As mentioned in a previous section, PROIV uses shared memory on systems to handle record locking.  This section explains how PROIV cleans up shared memory if a PROIV user terminates without exiting via the normal logging off process. This clean up is done by a PROIV utility called iscollect.

Iscollect is started automatically when the first user calls up PROIV.  Every 30 seconds it examines the shared memory tables for obsolete entries.  If a process is killed abnormally (for example, by the operator), and iscollect will detect that there are redundant entries in shared memory and will remove them.

Thus, under normal conditions, PROIV use of shared memory should be totally transparent to users.  Note that although iscollect ensures that records are not left locked, especially if a PROIV user terminates abnormally, it does not perform file rollback, or any other recovery technique.  Therefore, it is possible that abnormal termination of a PROIV process will result in corrupt data files. To check for this do the following:

  • Ensure that all users are logged out of PROIV.

  • Run ischk on all PROIV and application data files.

If all files are reported to be satisfactory, it is still necessary to check the data files for logical data consistency.  If there are doubts about the data consistency, the files should be restored from the previous back up.

Comment on this topic

Topic ID: 720102

 
 
 

Table of Contents

Index

Glossary

-Search-

Back