ShellMonkey

You can buy whole kits to build your own, the capacity to control what is made with the is about as much as if you tried to ban someone with a table saw from building a birdhouse.

Encrypt before send, and if you want to have protection against deletions of the data have a cold backup offline other than during the copy.

The purist part of my head wants to define self hosting as something done on your own hardware that no actor external to your net can influence directly, even to the point of requiring a licence check against an outside server is not ‘self hosted’.

By that definition it gets a bit dicey with a lot of projects that are at all complex though, so you have to decide your own line.

Firewalls are typically built using ‘in’ rules, but it’s not referring to ‘wan > lan’ but rather if the rule is processed on the ingress of the firewall or the egress.

Stateful firewalls sessions are always going to allow responses, the old request that someone needs a rule ‘both directions’ is something of a running joke at this point honestly.

A pcap would explain a lot, the fact that the error cites a specific domain does sound like an outbound attempt, but if it was a specifically inbound region block that wouldn’t add up.

Assuming that the ‘enter.domain.here’ is a masking of the real domain, I would have to wonder if the acme client is actually trying to use the http validation rather than DNS. You could pcap on the outside to look for it coming in to confir.

As for filtering to specifically allow their addresses, I think their system has a unique user agent that you could filter against, or use an inbound proxy to allow global access to the ‘/.well-known/acme-challenge/’ path while keeping the rest of the site blocked.

Could use something like nextcloud as a ready made option, not the greatest for things where there’s frequent updates on large files though since it doesn’t send just partial updates. To some extent it depends on what your use case is exactly.

I’d gotten a lifetime premiere key, mostly for the ldap functionality at the time. Something I can buy once and use rather than subscribing to everything is always a winner for me.

I poked at Jelly a bit ago, but as I recall the auto organize was abandoned (kinda a big deal to me) and something about conversions being a problem. Basically I just didn’t see anything that would make me inclined to go through the hassle of reworking what already works.

Any reason it wouldn’t work with Emby too? Have had it since before there was a Jellyfin spun off so I’ve just rolled with it.

My go to for messing with chat bots is Kobold that’ll let you split the work between multiple GPUs. I get the impression the actual processing is only done on one but it lets you load larger models with the extra memory.

I was using a Nvidia 3060 for a while, then had 2 in one box, then switched to a 3090.

The amount of vram is a big factor for decent performance. Getting it to not sound like a predictably repetitive bot though is a whole separate thing that is still kind of elusive.

Have run Lemmy and now Piefed, it’s nice to have things customized to your wants, but probably wouldn’t bother if it was setting up a host just for that.

Right, and if both sides have their public ally routable IPs on their respective firewalls it’ll work. If on gets put behind a NAT of some sort then it would be able to speak outward, but would require specific packet routing inward (port forwarding) to have someone connect in. Stateful sessions will be fine so long as the one inside a NAT is the initiator.

If only one side is behind a NAT then so long as that one initiates the tunnel it should work fine. NAT only really is a problem on the inbound side.

Snapshots largely, most everything is VMs and docker containers. I have one VM set aside for dev work to test configs before updating the prod boxes as well.

Sooo, they act as an admin and can assign mods instead of using someone else’s node and being a mod?

Aside from what I understand as an inability to actually remove bad content that gets in, how does that differ from something like hosting a fedi site?

Then you get things like the platform you’re on where in my case it resides in my house and lets me be that big scary admin/mod. Having the ability to purge bad content and actors from a central space is needed for anyone but the most thick skinned masochists to use a platform.

Plenty of people just want to go talk/post without wading through a swamp of the crap that one uncle brings up at Thanksgiving on a regular basis.

The system formerly known as Freenet has a module known as the web of trust that uses a similar model. It’s interesting but runs into a problem of forcing users/hosts to propagate content and messaging they don’t wish to be associated with.

There’s a reason places like gab or hexbear end up isolated islands, the general population has no desire to be preached to be the lunatic fringes.

Last I ever was hearing this pushed around the fedi the big ‘sell’ was that mods/admins can’t delete posts making it a ‘freeze peach’ platform.

The only people typically drawn to those are the people who tend to get banned for being intolerable on civilized platforms.

https://docs.mailcow.email/

Has worked well for me, but keeping in mind that email is probably about the hardest thing to properly self host.

DMARC/DKIM/SPF/MX records have to be set up properly, DNS for auto config and such preferable, and dealing with a myriad of mail IP blacklists (good monitoring service for free: https://mxtoolbox.com/) all come into play.