Troubleshooting
Problem
PDSEs in TWSz EQQJBLIB DD show massive growth caused by accumulation of obsolete members flagged for deletion, but not deleted because of connection to TWSz address space. Members are deleted when TWSz CONTROLLER is stopped.
Cause
This problem occurs because of incorrect processing by Computer Associates (CA) product CA-PMO Runtime Performance Optimizer. It can happen under the following circumstances:
1. A BLDL is issued by the TWSz address space specifying NOCONNECT, but the PMO product intercepts the request and changes it to CONNECT.
2. A member is read by the TWSz address space, establishing a connection to the physical location of that member in the dataset.
3. That same member is then edited or deleted by a DIFFERENT address space.
The "old" copy of the member would normally be physically deleted, but in this situation, it can't be deleted because of the connection, so it is flagged "pending deletion". It will be actually removed when the owning address space terminates.
.
However, the TWSZ controller started task frequently remains active for months at a time, thus leading to the accumulation of members pending deletion.
.
The BLDL NOCONNECT option should not be changed.
Environment
Any release of Tivoli Workload Scheduler when PDSE libraries are concatenated into the EQQJBLIB DD and CA-PMO is used to manage those libraries.
Diagnosing The Problem
Large increase in the size of TWSz Tivoli job scheduler PDSE JCL libraries and increased processing times for batch processes that are scheduled from TWSz.
Resolving The Problem
ISV Vendor Fix:
Vendor: Computer Associates
Product: PMO™Runtime Performance Optimizer
Release: 4.3
Fix Identifier: RO31224
TITLE: PDSE - PMO CAUSES EXCESSIVE MEMBER CONNECTIONS
Was this topic helpful?
Document Information
Modified date:
13 September 2019
UID
swg21507091