On Tue, 15 Jan 2013 07:01:44 +0100, [email protected] wrote:

>> As noted, this fails on the empty record problem.  However, for RECFM=U
>> (possibly via DD override) and z/OS-to-z/OS,
>>
>>     BINARY
>>     PUT
>>
>> works.  (Of course, DD override is not possible on GET.  Might need 
>> LABEL=SL.)
>> They seem to have fixed something!  IIRC, attribute override on the DD 
>> statement
>> used not to work; it seemed to look at the DSCB regardless.
>
>Not my recent experience, not even with a BLKSIZE less than 32K and 
>preallocated receiving data set. While the ftp finished, the receiving system 
>told me that it was not an adrdssu dump at all and would not read it.
>
You're right, of course.  On further review, I recognized that my one test case
(not an amaterse file) consisted of only a single block.

>I would also suggest to wrap this into an amaterse file. Have amaterse create 
>a *.trs file, ftp that *.trs file, unterse on the receiving side (back to 
>tape, if you want to), and then restore via adrdssu.
>
And if SPACE is a constraint, I believe the amaterse file can go directly to 
tape
as long as the untersed is not a PDS(E).

I hate MVS!  It too much conflates the solution space with the problem space.

-- gil

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to