Rational Development and Test boot process and helpful tips

Rational Development and Test (or RD&T for short), is a virtualized instance of a z/OS mainframe operating system running on a Linux host, on a normal Intel based CPU.  We’re assuming that you might already be familiarized with it with this article.

When starting the RD&T box, you should be using a linux startup script to start it. The one that is described in RD&T configuration guide is an excellent place to start. Part of that script includes an IPL statement, which actually launches the z/OS operating system:

ipl a80 parm 0a82DC

The ipl statement contains three important pieces of information. The a80 is the device address of the sysres1 volume, which is a bootable z/OS volume. This volume is found in the directory with all the other z/OS volumes such as sbsys1, sbprd1, USER00, etc. Next, 0A82DC indicates the (4 digit) device address of the IODF volume which holds IPL configuration files is 0A82 and that the LOADxx member that will be used is LOADDC. The LOADxx file contains the configuration to determine which system configuration files gets read (IEASYSxx). This member then determines which BPXPRMxx member gets started. This sequence diagram below shows you the general boot sequence for z/OS on RD&T starting with the Linux startup script.

If you get into a situation where you cannot boot your RD&T instance, due to an invalid BPXPRMxx member, you can change your startup script to use the CS as the last two digits of the load parameter. This starts the system in a simpler configuration and does a cold start. It also clears the entire JES pool, so for those instances where your JES spool is full (due to careful neglect), this will avoid having to reinitialize your system logs. The most common reason why your system will not start is if you have edited your BPXPRMxx and did not do a syntax check: before you IPL. Once you make your changes to your member, run the following command
setomvs syntaxcheck=(XX)‘ where XX is your two digit load parm (such as DB). You can run this command at the master console, or you can issue it through the convenient linux command oprmsg ‘setomvs syntaxcheck=(XX)’

*Note that nearly all master console commands can be entered using the oprmsg linux command. You just have to run a tail on the console log to see the out put. 

If the syntax is correct, you should see a return such as:
 BPXO039I SETOMVS SYNTAXCHECK COMMAND SUCCESSFUL.

Another helpful tip, is to split your BPXPRMxx into multiple members. Keep the basic configurations such as NETWORK and RESOLVER statements and the bare minimum FILESYSTEM statements in one member, and the ones that you add as you customize your instance in another. Good examples of this are when you are creating additional file systems for Team Concert, or the RDz host daemon, or if you need to run multiple versions of Java (i.e. Java 7) , and as such must mount them to different directories in z/Unix.
Some additional helpful links:

One Response to “Rational Development and Test boot process and helpful tips

  • Nice z/OS starter post for budding Sys Progs Kenny

    Unbelievably timely 🙂