Back in 2002 I wrote a tip in week 22 about how to identify a HSM backup for a dataset and how to restore it. What I didn't write about was how to create your own HSM backup of a dataset.

Most installations creates HSM backups of almost all datasets except DB2 datasets and work datasets, but there may be exceptions to these rules. Some installations don't care about personal datasets and some even deletes personal datasets after a while without asking for permission. Those datasets being backed up by HSM are backed up in the middle of the night, but maybe you need a backup in the middle of the day.

The answer to all these problems are that you can create your own HSM backup, but you can only backup datasets for which you have RACF ALTER access (see tip week 13/2001). A HSM backup is created using command TSO HBACKDS <datasetname>. Very simple, indeed. There is no need to create a new backup if a dataset hasn't been changed since the last backup was taken. You can use option CHANGEDONLY (shortest abbreviation is CH) which will prevent HBACKDS from taking unnessecary backups of unchanged datasets.

If you want to make a backup of all your personal datasets you can use the command TSO HBACKDS ** CH which will make a complete backup of all your personal changed datasets since the last backup. Fortunately HSM creates backups asynchroniously and notifies you when the backup is completed. If you want to wait for the backup process to complete simply add option WAIT. Please be aware that <datasetname> follow the rules for TSO dataset prefixing.