I have been asked to find a way to trigger events on the MF from non-mainframe systems. Users on a Linux system would like to be able to start and stop test CICS regions on demand. So I am looking for a way to receive communication from the external Linux systems. My first thought is SNMP, I have ENF SNMP running, I'm not familiar enough with it to know if it can receive inbound SNMP traps and create a log message. If any one has a thought or process currently in us, I would appreciate suggestions.
There are other ways to do this type of triggering. I'll suggest a simple approach using only FTP.
You could create a text file at the Linux environment with JCL statements to execute a batch job. Then open a FTP session via shell prompt to the mainframe. After authentication issue the FTP sub-command:
quote site filetype=JES
Issue the FTP put command above and then send the text file with the JCL statements.
Your job can be a simple IEFBR14 job so that you can intercept its execution with a CA OPS/MVS message rule and trigger the desired actions. You can also execute directly the action in the job itself if you prefer. If you use simple jobs you can interrogate the jobname for example to know what action to perform in the CA OPS/MVS message rule.
You can use one jobname to start the CICS regions and another one to stop the CICS region.
The Linux users would have to have mainframe credentials and authority to submit jobs.
Carlos Mario Silveira Filho
Principal Support Engineer CA Technologies A Broadcom Company
I'll assume that the CICS regions in questions are managed SSM resources in OPS/MVS. If so, the OPS team has recently build a plugin for Zowe CLI which allows you to start/stop managed resources though a scriptable command line interface which drives Rest APIs on the mainframe. Here is the link to the OPS/MVS plugin for Zowe CLI.
Alternatively, if the CICS region is not managed in OPS SSM, you can use Zowe CLI to issue an MVS Console command to stop/start the CICS region directly.
Here is the link to the console command group in Zowe CLI.
Thanks Carlos and Sujay, the FTP is a good option, my preference is to have as few moving parts as possible, and the FTP is doable. But I have also been talking to Don Fertenbaugh and he thinks that OPSMVS has the ability to receive SNMP traps, as long as the USS interface is configured and Event Manager is installed, he is checking on that for me. His other thought is to use Automation Point, but I believe that is a separately purchased product and that would not fly.
Sujay, I assume that the ZOWE plugin would require that we give end users access to specific tasks and that looks like it would take quite a bit of security work. It would be fine for a limited number of support personnel, but these are application programmers and contractors that change regularly. I don't have security 100% set up on OPS.
Zowe OPS/MVS plugin uses the OPS/MVS Rest APIs. So yes, it would be work to lock down what the end users can and cannot access within OPS.
If you're in favor a batch driven approach, FTP is falling out of favor with most customers because of security concerns. z/OSMF (through Zowe CLI) offers a HTTPS based approach to submitting the job: https://zowe.github.io/docs-site/latest/user-guide/cli-usingcli.html#zos-jobs
As an owner of OPS/MVS, I believe you have the rights to CA Automation Point as well. We can examine your requirements further while I am on your site next week.
Automation Point to OPS/MVS is a good way to go. Also, if you have ESP on the mainframe (or a scheduling package that can schedule to servers), it can watch for files put on the server to trigger an ESP event to start/stop the regions or tell OPS/MVS to do it. It could monitor for all files that start with CICS* maybe a file called CICS-START-REGION1 or CICS-STOP-REGION2 - the file name could be interrogated to determine the region name and if it should be started or stopped?