Concurrent copy improves backup-while-open processing by significantly
reducing the chances of the invalidation of a backup-while-open dump
because of updates to the data set. To use concurrent copy, specify
the CONCURRENT keyword when you dump backup-while-open data sets.
The following is a comparison of the various kinds of dumps you can
ask for:
- Normal dump. Use of the data set must be
quiesced. DFSMSdss obtains serialization, dumps the data set, and
then releases the serialization. The data set cannot be used for the
entire time.
- Concurrent copy dump. Use of the data set
must be quiesced. DFSMSdss obtains serialization, performs concurrent
copy initialization, releases serialization, and then dumps the data
set. DFSMSdss completes concurrent copy initialization within a very
short time (compared to the actual time to dump the data set). DFSMSdss
can use the data set as soon as the concurrent copy initialization
is complete.
- Backup-while-open dump. Use of the data
set does not need to be quiesced. DFSMSdss dumps the data set without
obtaining serialization. The data set can remain in use for the entire
time, but update activity can invalidate the dump at any time during
the dump.
- Backup-while-open dump using concurrent copy. Use
of the data set does not need to be quiesced. DFSMSdss does not obtain
serialization. DFSMSdss performs the concurrent copy initialization.
DFSMSdss completes concurrent copy initialization within a very short
time (compared to the actual time to dump the data set). The data
set can remain in use for the entire time. Like the Backup-while-open
dump, update activity during the dump can invalidate the dump. However,
only update activity that occurs before DFSMSdss
performs the concurrent copy initialization can invalidate the dump.
The chances of the update activity invalidating the dump are significantly
reduced because the concurrent copy initialization completes very
quickly.