IBM Mainframe Forum

Hi,

Request you to help me to understand the space calculation done on single volume or multi-volume datasets?

Is the formula same? Why do we prefer multi-volume over single volume? How does it help?

Thanks.

Space calculation estimates are easy -- take the number of blocks times the block size to get the total bytes used. If the data set is allocated on more than 1 volume, add up the number of blocks on each volume to get total blocks, then multiply.

If a data set requires more space than a single volume allows then you have no choice but use multiple volumes. This is quite common for tapes, and can occur for disk data sets -- a 3390 mod 3, for example, can hold about 2.8 GB of data; if your data set has more than that you have no choice but allocate additional volumes for it.

I'm not so sure anyone "prefers" that a data set be allocated on multiple volumes, but as Mr. Sample says, sometimes you have no choice in the matter. Multiple volume tape data sets are quite common. Multiple volume disk data sets are rather less common, at least by count.

yeah robert sample is correct

The biggest problem with space calculation is that before z/OS 1.7, the largest total space for a single volume and single data set was 65535 tracks. Until volumes larger than the 3390-3 (such as the 3390-9 and larger volumes on 3390 devices emulated on RAID) became available this point was moot since the total space on these volumes was less than 65535 tracks. To allocate "large" datasets (sequential data sets with more than 65535 tracks you specify DSNTYPE=LARGE when you allocate the data set.

This 65535 track limit still exists for traditional PDS data sets, though you can allocate "large" PDSE data sets.