Saying that the problem of multiple extents applies only to LNKLST data 
sets is, I believe, incorrect. If the data set can be extended after it 
has been opened, whether the data set is in the LNKLST or some other 
concatenation, a fetch of a module that is now in the new extent will fail 
because the directory entry will point into the new extent and that extent 
will not be defined in the DEB.

For something other than the LNKLST, the program could close and re-open 
(or the job could be started and restarted), but it still would have 
encountered the problem. For the LNKLST, if the job encountering the 
problem can be restarted, then creating and activating a new LNKLST set 
will work. If the job was already running and must stay running, then the 
unpredictably dangerous LNKLST UPDATE could be used after creating and 
activating.

IBM Health Checker for z/OS Check(IBMCSV,CSV_LNKLST_SPACE) alerts you when 
you have a data set with more than 1 extent.

FWIW, the limit of extents for the LNKLST is 255. The limit is something 
like 119 for normal concatenations (the number depends on how many data 
sets contributed to the total number of extents). And if you're wondering 
why it's different for the LNKLST, in part it relates to DEBCHK. The DEB 
for things opened by OPEN must fit within 4K. But the LNKLST is not opened 
by OPEN (and would not pass DEBCHK if that were tried). For the LNKLST, 
the limit comes from the number of extents being represented by a 1-byte 
field.

Peter Relson
z/OS Core Technology Design

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

Reply via email to