IBM Mainframe Forum

Hello,
I read from the Manual that the Dataset(PDS) to be added in the Linklist should have Secondary Space extend as 0

Can anyone say why and is that Mandatory that the HLQ of that Dataset should be SYS2 ?

Thanks & Regards,
Bharath K

HLQ can be whatever you want it to be. Secondary extents are not recommended as if you update the library and it goes into secondary extents the module will not be found until a LLA refresh/update is done. Also the LINKLST can only have 255 extents in total so it is advisable not to have secondary extents as it can lead to modules not being found if you go over the 255.

Thanks a lot Nevilh

nevilh wrote:HLQ can be whatever you want it to be. Secondary extents are not recommended as if you update the library and it goes into secondary extents the module will not be found until a LLA refresh/update is done. Also the LINKLST can only have 255 extents in total so it is advisable not to have secondary extents as it can lead to modules not being found if you go over the 255.

Hi Nevilh,

May I know what do you mean by 'The LNKLST' can only have 255 entents in total' ? i don't reach it , would you please help explain it? Thanks advance.

Rita Zhang wrote:... May I know what do you mean by 'The LNKLST' can only have 255 entents in total' ? i don't reach it , would you please help explain it? Thanks advance.

Every data set has at least one extent (not entent). An "extent" is just the data area on disk. A data area called a DEB is built by the system when it opens a data set. Most of the DEB is a map of the space allocated for the data set - or for a concatenated PDS all of the space allocated for the data sets. This space map is used by I/O to ensure a program does not use data areas allocated to other data sets. The space map has a maximum size of 255 extents; the number of entries in the space map is stored in a data area called DEBNMEXT. DEBNMEXT is defined as 1 byte, so the maximum value it can contain is X'FF', or 255. It has been this way for nearly 50 years; I can't imagine how many IBM and user programs would have to be changed if DEBNMEXT is expanded. A user program might use DEBNMEXT if it uses the space map to determine the number of tracks in an opened data set; this is much easier than trying to use the space maps in the Format 1 and Format 3 DSCBs for the data set, especially since OPEN has already done the arithmetic, which is kind of complicated.