The steps in this topic guide you through
the process of recovering DEDBs when the RECON data set cannot be
used.
The entire recovery procedure is explained
in three separate topics. The steps in this topic describe the first
task, from the start of the recovery procedure to the successful completion
of an IMS DEDB Fast Recovery job.
The following diagram depicts the flow of the recovery
procedure from the start of the recovery until the successful completion
of an IMS DEDB Fast Recovery job.
Procedure
Determine whether the IMS subsystem
can be started with the /NRE or the /ERE command.
If the /NRE or the /ERE command can be used to start the system,
use either of the commands to restart the IMS subsystem.
In this case, the IMS subsystem
does not need further recovery operations and you can exit your recovery
procedure.
If these commands are not available, continue with Step 2.
Determine the last and the previous OLDS.
Check the console log to determine the ddname of the OLDS
that was in use when IMS failed
and the ddname of the previous OLDS. If an emergency restart has already
closed the OLDS that was in use at the time of failure, determine
the ddname of the OLDS that IMS used
last and the ddname of the previous OLDS.
If the OLDS that was in use at the time of IMS
failure or the previous OLDS contains a log record X'4301' (OLDS entry table log record),
IMS DEDB Fast Recovery uses it to create an OLDS entry
table (DSET) and determines the required OLDS. However, if a DSET is not created, if an SLDS is
required, or if DBRC=N and AUTO=N are specified, you must specify the ddname and data set name of
all OLDSs and SLDSs after IMS recovery start checkpoint
(indicated by the message DFR3301I).
If SLDSs are required
and are not allocated dynamically, a DD statement must be specified
for each SLDS on the job control statement of IMS DEDB Fast Recovery.
The ddname must be eight characters. Data sets must not be concatenated
on DD statements.
To allocate SLDSs dynamically, provide the
SLDS names in the DFRXDYN0 user exit routine. For more information
about the DFRXDYN0 user exit routine, see DFRXDYN0 user exit.
If you are using an XRF complex, ensure that I/O prevention
is complete.
If I/O prevention is not complete, either
wait for the I/O prevention process to end or reset Central Electronic
Complex (CEC).
If you are using a two-phase commit process
environment, prepare Resync control statements.