If you believe the marketing from infrastructure companies these days, the golden age of sysadmins is in the past (term lifted from this RedHat blog post). But I believe this is a dangerous marketing strategy, especially for infrastructure companies.
What’s in a name?
In the olden times, we had sysadmins who managed servers, unless they managed Windows then we called them Windows admins. There were storage admins to manage storage, network admins to manage the network, and backup admins to manage backups. When virtualization became a thing we had vAdmins to manage that. In prehistoric times, we had mainframe administrators.
In short, sysadmins (and every other flavor of admin) managed the infrastructure used by engineers to build software.
Who manages infrastructure today?
Today, we’re seeing a shift in how organizations develop software. Instead of a single collection of code, developers are using containers to separate a workload’s code into discrete sections that can be maintained independently. Because of this, workloads are able to grow and shrink per business requirements. All of this requires a flexible infrastructure.
It doesn’t hurt that most modern developers are used to writing code in the public cloud. Since a public cloud provider maintains the underlying infrastructure (servers, storage, networking, virtualization), developers only need to manage public cloud contract, the virtual infrastructure needed for the workload.
But developers don’t typically manage this infrastructure. They still rely on an ops person (as in devops) to keep the cloud-first infrastructure running in a way that can support the workload.
While this ops person isn’t called a sysadmin any more, they need very similar skills. In fact, it’s really important to bring the old guard along into the brave new world of multi-cloud.
Dismiss the skills of sysadmins at your peril
I still use skills I learned as a sysadmin, even though its been decades since my shadow darkened a server room door. Are you learning how to do IaC for your cloud-native workload? I did that for pre-RHEL Redhat and Solaris systems with kickstart and jumpstart. You’ll never have to create documentation for FDA compliance where you had to sign off on every click made during an install, or probably even really need to use sed or awk.
But if you do, find a sysadmin.
If you were ever a backup admin, you probably understood the quirks of your datacenter and its workloads better than anyone else. That paranoia you learn about losing data is still important; SaaS and PaaS providers generally do not guarantee loss of data. You are responsible for that. Sysadmins know how to find the places that introduces risk to your data. Listen to them.
Are Sysadmins just too old?
Right out of college I worked for EMC in their training organization. They hired me because I could write well, and because I had experience using VI from a web admin internship. One of my first jobs was to teach mainframe SEs how to connect Symmetrix storage arrays to UNIX and Windows hosts.
Talk about sink or swim! I knew there had to be some bridge between mainframes and UNIX, but I was so green (although I made friends quickly with the very kind mainframe trainer). And the SEs knew they had to learn this or be out of a job.
The rumor is that’s why VMware changed VMworld to VM Explore, and why there is no more alumni status. Word on the street is they said vAdmins can adapt or retire. That sounds so short-sighted to me. You can’t do devops without ops, and vAdmins are ops. Why kick them to the curb like that? Even EMC had a training plan to help mainframe admins to transfer their skills to do sysadmin work. Isn’t that what we need at this stage of the digital transformation cycle?
And sysadmins are late career, but this is when we become the elder techies. Is there also some age bias going on here?
Ops = sysadmins. The infrastructure components evolving rapidly, and we probably haven’t even seen the extent of innovation yet that will transform workloads. But one thing is for sure, if all of the sysadmins retire we’ll be in a real crunch.
The skills on-premises sysadmins have are still valid in a multi-cloud world. We shouldn’t toss them to the side in our quest to conquer a new TAM.
I’d love to hear from you – what are you seeing out there when it comes to retooling sysadmin skills?