Tuning CA PMO › Addressing Special Tuning Concerns › Evaluating the Use of LLA and VLF › Handling LNKLST Libraries

Handling LNKLST Libraries

You can use the CA PMO managed list, LLA, or a combination of the two to handle LNKLST BLDL/FIND requests. In general, the CA PMO managed list takes less time and uses less processor storage than does LLA. Specifically, you have the following options:

• You can turn off LLA if you have CA QuickFetch (this product manages program fetch requests). If you do this, CA PMO will handle all directory searches, and CA QuickFetch will provide an alternative to keeping load modules in the VLF data space.
• You can use LLA and allow it to manage your LNKLST library directories. This occurs by default when you start LLA. In this case, LLA will back up the managed list, or it will replace the managed list if you exclude all LNKLST libraries from CA PMO management using the parameter MLSIZE=0 or EXCLLNK=*. LLA will keep load modules in the VLF data space. If you choose this option, LLA must be refreshed after an update to ensure that it has the most current copy of the updated load module.

LLA can be refreshed selectively so that the refresh affects only the in-storage directory entry for the library or member. Selective refreshes are more efficient than updating all of LLA, and they provide a large savings for VLF. Full LLA refreshes cause all load modules in the VLF data space to be purged. Selective LLA refreshes cause considerably fewer modules to be purged. VLF purges only the load modules that belong to the refreshed library; or, for a member refresh, VLF purges only the refreshed load module. After a refresh, the purged load modules can reenter the VLF data space when they are used again.

You can either conduct selective LLA refreshes manually, or you can allow CA PMO to handle them. If you are using LLA to manage any LNKLST library directories, or if you are using VLF to manage load modules, you should allow CA PMO to handle LLA refreshes automatically. (For details, see the CA PMO Product guide, the chapter on installation, the step Set up Automatic LLA Refreshes.)

Even if you use LLA with CA PMO, you can turn off VLF and use CA QuickFetch to manage your load modules. CA QuickFetch selectively purges updated modules from its address space upon fetch requests for the updated modules. As a result, CA QuickFetch can be more efficient than VLF. When an LLA refresh occurs for a library, VLF purges all load modules that belong to that library, rather than purging only the updated modules in that library. This can be very inefficient when you compress libraries, and, as a result of the compression, move only a few members in the library.

Note: If you want to manually initiate selective refreshes of LLA, see the IBM documentation. If you specify the parameter LLA=WARN, CA PMO will issue message PMO521I, and PMO/XSYS will issue message PMO751I to let you know that you need to refresh LLA. Message PMO522I or PMO752I will indicate which library or member was updated.