Failing to plan… is planning to fail!

Yes, it's that simple.

In the last two and a bit years, I have been involved in the migration of close to 400 vm’s onto the vBridge IaaS platform. 99% of these migrations had a detailed plan on how and when the workloads were going to be moved. As a result the number of workloads that needed to be rolled back was about 2%

Understand your environment!

A key success factor is for you to have a detailed understanding your environment, server and application interdependencies as well as networking and security. This is not only for when planning to migrate it somewhere else, but to quickly resolve issues when they arise. This is however not easy in some environments that have a high amount of technical debt or legacy infrastructure, but this is where as-built documentation and the upkeep thereof is crucial folks!

Right, story time…

I remember the first migration I was involved in was 2 months after I started working at vBridge. It was a relatively small number of servers but it had all the right ingredients to go south very quickly.

The IT manager had left the organization one month prior, he had been there for a long time and had all the infrastructure intellectual property locked in his head, and nothing was documented in the form of as-built documentation. The project manager was an external consultant so did not have a great understanding of the environment. We did not have the luxury of doing in-depth discovery either and the technical resource assigned to the project was also in the dark on how the environment was stitched together. Lastly, we also had the risk of failing infrastructure breathing down our necks.

But collectively we did the best we could and came up with a detailed plan on how we were going to undertake the migration, and mitigate as much risk as possible. The time came to start moving the workloads… 1:00 am on a Saturday morning we kicked it off, and I think everyone and sundry was puckering a bit. But guess what, by 1:30 pm Saturday afternoon we were done. Yes there were a few issues with some of the workloads when they came back up, but thanks to good planning we were able to address these and sort them out. Case in point - well planned and executed!

In essence, the result is migrations with poor planning, or a “lets wing it” approach are more likely to end up going very far south, with moments of “oh, is that what that server does - bugger” when something breaks, and that’s a position nobody wants to find themselves in.

Plan + Document + Execute = Success!