To protect the integrity of your user data sets, before CA Disk processes any data set, CA Disk verifies whether data sets are in use by other tasks. This is done by the system enqueue service. This service works well in both a single CPU environment and a multiple CPU environment. However, the multi-CPU performs better if you have a cross-system enqueue package installed.
In multiple CPU environments with cross-system enqueue packages, you must ensure that all enqueues are propagated across all CPUs. For auto-restores in particular, enqueues for DMSAUT0, DMSAUT1, DMSAUT2, DMSAUT3, and DMSAUT4 must be propagated accurately. These enqueues ensure that data integrity is maintained.
If you have multiple CPUs but do not have a cross-system enqueue package installed, you must run CA Disk at times when the other CPU does not need access to those data sets, or provide integrity through other means.
Enqueues using the normal system enqueue service are not always propagated across all systems in a JES3 environment. CA Disk solves this exposure by enqueuing on data sets using SVC 99. Enqueues resulting from a SVC 99 request are propagated in a JES3 environment and provide a more reliable means of protecting data set integrity.
Note:JES3 customers must specify a Y for sysparm ENQGDGBY because the CA Disk JES3 enqueue allocation technique automatically causes an enqueue of the GDG base. For details, see the sysparm description for ENQGDGBYn in the Systems Guide.
For any function (such as Archive, Restore, Release and Move) that requires writing to, deleting or a catalog change to a data set, CA Disk will attempt to get an exclusive enqueue on the data set. If any other task has a shared (DISP=SHR) or exclusive (DISP=NEW, OLD or MOD), CA Disk assumes that someone else is using the data set, and CA Disk will bypass its use of the data set.
For any function (such as Backup and Copy) that requires reading a data set, CA Disk will attempt to get a shared enqueue on the data set. If other tasks have shared (DISP=SHR) enqueues, CA Disk will process the data set. If any other task has an exclusive (DISP=NEW, OLD or MOD) enqueue, CA Disk assumes that someone else is updating the data set, and CA Disk will bypass its use of the data set.
CA Disk Backup and Copy enqueuing techniques follow the standard linkage editor enqueue technique for load libraries and the ISPF enqueue technique for most other partitioned data sets.
The PDS Compress function provides an additional facility to protect user data sets in the shared DASD environment. It allows a command parameter (RESERVE) to be specified that causes CA Disk to lock out all users from access of the volume on which the object data set resides. Lockout occurs during processing for each data set. When this occurs, all tasks in all other machines that need to access the volume must wait for the compress to complete. Check with your installation's systems programming department before using this parameter, as it can severely degrade your machine throughput.
The VBACKUP and VRECOVER functions optionally issue a RESERVE and an ENQUEUE on the VTOC. This locks out all users on other machines from accessing the volume and it prevents other users on the same machine from acquiring space on the volume. Because of reduced opportunity for a data integrity exposure, CA Disk simulate mode and Report processing generally do not issue enqueues, and reports on data sets even if they are in use by other tasks.
For more information, see Processing Clusters in Use by Other Tasks in the chapter "RESTORE/RECOVER."
|
Copyright © 2015 CA Technologies.
All rights reserved.
|
|