It’s sitting at around 46GB at the moment, not too bad.
Instance is a year and a few months old, so I could probably trim down the storage a bit if needed by purging stuff < 6 months old or something.
I think it initially grows as your users table fills up and pictrs caches the profile pictures, and then it stabilizes a bit. I definitely saw much more growth initially.
I subscribe to a few more communities and my DB dump is about 3GB plain text, but same story, box sits at 5-15% most of the time.
I’d at least get a plumber to check it out. You could snake it yourself probably but you could also make it worse. If the pipe’s broken, you might as well just get more debris falling into it and clogging it further.
A regular plumber visit/check usually isn’t that expensive. Not cheap but far from 20k expensive.
It could also be connected to your flooding too, so you probably actually want to at least evaluate the damage ASAP. If the pipe’s broken, you just have a convenient pipe to drain all the rain water straight to your basement.
Having the web server be able to overwrite its own app code is such a good feature for security. Very safe. Only need a path traversal exploit to backdoor config.php!
Yep, and I’d guess there’s probably a huge component of “it must be as easy as possible” because the primary target is selfhosters that don’t really even want to learn how to set up Docker containers properly.
The AIO Docker image is an abomination. The other ones are slightly more sane but they still fundamentally mix code and data in the same folder so it’s not trivial to just replace the app.
In Docker, the auto updater should be completely neutered, it’s the wrong way to update the app.
The packages in the Arch repo are legit saner than the Docker version.
I’ve heard very good things about resold HGST Helium enterprise drives and can be found fairly cheap for what they are on eBay.
I’m looking for something from 4TB upwards. I think I remember that drives with very high capacity are more likely to fail sooner - is that correct?
4TB isn’t even close to “very high capacity” these days. There’s like 32TB HDDs out there, just avoid the shingled archival drives. I believe the belief about higher capacity drives is a question of maturity of the technology rather than the capacity. 4TB drives made today are much better than the very first 4TB drives we made a long time ago when they were pushing the limits of technology.
Backblaze has pretty good drive reviews as well, with real world failure rate data and all.
Ethernet splitter
What kind of splitter? Not a hub or switch, just a passive splitter?
Those do exist to do 4x 100M links on a single pair each, but you can’t just plug those into a router or switch and get 4 ports, it still needs to eventually terminate as 4 ports on both ends.
If you’re behind Cloudflare, don’t. Just get an origin certificate from CF, it’s a cert that CF trust between itself and your server. By using Cloudflare you’re making Cloudflare responsible for your cert.
There’s also Cockpit if you just want a basic UI
I believe you, but I also very much believe that there are security vendors out there demonizing LE and free stuff in general. The more expensive equals better more serious thinking is unfortunately still quite present, especially in big corps. Big corps also seem to like the concept of having to prove yourself with a high price of entry, they just can’t believe a tiny company could possibly have a better product.
That doesn’t make it any less ridiculous, but I believe it. I’ve definitely heard my share of “we must use $sketchyVendor because $dubiousReason”. I’ve had to install ClamAV on readonly diskless VMs at work because otherwise customers refuse to sign because “we have no security systems”. Everything has to be TLS encrypted, even if it goes to localhost. Box checkers vs common sense.
LetsEncrypt certs are DV certs. That a put a TXT record for LetsEncrypt vs a TXT record for a paid DigiCert makes no difference whatsoever.
I just checked and Shopify uses a LetsEncrypt cert, so that’s a big one that uses the plebian certs.
Neither does Google Trust Services or DigiCert. They’re all HTTP validation on Cloudflare and we have Fortune 100 companies served with LetsEncrypt certs.
I haven’t seen an EV cert in years, browsers stopped caring ages ago. It’s all been domain validated.
LetsEncrypt publicly logs which IP requested a certificate, that’s a lot more than what regular CAs do.
I guess one more to the pile of why everyone hates Zscaler.
That’s more of a general DevOps/server admin steep learning curve than Vaultwarden’s there, to be fair.
It looks a bit complicated at first as Docker isn’t a trivial abstraction, but it’s well worth it once it’s all set up and going. Each container is always the same, and always independent. Vaultwarden per-se isn’t too bad to run without a container, but the same Docker setup can be used for say, Jitsi which is an absolute mess of components to install and make work, some Java stuff, and all. But with Docker? Just docker compose up -d, wait a minute or two and it’s good to go, just need to point your reverse proxy to it.
Why do you need a reverse proxy? Because it’s a centralized location where everything comes in, and instead of having 10 different apps with their own certificates and ports, you have one proxy, one port, and a handful of certificates all managed together so you don’t have to figure out how to make all those apps play together nicely. Caddy is fine, you don’t need NGINX if you use Caddy. There’s also Traefik which lands in between Caddy and NGINX in ease of use. There’s also HAproxy. They all do the same fundamental thing: traffic comes in as HTTPS, it gets the Host header from the request and sends it to the right container as plain HTTP. Well it doesn’t have to work that way specifically but that’s the most common use case in self hosted.
As for your backups, if you used a Docker compose file, the volume data should be in the same directory. But it’s probably using some sort of database so you might want to look into how to do periodic data exports instead, as databases don’t like to be backed up live since the file is always being updated so you can’t really get a proper snapshot of it in one go.
But yeah, try to think of it as an infrastructure investment that makes deploying more apps in the future a breeze. Want to add a NextCloud? Add another docker compose file and start it, Caddy picks it up automagically and boom, it’s live and good to go!
Moving services to a new server is also pretty easy as well. Copy over your configs and composes, and volumes if applicable. Start them all, and they should all get back exactly in the same state as they were on the other box. No services to install and configure, no repos to add, no distro to maintain. All built into the container by someone else so you don’t have to worry about any of it. Each update of the app will bring with it the whole matching updated OS with the right packages in the right versions.
As a DevOps engineer we love the whole thing because I can have a Kubernetes cluster running on a whole rack and be like “here’s the apps I want you to run” and it just figures itself out, automatically balances the load, if a server goes down the containers respawn on another one and keeps going as if nothing happened. We don’t have to manually log into any of those servers to install services to run an app. More upfront work for minimal work afterwards.
Yeah, that didn’t stop it from pwning a good chunk of the Internet: https://en.wikipedia.org/wiki/Log4Shell
IMO the biggest attack vector there would be a Minecraft exploit like log4j, so the most important part to me would make sure the game server is properly sandboxed just in case. Start from a point of view of, the attacker breached Minecraft and has shell access to that user. What can they do from there? Ideally, nothing useful other than maybe running a crypto miner. Don’t reuse passwords obviously.
With systemd, I’d use the various Protect* directives like ProtectHome, ProtectSystem=full, or failing that, a container (Docker, Podman, LXC, manually, there’s options). Just a bare Alpine container with Java would be pretty ideal, as you can’t exploit sudo or some other SUID binaries if they don’t exist in the first place.
That said the WireGuard solution is ideal because it limits potential attackers to people you handed a key, so at least you’d know who breached you.
I’ve fogotten Minecraft servers online and really nothing happened whatsoever.
If your stuff is all Docker then yeah, immutable makes sense as it makes the entire box declarative and immutable: you can get back the exact same operating Docker environment on the server, and then you can get back the exact same Docker workloads going with the Docker compose configurations.
If you ever need to run stuff you’d run on Debian, you can just shove it in a Debian container.
That said, if most of the stuff is containers, the risk of just the core Debian breaking is fairly low. Pick whatever is easiest for you to deal with based on your needs. Immutable distros have a bit of a learning curve.
I learned it accidentally trying to get root on an encrypted dataset working with systemd init without sd-zfs. This turns out to be how the zfs utility works internally to signal the driver “hey it’s okay, I’m a ZFS utility the user isn’t using mount directly”, and how you deal with mounting your root dataset to the temporary /sysroot while having its mountpoint set to / while in initramfs before pivoting root.
Obviously, don’t use that other than recovering your data, if you want to use this array you should figure out the mountpoints properly so ZFS does it automatically. It shouldn’t break anything but it’s gross, either set mointpoint=legacy and use fstab or set its mountpoint in ZFS and use zfs mount.
The trick for this one is mount -t zfs -o zfsutil internal /mnt/some/path
Assuming the root dataset is mountable. If you have a -o canmount=off on the dataset it will refuse to mount.
If it’s -o mountpoint=legacy then you don’t need -o zfsutil, but still need to provide both the source and destination paths. Otherwise you’ll get the fstab error because mount can’t figure out what to mount or where to mount it.
Yep there’s a reason I reached directly for that configuration. WireGuard uses UDP, that’s one of the first things that gets blocked.
Turns out that’s also the kind of protocol corporate VPNs use, reusing port 443 over TCP. They call those “SSL VPN”. They get to weed out all commercial VPNs used to bypass their firewalls as well as most torrent/game activity while still mostly catering to their business guests.
With Docker, the internal network is just a bridge interface. The reason most firewall rules don’t apply is a combination of:
The only thing that should be affected by the host firewall is the proxy service Docker uses to listen on a port on the host and send it to the container.
When using Docker, each container acts like an independent machine, and your host gets configured to act as a router. You can firewall Docker containers, the rules just need to be in the right place to work.