I fixed DNS

(My DNS queries were blocked by my ISP’s modem, I flashed OpenWRT on an old WiFi Repeater, and set up a DoH proxy)

Thanks for the advice. I also use a cheap domain with a wildcard, but use nginx instead. I just tried using Adguard and although it’s fascinating to see the insights of all the DNS requests, it didn’t really help fix the issue. However, since using DoH with Cloudflare in combination with setting it to the specific IP instead of my local device name and have 100% uptime now (since the last 10 minutes that is).

I’m using a public DNS record that points to a local device.

*.example.org → example.org
example.org → device_name.lan

Uptime Kuma seems to use nscd for caching internally and the default system DNS resolver. I’ve added a custom DNS resolvers to Uptime Kuma, and apparently it can get the records from Cloudflare (1.1.1.1) but it can’t get it from the OpenWRT router (192.168.1.1). 🤔

I’ve enabled a proxy on the router to force the use of DoH, maybe that will help if the ISP’s modem is at fault.

Since the records have TTL of 5 minutes wouldn’t dnsmasq have to reach to upstream DNS servers every 5 minutes?

If you self-host your own instance, make sure to disable image hosting / caching. I’ve had to DM a lot of people to inform them of “problematic” images hosted on their instance.

Good, I don’t get why so many people still use mailing lists. I don’t like Discourse that much, but it’s a big improvement nonetheless.

And without the unnecessary www.. This article could be shared with the authors name as user like jason-koebler@404media.co and the category (in this case generative-ai) as a moderator-only community like generative-ai@404media.co.

Which provider did you use? Also, Hetzner costs the same but with 8GB RAM.

I use Keycloak

Old PC’s and especially laptops (make sure to consider removing the battery though) make great homeservers. You can run dozens of services on old hardware.

Yes, but if you care about power efficiency then they really aren’t a great option. Most professional server hardware that you can get for a decent price uses significantly more power than an old mini computer or a cheap N100 PC. I own a proliant but rarely power it on due to the fact that I could rent an similarly performant VPS for 2x the power bill. Besides that many server CPU’s don’t have integrated GPU’s and will require additional hardware if you want to run something like Jellyfin.

const ref or unique_ptr if you need ownership

Yes I remember the lemmy.world servers being DDOS’ed every couple of days and having to switch between 3 clients and the webinterface because all of the apps were missing some features. The alternative frontends like photon and tesseract have really improved and imo should be the new defaults.

Some time ago I wrote a program in COBOL and half of it was just CALL "SYSTEM" BY CONTENT in the end.

Microsoft have been endorsing Linux recently for this reason.

I doubt that’s still the reason, I think they endorse it because they’re making a shit ton of money using it on Azure.

Even if the instance doesn’t have any rules about it?

It’s a spammer AFAIK, I considered the possibility of it being used to get IP’s and deanonimize users through the image but it seems this isn’t the case since the image points to a generic pictrs file from a Lemmy instance and most clients / frontends proxy image loading.

Have you used both languages before?

Rust code (and the ecosystem) seems to put a higher priority on describing code contracts through the type system and documentation. I personally feel you don’t need as much context to write functional code compared to C where every corner of the codebase might be hiding something that could impact the part your working on.