This message indicates that either some other task is currently updating the requested database file, or has been interrupted while updating a database file. CA IDMS uses a "lockout" approach to insure that the database files are only updated by one copy of the database software. When any piece of database software opens a database file, it sets ON a bit in that file (called the AREA‑IN‑USE lock).
A 0966 abend occurs when the database software looks to set this bit and discovers that it is already turned on.
The CA Dispatch main region (an CA IDMS central version) sets the AREA‑IN‑USE bit when it starts. All online users and batch jobs that run with CA Dispatch running will update the database through this copy of database software. The bit is cleared when the CA Dispatch main region is shut down cleanly.
A local batch job will also set this bit; the CA Dispatch main region must be shut down to allow the local batch job to run. The local batch job will get the 0966 abend if the CA Dispatch main region is still active or if the CA Dispatch main region has not ended cleanly.
Also, the CA Dispatch main region can get a 0966 if a local batch job did not complete.
Note: A 0966 abend is a signal that one of the jobs that updates the database has not ended cleanly.
The first step in correcting the situation is to determine which job did not end cleanly. If it was the CA Dispatch main region, restart that region. The warmstart facility will examine the journals and determine that the database is correct and ready for further updates.
If it is a local batch job, a restore of the affected database files is required before the batch job can be rerun or the CA Dispatch main region can be started.
This abend can occur in the DSEXJNL1 job
This is usually caused by a previous execution of this job abending or being canceled. The 2107 abend can be prevented by periodically checking the online journal status display DCMT D J and looking for abnormal journals. See the "Database Maintenance" chapter in the CA Dispatch Systems Programmer Guide. If an abnormal condition is found, immediately shut the system down with a STOPCADS. Verify that CA Dispatch came down cleanly.
Note: If CA Dispatch did not come down cleanly and DSEXJFIX is run, the result is likely to be a broken chain in the data base.
Run DSEXJFIX, DSEXPFIX, and DSEXBR01 after the system has successfully ended. Be sure your operators are aware that CA Dispatch will occasionally submit a batch job that will require a tape job.
They should not cancel this job or IPL while it is running.
To recover from a 2107 abend, first look at the journal report provided by the DSEXJNL1 job and find the journal in error. Change parameters as follows: change PROG to IDMSXJNL and MEMBER to DSJRNLn (n = 1 to 4, depending on the problem journal). Run DSEXJNL1 as a one time run only. If this run is successful, enter STOPCADS and after the system comes down cleanly, run DSEXBR01 and then restart CA Dispatch after the backup completes.
If the special run of DSEXJNL1 is unsuccessful and it abends, perform the following:
This abend can occur when DSEXJNL1 is submitted and there are no journals to unload to tape. No further action is needed.
|
Copyright © 2011 CA.
All rights reserved.
|
|