Using file-system-level suspension for Point in Time Copy

The use of file-system-level suspension through the mmfsctl command prevents incomplete updates in the FlashCopy® image and is the suggested method for protecting the integrity of your FlashCopy images. Issuing the mmfsctl command leaves the on-disk copy of the file system in a fully consistent state, ready to be flashed and copied onto a set of backup disks. The command instructs GPFS to flush the data buffers on all nodes, write the cached metadata structures to disk, and suspend the execution of all subsequent I/O requests.

1. To initiate file-system-level suspension, issue the mmfsctl suspend command.
2. To resume normal file system I/O, issue the mmfsctl resume command.

Assuming a configuration with:

• Storage subsystems – ESS 1; logical subsystem LSS 1
• LUN ids and disk volume names – lunS1 (hdisk11), lunS2 (hdisk12), lunT1, lunT2

  lunS1 and lunS2 are the FlashCopy source volumes. These disks are SAN-connected and appear on the GPFS nodes as hdisk11 and hdisk12, respectively. A single GPFS file system fs0 has been defined over these two disks.

  lunT1 and lunT2 are the FlashCopy target volumes. None of the GPFS nodes have direct connectivity to these disks.

To generate a FlashCopy image using file-system-level suspension:

1. From any node in the GPFS cluster, suspend all file system activity and flush the GPFS buffers on all nodes:

   mmfsctl fs0 suspend

2. Run the establish FlashCopy pair task to create the following volume pairs:

   lunS1 – lunT1 	(source-target)
   lunS2 – lunT2 	(source-target)

3. From any node in the GPFS cluster, resume the file system activity:

   mmfsctl fs0 resume