Multiple updates to a recoverable resource by the same distributed UOW

In a non-DPL environment, it is possible for multiple programs within one unit of work (UOW) to update the same recoverable resource.

For example, program1 might update Record1 in a recoverable file, then link to program2 , which could update the same record, Record1 , in the same file. This might not be good programming practice but it does work, because CICS® considers the owner of the resource to be the task, not the program.

However, in a DPL environment, where the programs involved are running in different CICS regions, it is not possible for multiple programs to update the same recoverable resource within the same UOW. Using the same example, program1 updates Record1 in a recoverable file, then links to program2 , which runs under a mirror task in another region. If program2 function-ships a file control request to update Record1 in the same file, the request hangs. It hangs because the mirror task that processes the file control request of program2 cannot get the record lock for Record1 . The lock is owned by the task under which program1 is running. Even though the file control mirror task and the task under which program1 is running are part of the same distributed UOW, CICS does not allow the update. This is because CICS uses the task, not the distributed UOW, as the basis for locking recoverable resources.



dfht1c00193.html | Timestamp icon Last updated: Monday, 5 June 2017