Wow, really interesting, thanks!

It didn’t fix everything, or even fully fix my Redis container. But it did improve things a lot. This is now my full Redis logs:

2025-12-29T17:01:52.277109213Z Redis has started
2025-12-29T17:01:52.746857972Z 7:M 30 Dec 2025 03:01:52.746 # WARNING: The TCP backlog setting of 511 cannot be enforced because /proc/sys/net/core/somaxconn is set to the lower value of 128.

Sorry, I’m not 100% sure I follow. Is this a quip about Nextcloud being difficult to run? If so, is it a loving one, or do you think it’s not worth the effort and I should try running something else?

The docker version says its API is on version 1.43. So I’ve fixed it by setting: DOCKER_API_VERSION: 1.43

Unfortunately I already tried this in my initial post, I’m pretty sure. It’s what I have with the errors shown in the OP.

Seems weird to me that there’s an AIO container that seems to contain other containers, but anyway I guess thats a synology thing.

No, that’s a Nextcloud thing. From what I can tell, it seems to be the preferred way of setting up Nextcloud these days. https://github.com/nextcloud/all-in-one

afaict it’s not that one container contains the containers, so much as one container is given control of the Docker socket so that it can create and control the other containers automatically.

all those containers are “starting” because theyre waiting for one other container to finish “starting” before starting up themselves

Apache and notify-push are waiting for Nextcloud. Nextcloud is waiting for Database. Whiteboard is waiting for Redis.

I have no idea what’s wrong with Database, Redis, or Collabora, but their errors aren’t obviously related to dependencies, to me. (Collabora’s could be, but it’s at least a different type of dependency since the logs are mostly [ remotefontconfig_poll ] ERR Remote config server has response status code: 502 (Bad Gateway)| wsd/RemoteConfig.cpp:133. I’ve not really started looking into it since it’s a rather downstream component and the core components failing is more important.)

Imaginary just says:

Imaginary has started

Is the 404 in the master container logs from you trying to access the instance in your browser?

Doesn’t seem to be. Seems to add a new log periodically even when I don’t try to load it up. I’m guessing the 404 comes from some kind of automated uptime checker?

deductive logic instead, or set theory, or group theory, or finite field topology

All of these are maths, though?

Also sometimes it actually just is good to know maths. Engineers, researchers, actuaries, accountants…there’s a huge range of practical applications for maths that’s more complicated than basic arithmetic in a very direct fashion, before you even get into jobs that more indirectly use abstract reasoning learnt through high school maths.

Really interesting! Thanks!

Part of the reason I ask: I bought a Synology (4-bay, 1 TB each in Synology’s RAID 5–like format) a while back with the intention of using it as a hybrid NAS & server. But have been repeatedly struggling to get it to do actually run the applications I wanted to run. And then started trying to get some of the things I wanted to run on my Synology to instead run on a Raspberry Pi…same problem. The weird architecture and distro makes some things not work smoothly. So have been thinking about getting a Mini PC with a standing x86_64 processor running a standard Linux distro to be a proper server, and using the Synology more exclusively for file storage & sharing.

Some of the things you listed there are definitely things that were on my radar to get around to running myself anyway (e.g. Immich, Watchtower). Others I had never heard of, but was already considering looking for an equivalent (Affine, Jotta—though for notesI was hoping Nextcloud would have a suitable app), and others I didn’t even think of (Linkwarden looks very interesting).

Since you’ve got a Synology already, I’m curious what value you see in the Mini PC running a Samba server? I’m also curious why you go with the Synology as your reverse proxy. Just the ease of the included tool? Does it also handle TLS termination, or how does that work in your setup?

I clicked into the comments for one fucking reason. And was left very disappointed.

What do you run on your home server? What sort of hardware is it running on?

Also Republicans:

But look, there was a shooting in Australia two weeks ago, so clearly gun laws don’t work!!1!11!!1!

I don’t particularly like the notion that it’s unreasonable to expect to be able to get healthy food at affordable prices without being required to use a car…

Or the conflation of using a cargo bike with being some impractical urban hipster.

Whoops. Just cleaning up some old tabs and realised I never responded to this. Thanks! It was some really interesting info.

dual booting, as a concept, almost always exists in relation with Windows

Hard to dispute this, except perhaps for the really niche situation of someone dual booting Linux on a Mac. Not especially useful very often, since Macs are a UNIX system. And because of that, not very common compared to Windows on a Mac, or dual booting Windows & Linux.

Ditto

They might have actually somehow managed to come up with worse logos than Google uses. I assume the top one is an email app, the bottom left is a file manager, and bottom centre is a password manager? But I’m less than certain about those, and everything else is a complete mystery.

@[email protected] is correct, but is perhaps not explaining it perfectly for the practical questions you seem to be asking.

If you have, say, two Docker containers for two different web servers (maybe one’s for your Wiki, and the other is for your portfolio site), you can have both listening on ports 80 and 443 of their container, but a third Docker container running a reverse proxy which has access to your machine’s ports 80 and 443. It then looks at the incoming request and decides which container to route the request to (e.g., http://192.168.1.2/wiki/%s requests go to the Wiki container, and all other requests go to portfolio site).

Now, reverse proxies can be run without Docker, but the isolation Docker adds makes it all a lot easier to manage, in part because you don’t need to configure loads of different ports.

Sorry mate, but you’ve got it wrong. The Prime Minister has specifically come out and said that this law is aimed only at companies, and that children or the parents of children who are able to get onto social media anyway will not be punished. Only the companies that let them slip through.

And you only need to read the legislation to see that that’s true. There are no penalties associated with accessing social media under the age of 16. Only with “a provider of an age-restricted social media platform…failing to take reasonable steps to prevent age-restricted users having accounts”. Or less closely related, “a provider of an age-restricted social media platform must not collect information…for the purpose of complying with [the above requirement] if the information is of a kind specified in the legislative rules”, and another similar “a provider of an age-restricted social media platform must not…collect government-issued identification material…for the purpose of complying with section [the above requirement]”, but this last clause “does not apply if…the provider provides alternative means…for an individual to assure the provider that the individual is not an age-restricted user”. It is also the case that a person who provides an age-restricted social media platform “must comply with a requirement…to give to the Commissioner, within the period and in the manner and form specified by the notice [about that person’s compliance with the law]…to the extent that the person is capable of doing so.”

That’s it. That’s all the new penalties that can be applied.

Here’s a page from the eSafety Commissioner that also confirms it.

Are there be penalties for under-16s if they get around the age restrictions?

There are no penalties for under-16s who access an age-restricted social media platform, or for their parents or carers.

The fact is, right now we know that Facebook has at times made a deliberate, conscious choice to leave in aspects of their algorithm that were causing harm. Their own studies have shown this. Making that practice illegal—knowingly causing harm with your algorithm—would be a good place to start with regulation.

no age can go on the internet.

I don’t think anyone had ever suggested anything like that.