For anything important, use matrix instead of lemmy DMs.
All good answers so far.
Basically your internet gateway doesn’t support what’s called NAT hairpin, also called NAT loopback.
The DNS solution is better anyway, and that particular setup is called split-horizon DNS, which basically just means your internal DNS requests get different results than ones from outside your network.
For one thing, it still has the option for an additional fingerprint lock to open the app.
Sure, my phone itself is locked anyway, usually.
I still like having the second lock built into the app itself for the odd time (however rare) that I hand my unlocked phone to someone.
I know they’re not snooping into my signal messages while I blink or misclick on a notification or whatever.
It may not be a big deal for some/most but I like that feature and I can’t imagine that’s something that needs much upkeep once initially implemented.
I’m sure it’s not the only feature, but it’s one I use often.
Signal (but actually Molly)
My customer decided to host their shit there ¯\_(ツ)_/¯
Coming soon:
Fortran code from 1969 that has been vibe coded since.
Yea, I don’t think I want the average YouTube commenter to come here.
Yea, I was referring to the OSI model.
Whatever you do, make sure you have working backups first.
I imagine you could copy the docker volumes over, but that’s more work than of they’re “mounts”, in which case you can just copy the corresponding on the host. Use scp or rclone or whatever to copy the files over
A single misconfigured thing can suck real bad as you’ve seen.
Selfhosting involves lots of things that can be misconfigured or go bad.
That’s not to scare you out of it out anything, merely to congratulate you in seeking knowledge first.
Disclaimer: I’m biased towards networks because I’m a network engineer, opinions may differ.
I would say… having at least a vague grasp of layers 1-4 of the traditional network model is a decent start.
You don’t need to understand everything, but knowing a minimum will help a lot imho.
It’s hard to point you in the right direction without knowing what you already know or not.
Honestly, I’d just use whatever the ISP provides.
Sure, it’s not open source and it kinda sucks…
But I mean, if you don’t trust the ISP modem, you can’t trust the rest of their infrastructure either anyway, so it’s kinda moot.
At least that way you have a vague chance of having a modicum of support when shit breaks.
If it can’t be put into bridge mode, it probably has some sort of DMZ function where it basically does port forwarding for any/all possible ports.
Double NAT isn’t as bad as it sounds these days.
Now to your question…
They exist, they’re mostly targeted at ISPs though, so might be a harder find than other things.
They might also be older, as basically all customers also want their ISP to provide Wi-Fi, which a bridge modem won’t.
Anyway,
You’ll have to know what DSL were talking about, there’s… ADSL, ADSL2+, VDSL, VDSL2, etc
One old-ass model we used to use back in the day was… a Siemens 5200, but that’s ADSL2 at best, definitely not VDSL.
I’ve not had good luck finding something that gives me the confidence to go about it
Now’s a good time to make sure you have good backups.
Knowing you can fallback to your backups helps a lot with confidence.
One this that’s really hard to replace is DDoS protection.
If it starts fine, but performance drops off a cliff after a while, it might be a heat issue.
I ended up putting such a flaky drive in a bag and wrapping a flexible icepack around it.
Not a great solution, but kinda good enough to get some data off of it.
I use this one:
https://play.google.com/store/apps/details?id=de.orrs.deliveries
It can usually get the status of most courriers unless they’re actively hostile against API/scraping.
Worst case you can still track status manually, or open the webpage.
It doesn’t track payment, but I don’t usually have a tracking number before paying.
I think you could technically add your own custom status entry.
I keep them as active until shipped and verified after which I mark them as completed.
Works well enough for my needs.
Aye, most of that activity was generated through user error on my end.
Jet is legit, but got caught by the bot or something and then the manual unban didn’t work right… so I site-banned them again (from sjw) and unbanned them, to try to make it federate, it was a bit of a mess. Lemmy generates community bans with site bans, not sure why that didn’t undo with the unban.
Either way, totally my fault.
Seems the content in the screenshot was from !networking@sh.itjust.works.
They were indeed still banned from that community, but shouldn’t have been, I’ve unbanned them manually.
I’ll check the other communities too.
You can use an adapter just fine.
Or use a 5.5" drive caddy, that’s just a little drawer that slides in and out.
Real question is it you have enough SATA connectors available.
It’s a bit advanced, but I have a GPU in my server that doubles as a home theater / couch gaming thing. I PCI passthrough the GPU to a windows VM where I get near baremetal performamce.
It runs ontop of proxmox and has a bunch of containers and VMs for my other stuff.
Currently, the linux part is headless, as in I don’t need a screen for the linux stuff on there, but it’d be a matter of hooking up the onboards graphics to another monitor input.
I like the idea because unlike dualboot, it all runs at the same time.
My other desktop’s mobo has issues with vfio and iommu groups so I can’t really do this on that other machine, but for my next build, good iommu groups will be a deciding factor.
I’d love to have a similar setup for my desktop and just switch monitor inputs or KB/mouse USB switch between both.
Warrantied drives still fail, they just happen to ship you a replacement.
Commercial drive trashing solutions are basically a smaller, fancier version of the mechanism in a log splitter.
You could probably rig a sketchy drive wedge/bending thing with a pump jack rather easily.
Wear PPE.
The odds of someone taking a failed drive and transplanting the platters to a working drive is pretty low to begin with.
Me? I don’t have tons of drives to destroy, so I just unscrew the thing, get the platters out and smash those.
I’m a big fan of mutual TLS, which I just do in nginx reverse proxy.
Client certs get installed on trusted devices and then it’s mostly seamless.
Not every app will support it, but I use Firefox as a fallback for stuff that doesn’t, or my VPN.
TBH, I’m just happy homeassistant supports it as that’s the one thing I’d rather be as seamless as possible.
Other things I don’t mind having to connect a VPN for, like jellyfin, which I don’t currently share with others.
For nextcloud I’m currently not running mTLS as it wasn’t supported back then and I had to share some links with some people where I wouldn’t deal with certs. I think mTLS support has been merged since, so if you just use it for devices under your control, or might be viable