Using DevOps practices and the Automation API from Control-M to reduce a 4+ month, thousands of person-hour project down to just 3 weeks with minimal support personnel.
A legacy Foreman system had to be upgraded to the latest release to support CARFAX infrastructure. Each server (physical or VM) would need to be touched. Due to the age of the system, and complexity involved, there was no clear cut upgrade path without having to run various scripts to migrate data, including having to have intervention for each server. It was estimated that even with scripting, this process would take over 4 months to complete with thousands of person-hours dedicated to the project. At one point, for one piece of the project, we looked at having nothing but an entire team doing nothing but “clicking a button” that was required as part of the migration efforts.
Automation to the Rescue
The project lead had previously worked with Automation team personnel on other projects and was aware of the capabilities that we had with the Automation API and Jobs as Code. Discussing the project, it was determined that we could utilize the Automation API and Jobs as Code to significantly reduce manual intervention, as well as make sure that CARFAX infrastructure remained stable during the transition and migration period.
Issues such as “How do I only take one server from a service down at a time?” was resolved through the use of Quantitative Resources; “How do I make sure that all steps are followed in sequence?” was resolved through the use of dependencies; “If I fail, how do I stop other servers in the same service from proceeding?” was resolved through conditional logic. Through the web portal we were even able to give System Admin personnel access to a red light/green light dashboard where they could view progress, set TLAs when necessary, and rerun any failed migration efforts.
While this reduced the project significantly in terms of person-hour time, we still had one piece that required us to login to a web page and push a button. With no API or other means available, Automation personnel used a robotic process automation tool to record and process this important step. With that in place, we then used the Automation API to again control this piece of the puzzle so that all parts were visible, manageable and being monitored through a single pane of glass.
The End Result
Using one master JSON file with 168 lines of “code” we were able to dynamically create over 16,000 jobs that controlled the migration process. Using the Automation API we were able to provide “black box” automation to pieces of the puzzle where no clear cut automation methods were available. Through robotic process automation, we managed to eliminate the final hurdle.
The 4+ month, thousands of person-hour project was reduced to 3 weeks. During this 3-week time, Control-M and the Automation API managed most of the migration efforts with systems personnel only having to respond to incidents and failures. Throughout the project, complete visibility and metrics were available thanks to the built-in web dashboard.
All this, thanks the power of the Automation API & DevOps. What will YOU do with the Automation API today?