How to select the appropriate catalog cache mechanism to improve the performance of access datasets on Z/OS

Catalog is a dataset used to index other datasets on z/OS, and many times the operation of accessing the dataset on the system is catalog, so increasing the performance of each catalog on the system can directly improve the performance of the Access DataSet. Catalog uses caching's structure to cache some catalog records that often have read and update requests, thereby shortening the time to access these catalog and reaching a system-level performance boost. Overall, the Catalog on z/OS has two types of cache: one is to allocate a portion of space in the Catalog address space to store records, called In-storage Catalog   (ISC) cache. Another is the use of z/OS VLF management of the data space to store records, called Catalog data cache (CDSC), two cache methods, because of the characteristics of the structure, suitable for the different functions of the catalog. Below we will introduce the next two kinds of cache mode, at the same time according to catalog function, in the choice of cache to put forward some suggestions.

ISC is stored in the private catalog address space, except for Master catalog (MCAT), each user catalog (UCAT) has a cache of its own fixed size space to store records. MacT has a certain privilege here, it also has its own ISC, but the size of this space is unrestricted. How much data is allocated to the cache of space. General ISC records the data once requested in this catalog, and every time a request is made for the catalog data, it is searched and accessed in ISC. However, the ISC cache mode is somewhat passive when dealing with the catalog update operation. For example, the user updates the records in this catalog, and the catalog management process releases the catalog ISC space corresponding to the dataset, regardless of whether the dataset is cached in ISC, A new ISC space is then allocated to the catalog to cache the data, so for those catalog with newer frequencies, the ISC is being redistributed in a constant release. In other words, this feature of catalog, with ISC is not significantly improved access performance. For small changes in catalog, such as Master Catalog,isc is the best way to cache. Can save a lot of time to visit catalog. In general, ISC is the default cache method. The user can use F catalog,noisc (catname) to close this cache mode, the same can also open isc,f Catalog,isc (catname).

CDSC is a cache that is managed in the VLF data space, defined by the user-defined VLF parmlib (COFVLFXX), which, compared to ISC, replaces every catalog ISC with a whole area, in terms of management records, It is based on each record, so it appears to be more intelligent. It can differentiate each different record in the cache, so that for catalog that have a lot of records that need to be updated regularly, CDSC can just update the different records instead of redistributing a large CDSC space. Direct reduction of the release, redistribution when the resource consumption. In addition, when the cache is full of space, CDSC will automatically delete some of the updated record, so as to provide space for the latest updated records, so this cache mode for those who often update some ucat, is the best choice.

The choice of cache mode is that each lpar is independent of each other, that is, the same catalog can be used in different cache ways on different lpars. Here is an example to illustrate how to set the CDSC cache mode. The ucat we picked were SYS1.QSTR02.ICFCAT.

1 First Use ' F catalog,allocated ' to see this ucat with the cache way

Figure 1-1

We can see at this time the cache mode is I (ISC)

2 set VLF parmlib (COFVLFXX) as shown in Figure 1-2

Figure 1-2

Where Emaj is used to indicate to use the VLF cache ucat name, Maxvirt is the most we introduced above the maximum allowable CDSC cache size, the unit is 4K, which is a total of 256*4=1024kb.

More Wonderful content: http://www.bianceng.cnhttp://www.bianceng.cn/OS/other/