Creating the integration node PDSE

Create the integration node PDSE, as part of creating an integration node on z/OS®.

About this task

Each integration node requires a PDSE or a PDS. A PDSE is preferable to a PDS because free space is available without the need to compress the data set.

If you have more than one z/OS image, consider how you will use the file system. You can share files in a file system with different members of a sysplex. The file system is mounted on one z/OS image. Requests to the file are routed to the owning system by XCF (Cross-system Coupling Facility) from systems that do not have the image mounted.

Moving an integration node from one image to another is straightforward and the files can be shared. However, there is a performance demand for using files that are shared between images in a file system, because the data flows through the XCF. This performance demand is true for trace and other diagnostic data.

For more information about mounting file systems and allocating space, including performance considerations, see Mounting file systems.

For details of the disk space that is required, see IBM Integration Bus system requirements.


  1. Create the integration node PDSE.
    For example, create the PDSE using option 3.2 on ISPF. The name of the PDSE must be the same as the JCL variable ++COMPONENTDATASET++.
  2. Allocate a data set that has the following characteristics:
    • Eight directory blocks
    • 15 tracks (or 1 cylinder) of 3390 DASD with a record format of fixed blocked 80
    • A suitable block size (for example 27920)
    • A data set type of library.

What to do next

Create the integration node directory.