I'm currently stuck in a situation at work where our entire platform is based on an on-premises setup, but it feels like we've just been adding things randomly without any plan. We have a couple of hundred physical servers spread across a few data centers, but there's no real management in place. We don't have any systematic way to handle deployments—it's basically chaos every time we need to put something new online.
We use some Docker, but it's standalone and not orchestrated with tools like Kubernetes or Swarm. Our virtual management lacks any centralized system, and most of our storage is direct-attached. Server installations happen manually, and monitoring is a hit or miss, usually limited to just checking if a service is running or a port is open.
I've noticed this inefficiency for a while now, and I'm eager to change things for the better, but my ideas often get overlooked due to budget and time constraints. All focus goes toward deploying new features, pushing foundational improvements aside.
If you were starting fresh with a few racks of servers and a tight budget, how would you tackle this? I've considered organizing into areas like Compute (using Proxmox or Kubernetes), Storage (maybe Ceph or Minio), and managed services for core tasks. Are my instincts on the right track? What strategies could I use to convince management to focus on these foundational issues?
3 Answers
You’re right—auditing your current setup is crucial. Check out every service and server, document their specs, and see what's necessary. From there, you can create a plan to minimize physical servers and transition services to virtual machines.
As for open-source software, while it can save costs, consider the support factor. Paying for software with more support might save you headaches in the long run.
Starting from zero might not be practical, but you can definitely create a roadmap for the next few years. Instead of overhauling everything at once, start to implement your vision gradually. As each server needs replacement, you can introduce virtualization step by step.
Keep it simple initially. Set up a VM host and some shared storage, and use that for any new services. You’ll be surprised how quickly you can consolidate things over time.
First off, it sounds like you’re in a bit of a mess with those many physical servers. The first step is to outline how much time and money the current setup is costing you, like tracking errors, outages, and troubleshooting time. Quantify this into numbers that resonate with management.
When presenting your findings, you should also highlight how the proposed solutions could save time and finances in the long run, even if they have an upfront cost—this is what management usually cares about.
Related Questions
Can't Load PhpMyadmin On After Server Update
Redirect www to non-www in Apache Conf
How To Check If Your SSL Cert Is SHA 1
Windows TrackPad Gestures