Huge, because logical view of the volume is still important.
Examples:
1. Some system datasets, like IODF *have to be single extent*.
2. Multiple extents could reduce total number of secondary extents and
maximum space. By secondary I mean the extents allocated when initial
allocation is filled up. Note, there are many way to relieve this problem.
3. IMHO it is still faster and more effective to write big portion of
data to single virtual disk area. I mean caching, VTOC operations, etc.
I believe the difference between the best and the worst case is small,
but noticeable.
--
Radoslaw Skorupka
Lodz, Poland
W dniu 05.09.2024 o 21:08, Michael Oujesky pisze:
Considering that most all DASD today is virtual (RAID) and going SSD,
how much difference does the lack of CONTIG make?
Michael
At 09:10 AM 9/5/2024, Radoslaw Skorupka wrote:
Content-Transfer-Encoding: 8bit
W dniu 05.09.2024 o 16:06, Lennie Bradshaw pisze:
A few questions for the VSAM experts here.
From memory, for a non-vsam data set the primary space request can
be split into up to 5 physical extents. This can be avoided by
specifying CONTIG on the SPACE keyword.
There is no CONTIG equivalent for VSAM data sets.
Can the primary space allocation be split as well?
If so, is there some way to avoid it? (use an empty volume ðY~S).
Does the SMS storage class option Guaranteed_Space affect things?
Yes, there is no CONTIG option for VSAM. Yes, VSAM can be allocated
with more extents.
No, it cannot be avoided.
Yes, it is some pain i.e. for IODF allocation.
--
Radoslaw Skorupka
Lodz, Poland
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN