Thanks. Not much there about mirroring, secondary volumes or telecommunication 
then. The padding I guess is the bytes with that funny value in each suffix, if 
it's not there the block didn't get written correctly. I'd guess length and 
block number are to enable blocks to be validated somewhere as well. 
Interesting how one slowly discovers things which aren't really documented. 
Although I don't have the details to hand (or in my head) just now I spent some 
time a few years back reading all the various dataset types with EXCP to try to 
develop code to validate all kinds of blocks, such as in an extent in the 
middle of a dataset. It turned out to be quite complicated. I worked out most 
of what is in these suffixes but I did wonder about those bytes.

-----Ursprüngliche Nachricht-----
Von: IBM Mainframe Assembler List [mailto:[email protected]] Im 
Auftrag von Bill Fairchild
Gesendet: Donnerstag, 13. September 2012 20:55
An: [email protected]
Betreff: Re: Data spaces vs hiperspaces

According to slide 4 of z/OS SAM extended format and compression (a 12-slide 
slide show at 
http://publib.boulder.ibm.com/infocenter/ieduasst/stgv1r0/index.jsp?topic=/com.ibm.iea.zos/zos/1.0/DFSMS/V1R0-SAM-Extended-Format-100808/player.html),
 the suffix contains the length of the user data, the relative block number, 
and three bytes used to detect control unit padding.

According to z/OS DFSMS Using Data Sets, control unit padding occurs when the 
block is written during one of the following four situations:
when the processor loses electrical power while writing a block.
when an operator issues the CANCEL command.
during a timeout.
during an ABEND when PURGE=QUIESCE was not specified on the active ESTAE macro.
When read back in, if control unit padding is detected then an I/O error is 
signaled.

Bill Fairchild
Programmer
Rocket Software
408 Chamberlain Park Lane * Franklin, TN 37069-2526 * USA
t: +1.617.614.4503 *  e: [email protected] * w: 
www.rocketsoftware.com


-----Original Message-----
From: IBM Mainframe Assembler List [mailto:[email protected]] On 
Behalf Of David Stokes
Sent: Thursday, September 13, 2012 1:10 PM
To: [email protected]
Subject: Re: Data spaces vs hiperspaces

The 32 bytes is a Media Manager thing, which is a level below the access 
method, it's nothing to do with 64 bit or not. AFAICT there's a count for 
record number in a file, I don't see any connection with data transfer since 
there's no time dependency here, also a network protocol can add all kinds of 
metadata, no need to save it in the physical file. Do you have any reference 
for this? Also user block lengths are simulated in extended files. (One of the 
more interesting items I've discovered in this area btw is the Berger/Bruni 
RedPaper on MIDAW). Anyway, there's no point I can see in adding it yourself 
since either MM does it for the datasets it handles or it would be assumed to 
be just part of the user record.

Reply via email to