z/OS DFSMShsm Implementation and Customization Guide
Previous topic | Next topic | Contents | Contact z/OS | Library | PDF


Specifying data set serialization

z/OS DFSMShsm Implementation and Customization Guide
SC23-6869-01

When DFSMShsm is backing up or migrating data sets, it must prevent those data sets from being changed. It does this by serialization. Serialization is the process of controlling access to a resource to protect the integrity of the resource. DFSMShsm serialization is determined by the SETSYS DFHSMDATASETSERIALIZATION | USERDATASETSERIALIZATION command.

Note: In DFSMS/MVS Version 1 Release 5, the incremental backup function has been restructured in order to improve the performance of that function. The SETSYS DFHSMDATASETSERIALIZATION command disables that improvement. Only use the SETSYS DFHSMDATASETSERIALIZATION command if your environment requires it. Otherwise, it is recommended that you use the SETSYS USERDATASETSERIALIZATION command.
If you specify . . . Then . . .
SETSYS DFHSMDATASETSERIALIZATION DFSMShsm issues a RESERVE command that prevents other processing units from accessing the volume while DFSMShsm is copying a data set during volume migration. To prevent system interlock, DFSMShsm releases the reserve on the volume to update the control data sets and the catalog. After the control data sets have been updated, DFSMShsm reads the data set VTOC entry for the data set that was migrated to ensure that no other processing unit has updated the data set while the control data sets were being updated. If the data set has not been updated, it is scratched. If the data set has been updated, DFSMShsm scratches the migration copy of the data set and again updates the control data sets and the catalog to reflect the current location of the data set. Multivolume non-VSAM data sets are not supported by this serialization option because of possible deadlock situations. For more information about volume reserve serialization, see DFHSMDATASETSERIALIZATION.
SETSYS USERDATASETSERIALIZATION Serialization is maintained throughout the complete migration operation, including the scratch of the copy on the user volume. No other processing unit can update the data set while DFSMShsm is performing its operations, and no second read of the data set VTOC entry is required for checking. Also since there is no volume reserved while copying the data set, other data sets on the volume are accessible to users. Therefore, USERDATASETSERIALIZATION provides a performance advantage to DFSMShsm and users in those systems equipped to use it.
You may use SETSYS USERDATASETSERIALIZATION if:
  • The data sets being processed are only accessible to a single z/OS image, even if you are running multiple DFSMShsm hosts in that single z/OS image.

    OR

  • The data sets can be accessed from multiple z/OS images, and a product like GRS must be active and is required in a multiple-image environment.

Go to the previous page Go to the next page




Copyright IBM Corporation 1990, 2014