New Co:Z Release 8.0.0 is available. Review the Release Notes and Download.
New Co:Z Release 8.0.0 is available.

z/OS Hybrid Batch Processing: Updating a Linux database from z/OS batch

In a typical zEnterprise ecosystem, it is common to have Linux virtual servers that run customer facing web applications. In some cases, these applications use z/OS DB2 for their back end data store, but it is also typical to have this role fulfilled by an Oracle database on the Linux system itself.

Even if the database is local to the Linux system, however, it may need to be updated with data that is produced and/or held on z/OS. Traditionally, the process of applying these updates has involved a file transfer of a dataset containing the updates followed by a trigger that kicks off the utility (typically SQL *Loader) to load the database.

This approach has several drawbacks:

  • Coordination between the file transfer and SQL *Loader can be non-trivial
  • The update data needs to exist in a file on the remote system before SQL *Loader can begin (i.e. unwanted "data at rest"). If it is sensitive, access policies need to be carefully considered.
  • SQL *Loader status (error messages, logs, return codes) has to be checked on the remote system. This can be problematic for operations staff that are z/OS centric.

It would be ideal if this maintenance process could be scheduled and managed from z/OS, with just the database update program being present on the Linux system. In order to make this possible, the following considerations need to be addressed. How will SQL *Loader:

  • access required input and output data, which is on z/OS?
  • be scheduled, possibly so as to fit into the context of a larger job stream?
  • direct status (log messages, monitoring, completion status, etc.) to z/OS so that it can be observed and acted upon?

The Co:Z Co-Processing Toolkit provides z/OS users with tools that mitigate these considerations while making it easy to exploit remote virtual servers (Linux on System z or Linux on zBX) from traditional batch workloads - an idiom we refer to as z/OS hybrid batch processing.

View the full paper: z/OS Hybrid Batch Processing: Updating a Linux database from z/OS batch [PDF]