Nonrecoverable files

For BDAM files that are nonrecoverable (that is, LOG=NO is specified in the FCT entry), CICS® does not lock records that are being updated.

By default, you get BDAM exclusive control, which operates on a physical block, is system wide, but lasts only until the update is complete. If a transaction reads a record for update under BDAM exclusive control, and the transaction subsequently decides not to change the data, it must release the BDAM exclusive control. To do this, issue an EXEC CICS UNLOCK command, which causes CICS to issue a RELEX macro.

If you don’t want BDAM exclusive control, specify SERVREQ=NOEXCTL on the file entry in the FCT.

For nonrecoverable VSAM files accessed in non-RLS mode, VSAM exclusive control locks the control interval during an update. For nonrecoverable VSAM files accessed in RLS mode, SMSVSAM locks the record during the update.

Figure 1 illustrates the extent of locking for nonrecoverable files.
Figure 1. Locking during updates to nonrecoverable files. This figure illustrates two tasks updating the same record or control interval. Task A is given a lock on the record or control interval between the READ UPDATE and WRITE commands. During this period, task B waits.
The diagram shows task A and Task B running concurrently. Task A issues a READ for UPDATE command against a nonrecoverable file, but before it can rewrite the record, task B also issues a READ for UPDATE against the same file for a record in the same control interval. The diagram shows task B in a wait state until task A rewrites the record and releases the lock. Task B is then given a lock for the duration of its update request. The diagram illustrates that locks are held only until the updated records are written back to the file.
Figure 2. Locking (enqueuing on a resource) during updates to recoverable files. This figure illustrates two tasks updating the same record or control interval. Task A is given an exclusive lock on the record until the update is committed (at the end of the UOW). During this period, task B waits.
The diagram shows task A and Task B running concurrently. Task A issues a READ for UPDATE command against a recoverable file, but before it can rewrite the record, task B also issues a READ for UPDATE against the same file for a record in the same control interval. Task B has to wait because of the lock held by task A. Task A rewrites the updated record, but the diagram shows task B continuing to wait until task A reaches the end of its unit of work, when the update is committed and lock released. Task B is then given a lock which lasts until the end of its unit of work.


dfht2c00202.html | Timestamp icon Last updated: Monday, 5 June 2017