I’ve been running two piholes for a few years:

I let the router handle dhcp, while having it hand out each of the piholes IPs as DNS #1 and DNS #2. Those piholes settings are mirrored using Nebula-Sync, so any change made to one pihole gets copied over to the other automatically.

Both of those piholes are setup to wrap outbound DNS reqests in DOH as well, with the router blocking port 53 outbound. So all DNS traffic leaving the network has to be DOH/DOT or it just gets dropped. Currently using NextDNS and quad9 as my upstream DOH providers.

Note: the cloudflared method for DOH is outdated. I’m still using it via the docker image “image: cloudflare/cloudflared:2025.11.1”

I believe This (dns-crypt) is the more recent method.

For a while, the windows machine I was using kept randomly becoming unresponsive (both via network, as well as the keyboard+mouse) and I just could not figure out why. Everytime it did, I’d have to force a shutdown by holding the power button until it shutoff, then press it again to start back up. Sometimes this happened while I wasn’t home, so it had to just stay offline until I was.

I got sick of having to do this; so I replaced the power button with a transistor and a RPI. The PI would ping the server every 5min; if it failed to get a response 3 times in a row, it would trigger the transistor for 10sec, release for 3sec, then press and release again; forcing a poweroff then starting the machine again. It’d also write these events to a log file. I think it was around twice a week ish.

Never did figure out what the cause of the lockup was; but it stopped when I replaced windows with Debian.

I guess it automatically locallizes the price. Shows me 23.99CAD.

Doesn’t show me a captcha either.

I don’t see why not, and for $24 you may as well give it a go.

That’s one of the joys of FOSS, when the dev for a popular project can’t continue; there’s often someone willing to pick up the torch and carry it forward.

I switched to this fork when the one you’re thinking of ended development.

https://github.com/nicholas-fedor/watchtower

Last update 3 days ago.

Homepage is my main front end/landing page for reaching/monitoring all my services; but I’ve got both my dockge instances linked together as well, so all my services show up in both.

I run watchtower on a cron schedule. It had already fired for today, so it wouldn’t have run again until tomorrow. I could have waited, but manual is just as easy.

I do love docker + a front end like dockge.

Saw this post, switched to my browser, clicked three buttons; vaultwarden is now updated. Didn’t even move my lazy ass off the couch. (I’m on my phone)

Thanks for the PSA! :)

They’re being charged in the UK. All Trump can do is refuse to extradite; but that just means they’re bared from the UK and anywhere else that has an extradition treaty with the UK.

Watch Moscows AA shoot his plane down 😂

Fair points.

I’ve been lucky enough to have never been behind cgnat, so I keep forgetting about it.

My bigger concern with tailscale is being required to install software on the client. Not every device I use, I have permission to install a vpn client, nor would I want to.

For example, I have a fileshare using Filebrowser where I store work related files that I don’t want to loose access to or need access to from multiple machines (non proprietary info, stuff IT/MGT wouldnt get mad at me for ofc. I’ve actually cleared it with my managers, so no worries). That’s also a handy way to (temporarily) share large files with people or provide a way for friends to upload large files to me.

I also like to access my emby server (using sufficiently limited accounts), from things like the TV in the work break room, or a friends PC while I’m visiting.

Tailscale is a hurdle that I just don’t need/want.

Glad I could help. I’m not always immediately available, but I don’t mind answering questions if you run into troubles. Just send me a DM and I’ll do what I can. :)

You don’t need a static IP, you just have to keep track of what your current dynamic IP is.

You can do this with either a free or a paid DNS service.

There are a few different ‘free dns’ services that will delegate a subdomain of theirs to you at no cost. Admittedly, I’ve never actually used one of these so their names escape me. Hopefully someone else can point one of those out if that’s what you really want.

I purchased a domain via google domains, when they existed. It’s now transferred to squarespace, because they bought out google domains a few years ago.

It was around $13/year when I first got it a decade ago. It’s now around $28/year.

This allows me full control over the domain: I can use as many subdomains as I want to give each service I use it’s own unique name. (Instead of using their own separate ports that you’ve gotta remember) My domain will also forward all inbound email to my gmail account; this lets me use email addresses like <servicename>@mydomain.example. This way, I don’t share my real email and can immediately tell who sold my info to the highest bidder when I get spam. (I could also host my own email service if I really wanted, but I haven’t bothered)

Add Cloudflare ontop (for free); and it can filter out known attacks, ddos attempts, geofence your services to regions you’ll actually be in, provide/autorenew ssl certs for https, show you usage analytics, cache static data reducing server/network load, etc.

Ultimately, the paid option is well worth it IMO. $2/month (which I typically pay in 3-10 year blocks) is hardly anything.

/edit; vpns are good and all, but they require you to setup software on the remote device to connect to it, and that typically routes most if not all your traffic back to the vpn server then out to the internet. That can create speed/bandwidth issues.

A domain allows you to access your services from any Internet connection with 0 configuration on the client side. Just accessing it like any other website.

I also host a vpn directly from my network, that I access/find via my domain. This means I’m not dependent on a public service like tailscale, but can still add additional security to access private only services (stuff I don’t expose to the open internet)

https://en.wikipedia.org/wiki/Network_address_translation#NAT_hairpinning

TL;DR Your router sees you trying to reach your external address and routes the connection back to your LAN without leaving the network.

This does still depend on a functional internet connection however, as your client gets your public IP from a public DNS server over the Internet.

If you were to run a DNS server locally (I use pihole for this), you could have that DNS respond with your local IP, allowing clients within your LAN to resolve the name without needing to reach out to public DNS. This means your local connections will still work when your internet is down; it also provides more privacy by keeping those requests local and can let you make local-only names that aren’t publicly listed.

Of the ~28 FQDNs in my setup, only 4 are public. The rest is local/vpn only and not publicly listed due the above. The reverse proxy then drops all connections that don’t use one of those recognised names, before even completing the TLS handshake. (So direct connections from someone port scanning my IP or using a domain name someone else has pointed at my IP are completely ignored/dropped without response. The server doesn’t even send the TLS cert so as to not expose the names defined in it.)

I try to be in the habit of making a full image onto a demonstrably working spare card every couple weeks.

That’s a whole lot of writing to an sd card, wearing it out. It may fail by the time you want to read it. You also destroy each previous backup by creating a new one.

Each of my rpis backup to my main server nightly using dd via ssh. The server then keeps historical backups of those .img files via Borg so I can pull any version from any day of the last year or so.

I use Emby instead of Plex or Jellyfin; mostly because it has an Xbox client, and I’ve already got a lifetime licence. One of my most active users only watches via Xbox.

Really don’t like Plexs centralised user system or the overall direction they’ve been headed for years, so I moved away from that long ago (8+ years ago at least). Jellyfin wasn’t up to par at the time (though they’ve made leaps and bounds of progress in that time), and Emby has always supported more types of devices\clients. Their device limit (the client count limit with premeir) has never come into play for me, but I know there are larger user bases out there where that is a problem.

Embys development is extremely slow though, taking YEARS to implement simple features or even address major concerns. Plus their support sucks without the community stepping in and providing it on behalf of the staff. Luke (the main dev) is better at copy+pasting candid responses than he is at actually interacting with human beings.

That icon didn’t show up until I installed another voice app. FUTO voice is what I’m using now alongside Heliboard. Seems to work nicely so far.

Thanks! That seems to integrate into Heliboard nicely, activating with the voice button that’s now on the toolbar where I expected it.

I wasn’t expecting this to be two separate components/apps, but I guess that makes sense.

This seems to be kind of the consensus. I’m not interested in moving from one theiving shitstain of a company to another. So gboard is a no-go.

I can’t get voice input to work on Heliboard though. The option is there in toolbar settings, but the button doesn’t show up on the toolbar and there’s no microphone permissions…

Do you have to do something to make voice input exist?

It’s listed as an option in the toolbar settings but the button doesn’t appear on the toolbar. Even with everything else disabled and just the voice input button enabled: the toolbar is totally empty…

There’s also no microphone permissions…