Considerations for recovering IMS without using RECON data sets

It is possible that the DBRC RECON data sets required by IMS DEDB Fast Recovery will be unusable, or that DBRC abends because of a DBRC function failure. In that event, you can run IMS DEDB Fast Recovery with DBRC=N and AUTO=Y.

When DBRC=N, AUTO=Y is specified, IMS DEDB Fast Recovery sorts the OLDSs to determine the last- and previously-used OLDSs. IMS DEDB Fast Recovery creates an OLDS entry table (DSET) to determine the log data sets needed to complete recovery.

If neither of the selected OLDSs contain the X'42' or X'4301' log record, IMS DEDB Fast Recovery terminates with a return code of 16. If this happens, rerun IMS DEDB Fast Recovery with AUTO=N.

OLDSs that were in use by the IMS to be recovered are treated as follows:
  • OLDSs to which log records have already been written must be within the sorting group. That is, even the OLDS that cannot be the last OLDS must also be specified on a DD statement or allocated dynamically.
  • OLDSs that were started when IMS failed must contain IMS log records, at least in the first block. OLDS to which no log records have been written are used as follows:
    • If the first block can be read, it is used. (This data set cannot be the last OLDS since an old record is read.)
    • If an end-of-file is recognized, or if DUMMY DD has been specified, the job continues.

OLDSs that were not in use by the IMS to be recovered might be contained in the sorting group, but they will not be misused in the subsequent process.

After running IMS DEDB Fast Recovery and before the IMS cold start, you need to recover the DBRC RECON data sets, even if they are unusable.