When a problem occurs, you can replace the original sphere with
the updated shadow copy.
Using the shadow copy can reduce the time it takes to perform a forward
recovery because most of the log records have already been applied.
Follow these basic steps to recover from a shadow copy if the sphere is
accessed as an RLS data set:
- Run shadow forward recovery one more time to be sure that the shadow has
all updates applied. If any change accumulation data sets are available,
they are used in addition to the forward recovery log.
- Run IDCAMS SHCDS with the FRSETRR keyword to mark the original sphere
as being under maintenance. This makes it unavailable and also allows the
following unbind operation to succeed.
- Run IDCAMS SHCDS with the FRUNBIND keyword to unbind any retained locks
against the original sphere. This enables SMS/VSAM to preserve the locks ready
for re-binding later to the recovered sphere.
- Run IDCAMS DELETE to delete the original sphere and define an empty sphere
with the same name.
- Run IDCAMS SHCDS with the FRSETRR keyword to mark the new empty sphere
as being under maintenance. This makes it unavailable while the REPRO from
the shadow copy is in progress. This is also necessary to allow the later
re-bind operation to succeed.
- Run IDCAMS REPRO to copy the shadow copy to the empty sphere, hereafter
called the recovered sphere.
- Run IDCAMS BLDINDEX to rebuild the AIXs for the recovered sphere. This
step is only needed if the shadow copy was created with non-reusable AIXs.
It can be skipped if the shadow copy was created with reusable AIXs.
- Run IDCAMS SHCDS with the FRBIND keyword to re-bind to the recovered sphere
all the retained locks that were unbound from the original sphere.
- Run IDCAMS SHCDS with the FRRESETRR keyword after the re-bind to reset
the maintenance flag and enable the recovered sphere for use.
Note: If your shadow copy is not accessed as an RLS data set,
you only need to perform steps 1, 4, 6, and 7.