Hi Folks,
The device type doesn't even matter here, and we shouldn't get
sidetracked by 3380's or whatever. I'll explain.
The fact is that a free Broadcast Dataset record, with key X'FF' has
to have the absolute record number, starting with X'01', in the first
data byte of the record. If a (candidate for a) Broadcast Dataset does
not have this value hard-coded in the first data byte of the X'FF'
records, and the system tries to use that dataset as a Broadcast
Dataset, all mayhem breaks loose. My (commercial) program BDMDSFIX will
solve the situation by resetting the first data bytes of all X'FF'
records to the relative record number on the track. The same result can
be accomplished with my free package from File 247 of the CBT Tape
(Updates page, please), by running BCMDUMP followed by BCMREST, and
after the dump/restore, the resulting Broadcast Dataset will have the
same data as before, but all these hard-coded data bytes (containing the
record position) will be fixed up. ALL THIS HAS NOTHING TO DO WITH
DEVICE TYPES, at least so far!
An easy way of CREATING THE PROBLEM is by copying a Broadcast
Dataset that was formatted on a 3380, to a 3390, which actually holds
FEWER RECORDS PER TRACK, of the unblocked and keyed Broadcast Dataset
records. The resulting (copied) dataset then has the problem, because
the first data bytes of the X'FF' records are out of sync with the
record number of the track that they are on. That's why IBM says you
can't move a Broadcast Dataset between different device types (but I can
do it with my tool packages). The idea is that the device type
situation came in peripherally, and it wasn't the real problem here. So
if it's 3380's or something else (3375's?), it doesn't matter for our
discussion.
Anyway, even though most of the world now seems to have standardized
on DASD images and devices that have 3390 geometry, 3390's are by no
means entirely universal, even nowadays. And I've just tried to make
people aware of the weird situation in the Broadcast Dataset, where the
X'FF' records have to contain hard-coded record positions within their
data. THAT is the actual issue here.
All the best of everything to all of you.
Sincerely, Sam
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
