Standard image copies
Online Consistent Copy can also use standard image copies, in addition to Instant Snapshots, to make consistent copies.
Standard image copies may be the most efficient way to make consistent copies of a large number of objects, when a relatively small percentage of the data sets involved are very large. (An Instant Snapshot that takes only 3 seconds per data set is great for a 2 GB-data set, but this time is not optimal for 300 40 KB-data sets.)
Use of standard image copies can have the following advantages:
- You are not required to have special hardware or software.
- If you are copying a lot of small data sets, standard image copies may improve performance because Online Consistent Copy can currently do more standard copies in parallel by fully exploiting
subtasking than Instant Snapshots. (However, for large data sets, Instant Snapshots are faster.)
Use of standard image copies can have the following disadvantages:
- The process creates an additional copy of all small data sets.
- The process creates issues with volume selection for new data sets.
To make VSAM consistent copies by using standard image copies, you use the OUTSIZE option and BIGDDN option (described in the
) to automatically make standard copies for small data sets and Instant Snapshots for large data sets.
The OUTSIZE option works with the BIGDDN option to determine when a data set is sufficiently large to merit a different sort of output allocation. One way to use these options is to direct very large data sets to an OUTPUT command descriptor defined with DSSNAP YES. In this way, only those data sets that pass a user-specified size threshold are copied with Instant Snapshot.
A requirement to use standard image copies in Online Consistent Copy is that the job must specify a BIGDDN value that corresponds to an OUTPUT descriptor that includes a data set name (or model) to use for the output VSAM copy.
For example JCL, see Example-6-Using-standard-image-copies-and-Instant-Snapshots.