PDSE.SHARING = YES | NO
This parameter indicates if the zOS PDSE sharing feature is supported by IBM® Connect:Direct®. The keyword, PDSESHARING in the IGDSMSxx member in SYS1.PARMLIB , which defines the level of PDSE sharing across subsystems of a sysplex, has two possible values:
- NORMAL sharing allows users to share a PDSE only at a data set level.
- EXTENDED sharing allows users to share a PDSE at both a data set level and member level.
For more information on PDSE sharing, including requirements, and test scenarios and results for both the normal and extended mode, refer to the , which you can find at IBM Redbook Partitioned Data Set Extended (PDSE) Usage Guide.
Because PDSE sharing allows multiple users to open the same PDSE member for output, some operations may destroy directories and create data integrity problems. The last user to issue the STOW macro, which replaces an entry on the directory, gets their update permanently applied to the member. To ensure that updates are not lost, keep this consideration in mind so that users can take the appropriate steps.
When you initialize IBM Connect:Direct, the IGWLSHR callable service is used to verify that the operating system can support PDSE sharing. If the operating system cannot support this PDSE sharing level, the SITA641W error message
will display and initialization will continue as if the PDSE.SHARING parameter had been specified as NO.
Level of PDSE.SHARING is not supported
|YES||IBM Connect:Direct will
support PDSESHARING as defined by the operating system.
Note: You must specify SHR as the disposition (DISP) keyword so that you can share a particular data set with other jobs. All other dispositions – OLD, NEW, RPL, and MOD – will continue to serialize the PDSE to ensure that multiple IBM Connect:Direct processes cannot share the same PDSE.
|NO||IBM Connect:Direct will
not support PDSE sharing. If multiple processes attempt to open the
same PDSE for output at the same time, the processes terminate with
one of the following error messages:
In addition, the processes are placed in the appropriate queue and retried according to any initialization parameters specified.
Modifiable through MODIFY INITPARMS command: YES