When using DFSMSdss as the data mover, the BACKUP COPY TECHNIQUE
management-class attribute that is assigned to the control data sets
informs DFSMShsm whether to direct DFSMSdss to use the concurrent
copy function when backing up those data sets. (The attribute is used
for a similar purpose during automatic or volume backup processing.)
The BACKUP COPY TECHNIQUE attribute is used as follows:
- CONCURRENT PREFERRED (P) - specifies that the concurrent copy
facility is preferred. If a concurrent copy session can be established,
the data set is backed up using the concurrent copy facility. Otherwise,
data set is backed up by DFSMSdss logical dump without concurrent
copy.
- CONCURRENT REQUIRED (R) - specifies that the concurrent copy facility
is required. If a concurrent copy session cannot be established, the
data set will not be backed up.
- STANDARD (S) - specifies that the concurrent copy facility is
not to be used, even though it may be available for the data set.
Normal DFSMSdss logical dump backup processing is used without concurrent
copy.
- VIRTUAL PREFERRED (VP) - specifies that the virtual concurrent
copy is preferred. If a virtual concurrent copy session cannot be
established the data set is backed up by DFSMSdss logical dump without
concurrent copy.
- VIRTUAL REQUIRED (VR) - specifies that the virtual concurrent
copy is required. If a virtual concurrent copy session cannot be established,
the data set will not be backed up.
- CACHE PREFERRED (CP) - specifies that the cache-based concurrent
copy is preferred. If a cache-based concurrent copy session cannot
be established the data set is backed up by DFSMSdss logical dump
without concurrent copy.
- CACHE REQUIRED (CR) - specifies that the cache-based concurrent
copy is required. If a cache-based concurrent session cannot be established,
the data set will not be backed up.
If the management class associated with a data set specifies that
the concurrent copy BACKUP COPY TECHNIQUE be used, DFSMSdss is invoked
specifying the CONCURRENT parameter.
Note: Use these additional considerations when using cache-based
concurrent copy to process large data sets or volumes, or when using
multiple tasks to process a large number of data sets:
- Use concurrent copy only during periods of light update activity
for the data sets or volumes involved.
- Spread multiple concurrent copy jobs across as many LPARs as possible.
- Ensure that jobs and address spaces that will use concurrent copy
are assigned a WLM service class with a high execution velocity. Never
assign a discretionary goal to concurrent copy work.
- Ensure that volumes that will be processed will not be reserved
by another system during the concurrent copy.