Hi Folks,
Please download File 247 from the Updates page of www.cbttape.org to
get a fresh load module of my new commercial version of the old
"BRODSCAN" program to analyze your Broadcast Dataset (SYS1.BRODCAST or a
copy thereof). The BDMSCAN member, and the sample JCL member BDMSCAN$
will guide you through the process of analyzing your Broadcast Dataset
(i.e. SYS1.BRODCAST). I give permission for anybody to use this program
anywhere, no strings attached, subject only to the conditions and
disclaimer rules of the CBT Tape collection. See www.cbttape.org for them.
Please have a look at my website, www.brodmstr.com , which talks
about my commercial Broadcast Dataset product. If you are interested in
my commercial product or you want to find out more about it, please
email me at [email protected] . It will be VERY inexpensive.
This new version of the BDMSCAN program will tell you how many
Notices records are active, and it will also reveal an error condition
which can result when you copy a Broadcast Dataset from one device type
to another (something IBM says you shouldn't do, but it might happen
accidentally). The reason for the error condition is as follows:
Empty message records on the Broadcast Dataset (records with Key
X'FF') have to have as their first data byte, the exact record number on
the track (the R of the CCHHR of the record). If even one of the X'FF'
records doesn't have the R of the CCHHR as its first data byte, then the
Broadcast Dataset (despite any good data it may contain) is not usable
by the system. How can this happen? I'll give an example.
Suppose you use FDR, for example, to copy a Broadcast Dataset from a
3380 to a 3390. (I think ADRDSSU will stop you.) On a 3380, there can
be 53 records in a track for the Broadcast Dataset. On a 3390, there
can be only 50 records on a track. All X'FF' records have the record
number for that record, hard coded into the first data byte. Some of
these records might have X'33' thru X'35' in there, if the dataset was
formatted (by ACCOUNT SYNC) on a 3380. This cannot occur for a 3390
(where the record numbers only go up to X'32'). So the X'FF' records of
the copied dataset will have to be wrong, and a serious system error
will result if you try to use the copied Broadcast Dataset as your
SYS1.BRODCAST dataset, or if you try to switch to it using PARMLIB
UPDATE(xx).
BDMSCAN will now detect these Broadcast Dataset errors. You'll now
see the following messages on a bad dataset:
Count of BRODCAST Header Record (Should be 1) 1
Number of BRODCAST "Notices" Directory Records 10
Number of BRODCAST "Notices" Message Records 250
Number of USER "Mail" Directory Records 116
Number of USER "Mail" Message Records 0
Count of Free Search Record (Should be 1) 1
Number of Empty USER "Mail" Record Slots 9003
Number of Bad empty "Mail" Record Slots 8853
>>>>>> (BRODCAST DATASET UNUSABLE! Fix with BDMDSFIX pgm.) <<<<<<
On the other hand, a "corrected" Broadcast Dataset will show the
following messages:
Count of BRODCAST Header Record (Should be 1) 1
Number of BRODCAST "Notices" Directory Records 10
Number of BRODCAST "Notices" Message Records 250
Number of USER "Mail" Directory Records 116
Number of USER "Mail" Message Records 0
Count of Free Search Record (Should be 1) 1
Number of Empty USER "Mail" Record Slots 9003
Number of Bad empty "Mail" Record Slots 0
and no error message will follow.
My BDMDSFIX program (which is part of my commercial product) will
fix the errant Broadcast Dataset directly. However, without my
commercial package, you can indirectly fix the problem, without losing
any data, by running (from CBT Tape File 247) the BCMDUMP and BCMREST
programs. After the dump and restore, the restored dataset will have
all the X'FF' records fixed with the proper data in the first data byte,
and the new dataset will then be usable by the system.
I hope this shows you some valuable information about the Broadcast
Dataset, which most of us take for granted, and maybe this information
will help get you out of a pinch someday. This might happen if you're
switching your DASD to a device of different geometry.
All the best of everything to all of you.
Sincerely, Sam Golob
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
