CRU205I   SMF INSERT IS MOST CURRENT BUT IS PRECEDED BY AN SMF INSERT FOR ( t )  catalog entryname /nn NEWER: smftype FROM SYS sysid AT
hh:mm:ss.hh ON mm/dd/yy (yy.ddd)OLDER: smftype FROM SYS sysid AT
hh:mm:ss.hh ON mm/dd/yy (yy.ddd)TIMES DIFFER BY hh:mm:ss [AND
dddd DAYS][NO] SYNCHRONIZATION CHECK
{BASED ON INTERVAL OF ssss
SECONDS}{SINCE ONLY ONE SYSTEM IS
INVOLVED}{SINCE CLOCK DIFFERENCE =
'spec'}NEWER SMF INSERT RECORD IS
WRITTEN TO THE NEW EXPORT
DATA SET

Explanation

The record is for an entry of type t named catalog entryname, extension number nn. For t, A=nonVSAM, B=GDG, C=cluster, E=VSAM extension, G=AIX, J=GDG extension, R=path, T=truename, U=catalog connector, X=alias. nn=00, except for types E and J.

The most current SMF record with the NEWER, highest date/time stamp for this catalog entry is an INSERT resulting from an smftype (DEFINE or ALTER) by the system with identifier sysid at the time and on the date indicated on the NEWER message line.

The SMF record with the OLDER, next-lower date/time stamp for this catalog entry is an INSERT resulting from an smftype (DEFINE or ALTER) by the system with identifier sysid at the time and on the date indicated on the OLDER message line.

The TIMES between these two SMF records DIFFER BY hh hours mm minutes and ss seconds (AND dddd DAYS). Synchronization checking for multiple systems is based on the (nonzero) difference in seconds, ssss, supplied by the user. A SYNCHRONIZATION CHECK occurs when the time difference between the two SMF records is less than ssss. NO SYNCHRONIZATION CHECK means the SMF time difference is larger than ssss, or that synchronization checking was bypassed because ONLY ONE SYSTEM IS INVOLVED or because the CLOCK DIFFERENCE was defaulted, spec = NONE, or was specified as spec = 0000.

A record for deletion of this entry should have appeared between these two records, but none was found for one of the following reasons:
  • Some necessary SMF data was not included in the input.
  • In a multi-system environment, the clocks were not synchronized more closely than the interval between these changes to the catalog, resulting in the incorrect ordering of an intervening delete by a different system. (The clock on the system performing the deletion was behind the clock of the OLDER sysid by more than the interval between the update and the delete or ahead of the clock of the NEWER sysid by more than the interval between the delete and the insert.)
  • Some necessary SMF data was lost.

System action

After taking the action indicated in the last message line, the condition code is set to 8 (if not already higher) and processing continues.

Operator response

Take the following actions:

  1. Review the reports and messages from CRURRSV, Record Selection and Validation, for lost or omitted SMF data. If SMF data was omitted, supply the missing data and re-execute this recovery.
  2. Review the chain of messages for this entry, looking for the missing DELETE. If a DELETE appears near the top of the chain, probably with a synchronization check or another error message, the records are most likely out of sequence and the NEWER INSERT may be assumed to be the most current record. If you are uncertain that this is the case, continue with the next step.
  3. If neither of the above apply, assume that SMF data has been lost. Further assume that one of the missing records is for this catalog entry and save this log for use with the diagnostic information to be gathered after the output data set is imported. When the output of IDCAMS LISTCAT and IDCAMS DIAGNOSE is available, proceed as follows:
    1. If the entry appears in the IDCAMS LISTCAT and if IDCAMS DIAGNOSE does not find it to be in error, confirm that the data set or each component of a VSAM sphere is actually on the volumes indicated by LISTCAT. The IDCAMS DIAGNOSE with the COMPARE option will accomplish this for VSAM entries. For nonVSAM entries, check the VTOC for disk data sets. For data sets on tape, check the tape data set inventory, if a tape management system is in use, or actually check the tape volume. You could also run IDCAMS PRINT IDS(entry.name) COUNT(1) to DD DUMMY for all data set types, if you use standard-label tapes and if you don't have so many tape data sets as to make the number of mounts intolerable.
    2. If the data set and each component is present on the volumes, then assume that this entry is the most current one and no further action is required.
    3. If the entry does not diagnose correctly, or if it mismatches the data volumes, then SMF data has been lost and this entry is not the most current one for one of the following reasons:
      • The data set, sphere or component no longer exists and the catalog entry should be deleted.
      • The data set, sphere or component now exists on different volumes and a correct catalog entry for this data set or VSAM sphere must be built.
    4. Make a note of the volumes on which the data set was last known to reside and then delete the existing catalog entry specifying NOSCRATCH.
    5. If the component is not subsequently found on any volume, then we are finished.
    6. If the data set name tells you that this was a data set that can be easily recreated or is otherwise not essential, allow volume cleanup processing to scratch the data set when it is encountered.
    7. If is necessary to locate a disk data set, examine the VTOCs of all volumes that might contain the data set. For VSAM data sets, IDCAMS DIAGNOSE VVDS will do this. For nonVSAM data sets, use IEHLIST LISTVTOC ...,DSNAME=.... You could also use DFSMSdss with the NORUN option to DUMP ...BY(CATLG,EQ,NO)....
    8. If is necessary to locate a tape data set, look for it in the tape management inventory.
    9. When the data set is found, redefine (DEFINE NONVSAM or DEFINE CLUSTER RECATALOG) the entry.
Parent topic: CRU messages