Previously I ran a server rack on a 1000/1000mbps symmetrical service to my home (Ethernet provisioned from an ISP) to provide a number of personal media services to myself, friends and family.
Having had to relocate across 3 properties in the last 12 months however, drastically changed things – and I’ve since fallen back to a single 4RU server that’s packed to the gills with storage, and uses generic off-the-shelf consumer parts to keep going. There’s been a lot of benefits to this actually.
- It’s simple to get replacement parts
- Noise is kept to a minimum
- Reliability and up-time is a little less than enterprise hardware, but otherwise acceptably usable
With my recent trip to Japan and return – I’ve found the need to actually back up all my photos and media into a singular place that’s much easier to manage. So now I just run a single monolithic server called JT-SERVER which provides all the services I need.
Services I’ve built up to use nowadays are:
- Plex Media Server
- Nginx Proxy Manager – which routes to:
That’s it – nowadays there’s very little I do in the way of other things. A game server is still desirable but it’s not the highest priority for now. Maybe in the near future I might run:
- Minecraft (resuming previous efforts)
- Satisfactory (occasional)
- Factorio (occasional)
But otherwise most gaming requirements and needs for my “village” of friends and family aren’t necessary.
I think over the course of 2024, I’ve had to be much leaner, out of necessity. And I’ve come to find that some things just aren’t that necessary for “me and mine”. As I go into 2025, I head in with the goal of not buying any new technology, but rather just maintaining the status quo of what I have, and building a necessary platform of services for myself and my loved ones. It’ll be minimal, and a lot more wallet friendly, I think.
January and the start of February also brought on a look at better brain dumping knowledge from inside my head, into something tangible. Sort of a legacy thing, I think.
Obsidian has been a note taking app that’s based around the markdown plaintext format that’s been around for years. A lot of my friends with a hyper technical background have been huge advocates of Obsidian and in an effort to find something that’ll let me brain dump with an intelligent linking method that is code friendly — Obsidian came up time and time again.
This blog post is generally just me rationalizing why I’m switching to it, over the two existing services I use (and pay for) – Notion and Clickup – both of which are fantastic apps for people who need something a little more fancy. But up front, I think it’s best to talk about the cost of these apps. Both of these apps are wonderful; they cost money however, at $5 USD and $12 USD respectively, and this adds up to a little over $200 USD per year. More than one might think to affect one’s finances in these trying times!
And so…! In a bid to move towards reducing my overheads, I thought I’d look into DIY solutions that I can integrate or piggy back on more critical services. In this case, Obsidian – which can use iCloud Drive to store itself works well. iCloud isn’t a service I can easily get rid of – my mobile phone, my tablet are both rooted deep in the Apple ecosystem, as there’s health related devices and apps that are better on iOS than in Android or Windows for my situation (your mileage may vary of course). Lucky for me though, that this is still usable across Windows – meaning I technically don’t need to worry about something like the paid Obsidian Sync service.
Security is also another thing I find myself concerned a little bit about. There’s not much I can do about state level bad actors gaining access to my data (and I don’t think anyone’d find use for it) – but your typical cyber criminal is still a concern because they’re on an interpersonal level. Last thing I need is sensitive data (like health records) getting compromised and having them leveraged against me. But to make things worse, it turns out that Notion isn’t encrypted on any level – which kind of explains why it’s so easy to publish something directly to the web.
Scary.
Clickup is also web based and doesn’t do much better. I feel like Notion and Clickup don’t have the resources to build privacy on a level that Apple does with its iCloud services. Having been subject to some very public breaches of customers data (not Apple’s fault – they got socially engineered) – Apple has no doubt more than doubled down to make sure it never gets the blame for any cyber security breaches.
So, all in all – I’ve moved to Obsidian and as of the time of this post, it’s been almost 2 weeks. So far I’ve started to slowly port across the knowledge dumped in Notion into it. It’s a long, slow and tedious process, but the beauty of the way Obsidian draws links between articles (Wiki-esque) means that I don’t have duplicate style documents, unless I make them forcibly within the file structure of the Vault itself.
It’s also nice that I can write SQL-esque “Dataview” queries that can generate lists of pages within things. It feels a lot more like a programmer’s knowledge assistant than a “Note taking” application.
It feels natural using Obsidian now, and I keep improving how I use it as I go along, it’s still got that shiny new “Learning new hacks all the time” feel to a new application.
It feels like The Right Move™
With the implementation of my JT-LAB homelab; it stands to reason I should probably self-host whatever I can to try and get a decent use out of the stupid amount of money I’ve poured into the project. Being able to ensure I’m only sharing the data that I want to share as well (for whatever reason) – is pretty important to me as well.
Normally, the majority of my code to-date, has been stored on Github (which is fine, it’s fantastic and it’s an amazing free resource for the world). But my workplace had an implementation of Gitlab that I thought was done pretty well.
So, I’m going to take it upon myself to implement GitLab into JT-LAB, and make sure there is a version of my work that runs from GitLab. Github will essentially become my backup for code (Github being way more reliable in uptime than anything I’d run like GitLab etc.)
Why is this so important?
GitLab is going to work as my core repository and project management system; with it, I’ll be able to store and update my code for various projects as previously mentioned
What’s the challenge?
Well, after some wrestling it’s implemented – however the one area that I’m most unsure with, is the AutoDevOps feature of GitLab.
Lots to learn!