DFHSM conversion to CA Disk improvement

Document ID : KB000018199
Last Modified Date : 14/02/2018
Show Technical Document Details


While running the conversion from DFHSM to CA Disk, this can takes a lot of time to be done. It exists a way to introduce parallelism and so reduce that conversion time in a huge way.


In the step SORT2 of HSMGEN1, you can select from the MCDS REPRO sequential file, the records longer than 280 and they are sorted by VOLSER. The output is written in a file 'xxxx..DISK.OUTPUT.SORTM'. This file will contain all the datasets which should be moved from DFHSM to DISKOS with the related tapes.

Now, if for example these records are 3000, in order to do parallel convert jobs , you can break this file xxxx..DISK.OUTPUT.SORTM down to 3 new files , each containing 1000 records. So you create 3 files SORT1 SORT2 and SORT3 that will be the input of 3 parallel convert jobs which will be CONVERT1 CONVERT2 and CONVERT3. All are the copy of HSMGEN1 but divided in 3 jobs each using SORTn created previously.

In this way each job will manage separated data and if one of them would face any error, only this one will be hung up, not the whole conversion.

Using that way will decrease the conversion delay in a very great way.