The Makers of FDR®


FDRMOVE Technical Description

FDRMOVE has two modes of operation: MOVE and FASTMOVE.

MOVE

The MOVE operation uses high-performance I/O to copy selected data sets directly between the source and target disks as soon as the data sets become available.
  • When a MOVE is initiated (either manually or through your z/OS scheduler), it will immediately move all the selected data sets that are inactive, holding an exclusive SYSDSN ENQ on each data set until the move operation on that data set completes. This includes scratching the data set from the original volume and re-cataloging it to the new volser. Once a data set has been moved, it is immediately available for use by users and applications.
  • If a data set is active, MOVE frequently checks the ENQ status, and as soon as the data set becomes inactive (e.g. when a batch job finishes with it, or when a TSO user logs off) it is immediately moved to the nominated target device(s).

  • A MOVE will automatically terminate once all of the selected data sets have been moved. It can also be stopped manually at any time (or automatically with the STOPAFTER= operand), although you can, if you wish, leave a MOVE task running for several hours, or even days. This will allow it to wait for each active data set to become inactive, giving it the opportunity to progressively move all of your selected data sets across to the target volume(s) as soon as they become inactive.
In many installations, a large percentage of data sets become inactive (not in use) over a period of time, and these data sets typically remain inactive for some time (e.g. overnight). MOVE is, therefore, ideal for automating the movement of these types of data sets with little or no impact on day-to-day operations.

Performance Guideline
In a recent test we used MOVE to move 2.5GB of data (spread between 600 data sets) in around 4 minutes.

FASTMOVE

If an application operates on a near 24/7 basis, it may not be able to wait the time required for a MOVE to move the data. In those instances, a FASTMOVE operation may be more appropriate.

FASTMOVE is designed for moving data sets that belong to applications that cannot tolerate more than a few seconds of downtime. FASTMOVE operates in three phases:
  • Phase 1. FASTMOVE uses a special FDRPAS job to non-disruptively move the source volume(s) containing the selected data sets onto temporary “transit station disks” in the target subsystem.
  • Phase 2. Once a source disk has been transferred into the transit station in the target subsystem by FDRPAS, the FASTMOVE phase then starts up, and acts very much like a MOVE in that it moves and re-catalogs all the selected data sets that are inactive, and then waits for active data sets to become inactive before automatically moving and re-cataloging them as soon as they become available. However, unlike MOVE, which uses high-performance I/O to move the data, FASTMOVE uses instant replication technology (e.g, FlashCopy, EMCSNAP or SnapShot Copy) if present within the target subsystem to move the data sets to their final location on the requested target volume(s). This data movement is done at ultra-fast speeds.
  • Phase 3: Once the FASTMOVE phase has finished moving all of the selected data sets from a transit volume, FDRPAS is then used to automatically move the volume back to its original location in the source subsystem.
Performance Guideline
In a recent benchmark test, FASTMOVE was able to move 1Tb of data (in 250 data sets, spread across 16 volumes) in just 1 minute.

return to top     return to overview