute the
CA SMF Director Extract job.
4) To obtain the required date and time parameters for the
extract job, you will need to run a CA SMF Director LISTH job
that will report on the contents of the SCDS.
See the following example:
. . . JOB CARD . . .
//STEP1 EXEC PGM=SMFD,REGION=0M,PARM='READ,PERMIT'
//STEPLIB DD DISP=SHR,DSN=caiprod.smfdir.r11.cailib
//SCDS DD DISP=SHR,DSN=scds.name.from.d.record
//SYSPRINT DD SYSOUT=*
//SYSXDIAG DD SYSOUT=*
//XDCCDF DD DUMMY
//SYSIN DD *
LISTH SID(sid) FROM(mmddyy).
/*
//
- Modify the SCDS DD to refer to the name of the SCDS found
in the 'D' record in the duplicate split index.
- Modify the LISTH parameter and substitute the Orgsysid
identified in the 'X' record for the lower case sid.
- To reduce the amount of output produced by the LISTH job,
enter the date from the start date field found in the 'D'
record into the FROM parameter of the LISTH job.
Note:
- You only need the STEPLIB if the CA SMF Director load
library is not in linklist. Modify it as needed.
5) Submit the LISTH job.
6) Using the LISTH job output, locate the FROM date and time
entry on the LISTH report that matches the START date and
time from the duplicate split index 'D' record for the
same orgsysid (SID).
You should find an exact match between the START date and
time in the LISTH report and the 'D' record. The TO date
and time columns from the LISTH report will reflect the
highest date and time for the SMF dump process and should
be used to complete the TO parameter of the extract
statement for the CA SMF Director Extract job.
Note: If there are multiple LISTH entries subsequent to
the START date and time from the D record, you will use
the START date and time from the 'D' record and the END
date and time from the last LISTH entry, to identify the
missing date and time range.
7) Below is an example of a CA SMF Director Extract job.
Lower case characters represent fields that require
modification or review.
//. . JOB CARD . . . <==change
//EXTRACT1 EXEC PGM=SMFD,PARM=READ
//STEPLIB DD DSN=hlq.smfdir.r11.CAILIB,DISP=SHR <==change
//SMFOUT DD DSN=dsn.from.BAS64026E.msg, <==change
// DISP=(,CATLG,DELETE),
// UNIT=sysda,STORCLAS=storclas, <==change
// SPACE=(CYL,(nn,nn),RLSE), <==change
// DCB=(RECFM=VB,LRECL=32756,BLKSIZE=32760)
//SCDS DD DISP=SHR,DSN=scds.from.d.record <==change
//DUP DD DISP=SHR,DSN=dsn.from.BAS64022i.msg <==change
//SYSPRINT DD SYSOUT=*
//SYSXDIAG DD SYSOUT=*
//SYSUDUMP DD SYSOUT=*
//SYSIN DD *
EXTRACT
FROM(mmddyy hhmm) TO(mmddyy hhmm) <==change
SID(orgsysid) SELECT(rectype) <==change
INDEXDSN(hlq.smfd.index) INDEXDUP(DUP) <==change
TODD(SMFOUT) .
/*
//
- Modify the SMFOUT DD to refer to the name displayed in the
BAS64026E and/or the 'X' record. You MUST create the
recovered file under the exact same data set name which
originally failed, including the generation number.
- Modify the SCDS DD to refer to the name of the SCDS found
in the 'D' record in the duplicate split index.
- Modify the DUP DD to refer to the name of the duplicate
split index identified in the BAS64022I message.
- Modify the EXTRACT parameter with the following:
- FROM should be the START date and time from the
LISTH output for this ORGSYSID (SID) that matched the
'D' record start date and time.
- TO should be the END date and time obtained from the
LISTH report output analysis as described in the
preceding note in step 6.
- SID(ORGSYSID) should be the orgsysid identified in the
'X' record.
- SELECT(rectype) should specify the record types needed
for this CA MICS component step. See the
sharedprefix.MICS.PARMS(cccSMFD) member for the list of
record types used by this component, ccc equals the three
character CA MICS Component ID, if needed.
- INDEXDSN is the data set name following the name of the
SCDS in the 'D' record of the duplicate split index.
- INDEXDUP should refer to the completed DUP DD statement.
Note:
- You only need the STEPLIB if the CA SMF Director load
library is not in linklist. Modify it as needed.
- The CA SMF Director duplicate split index is ALWAYS
allocated with a disposition of SHR only. This avoids
contention with CA SMF Director.
- the period terminates the extract statement and is
required.
8) Submit the Extract job and ensure it completes
successfully. Upon completion, the duplicate split index
should contain a 'C' index entry for the SMF split data
set that had previously failed.
9) Restart the failed CA MICS update job, be it the DAILY or
the incremental update job.
