MVSFORUMS.com

[warp5]

This is an information thread showing some of the hurdles you may meet when changing to larger disks or data sets.
Up to now we have been using 3390 model 3 disk drives. Because we are running out of adresses we are changing to model 9 (3 times the data) and model 27 (9 times the data).
The first hurdle we hit is that now large datasets could allocate over 65535 tracks on a volume. This problem can be bypassed by using the DSNTYPE=LARGE or DSNTYPE=EXTPREF in your JCL. The 65535 tracks is not a problem any more.
The second hurdle (it may be your first) is that we reduce the number of volumes in a storage group, thus there may not be enough volumes for your allocations to be made - you are limited to 16 extents per volume, with lots of volumes no problem, but when you now have a model 27 instead of 9 model 3 disks then the allocation may be too small and you have too many extents (up to 16), but not enough to hold all the data. The solution here is to make larger allocations or to use DSNTYPE=EXTPREF which allows up to 123 extents on a volume. The disadvantage is that some programs will not work with an extended data set, for example ADRDSSU can not dump and restore to or from an extended data set. You can solve this problem by using DSNTYPE=LARGE or DSNTYPE=BASIC (like in old times, with out extras).
And that is the good part, you can override the options at any time using the DSNTYPE=EXTPREF/LARGE/BASIC.
Notice that there are new fields in your data class definitions, I will not explain that here, the help function is very good.
Here is an example: