Overengineering my homelab so I don't pay cloud providers

Google One for 10TB is 274,99€/mo (at least in my country) so you'd make the entire nas price and subscription cost within a few months, let alone years.

There just aren't compelling public cloud for large sizes (My NAS is 30TB capacity and I'm using 18 right now) and even if you go the more complex loops with like S3 and whatnot you still get billed more than it's worth. Public cloud is meant for public files, there's a lot of costs you're paying for stuff you don't need like being fast to access from everywhere.

On my homelab, I update everything every quarter and it takes about 1 hour, so 4 hours a year is pretty reasonable. Docker helps a lot with this.

And I’ve almost never run into trouble in years, so I have very few unexpected maintenance tasks.

EDIT: I am referring to a homelab that is only accessible for private purposes through a VPN.

VPS are very expensive for what you get. If you have the capital, doing it yourself saves you money very quickly. It's not rare to pay $50 for a semi-decent VPS, but for $2000 you would get an absolute beast that would last 10 years at the very least.

With Docker, maintenance is basically zero and unused services are stopped or restarted with 1 command.

However, it also depends on how you use that data. In my case, I'm a Sunday photographer, so I tend to wrangle multiple GB of data at a time. I usually edit my photos locally, but I sometimes will want to revisit older stuff. I can download it, but it's a PITA and s_l_o_w. Google drive file stream is terrible for this, you never know if the files are uplaoded or not. Onedrive isn't any better. I haven't tried dropbox.

Hetzner has some storage box which exposes SMB but doesn't seem to enforce encryption nor IP filtering, so I'm not very comfortable with that.

Also, my internet connection pretends I have 5 Gb down, 0.5 up. The down part is usually as expected (my machine only has a 1 Gb nic), but upload is sometimes very slow. Running a local NAS is much faster. It's ZFS, so backups are trivial to send to encrypted offsite storage.

It also doesn't need to run 24/7, which helps with power usage (0.22 €/kWh here).

> I'll just add, as you grow older, you come to realize that time is a finite resource, and while money may seem like it is finite, you can always make more money.

Indeed. Waiting around for files to transfer gets old quick. I have better things to do with my time. My NAS needs a whopping five minutes of my time every now and again when a new kernel comes out.

The thing is, it's worth it to learn. Do you know the basics of how to set up a completely redundant environment? There's no conceptual difference between setting one up at home by using consumer equipment and setting it up in a data center. You can get pretty capable equipment (Mikrotik) for less. The enterprise stuff has more configuration options, but it's doubtful that you'll use most of them.

Set up backup WANs, redundant routing, DNS, power, etc is fun. Setting up redundant load balancers, backend services, databases, etc is also fun. It's not hard to do, it's just hard to get it all right. There are probably a zillion configuration parameters you can mess with, and only a few sets actually work. Unfortunately, the sets that work in your home won't be the ones that will work in production, but you could possibly run load tests etc to simulate a real environment (though simulating multiple clients from multiple endpoints is harder than you think).

And of course, getting production equipment is hard. Nobody has 2 F5s lying around. And you really need at least 4 F5s, because you have redundant locations. That's a lot of cashola. And in most environments you wouldn't want some random person messing around with the production (or test) F5s. It's the same with NAS, VM servers, docker registries, etc.

I suspect getting the whole end-to-end setup isn't something people experience anymore, because small companies have (or at least should have) moved to the cloud by now.

Even at the high end estimate the homelab is giving you several times the storage for the same cost.

My homelab is my hobby. I maintain it for my pleasure and to learn new skills. We have an infra nerds club with a few colleagues and we're having a lot of fun comparing our approaches!

It’s also an excuse for me to stay in most summer days.

I, for one, don't want to have Google, etc. as a dependency[1], so I will pay some energy cost to do that.

1: see: https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que...

Who are you to tell people how to spend their time? Let people have hobbies ffs

Every homelab I have come across is a hobby project and time sink that is more like a backyard garden or classic car.

Agreed, but it doesn't have to take time from your family. I'm on a small team that self-hosts internal services to lower costs/risks. It takes very little time to maintain, and maintenance windows happen on our terms. Our uptime this year is better than "Github Actions", the latency is incredible, and we've had no known security issues.

There are two keys to doing this successfully: (1) don't deploy anything you don't understand (so it won't take you long to fix), and (2) even then, aggressively avoid complexity (so it doesn't break in the first place.)

For example, despite significant network expertise, we stuck to a basic COTS router and a simple IPv4 subnet for our servers. And the services we run are typically self-contained golang binaries that we can deploy with bash onto baremetal. No docker, kvm, ansible, or k8s.

This DIY setup saves us considerably more than it costs. Not for everyone, but with proper scoping, many readers of hacker news could pull this off without losing time with their loved ones.

To some, spending time hunched over servers is doing things they love.

I mean, each and every thing you said about maintaining a home lab you can also say about maintaining cloud infrastructure.

There was a time when having hobbies was normal. It seems nowadays some people mistake hobbies for work after hours? Where is that hacker mindset?

don't spend your time cooking food, pay for others to prepare it for you.

don't spend your time maintainig a house, rent and let someone else do the maintenance.

just lease a car and get a new one automatically every 3 years.

honestly, everyone has their own setpoint for things. and there are degrees of solutions for every point you make.

I think most people would benefit from being just a little bit self-sovereign.

Personally, "majority of people" could use one low power fanless server with 1tb for the few things most people need continuously online.

And separately a server you turn on occasionally with lots of storage, like maybe

https://www.amazon.com/dp/B09TV1XPDD

I'm reminded of jwz's backups info www.jwz.org/doc/backups.html

"RAID is a goddamn waste of your time and money"

If your OS is using systemd, you can fix that pretty easily by adding an After=network-online.target (so the ExecStart doesn't even try to check if there is no networking yet) and an ExecCondition shell script [1] to actually check if nfs / smb on the target host is alive as an override to the fs mounts.

Add a bunch of BindsTo overrides to the mounts and the services that need the data, and you have yourself a way to stop the services automatically when the filesystem goes away.

I've long been in the systemd hater camp, but honestly, not having to wrangle with once-a-minute cronjobs to check for issues is actually worth it.

[1] https://forum.manjaro.org/t/for-those-who-use-systemd-servic...

So even if your local node could transmit, none of the others could, and they can't buffer either.

To mitigate power outage, you would need both power, and a cellular connection, and that connection would only be good for 2-3 hours (Cell tower backups), and those would require a Cradlepoint.

What I don't want though is a power outage putting my server offline while I'm on holidays, and not be able to access my services at all.

My ISP-provided router supports Wireguard, so I can use that to connect to my KVM and send the Wake on LAN packages.

Do you have power outages often? Even if I have one, my services can come up automatically without doing anything, when the power is restored.

Can you share more about this? I have a APC Back UPS PRO USV 1500VA (BR1500G-GR) and it would be nice to know if this is possible with that one as well.