z/OS DFSMS Using Data Sets
Previous topic | Next topic | Contents | Contact z/OS | Library | PDF


Increasing Virtual Storage for Index Set Records

z/OS DFSMS Using Data Sets
SC23-6855-00

To retrieve a record from a key-sequenced data set or variable-length RRDS, or store a record using keyed access, VSAM needs to examine the index of that data set. Before your processing program begins to process the data set, it must specify the amount of virtual storage it is providing for VSAM to buffer index records. The minimum is enough space for one I/O buffer for index records, but a serious performance problem would occur if an index record were continually deleted from virtual storage to make room for another, and retrieved again later when it is required. Ample buffer space for index records can improve performance.

You ensure virtual storage for index-set records by specifying enough virtual storage for index I/O buffers when you begin to process a key-sequenced data set or variable-length RRDS. VSAM keeps as many index-set records in virtual storage as possible. Whenever an index record must be retrieved to locate a data record, VSAM makes room for it by deleting the index record that VSAM judges to be least useful under the prevailing circumstances. It is generally the index record that belongs to the lowest index level or that has been used the least. VSAM does not keep more than one sequence set index record per string unless shared resource pools are used.

Go to the previous page Go to the next page




Copyright IBM Corporation 1990, 2014