Hey, I have the same thing for my second router that works as an extender, to cover some remote area. I auto-reboot it every 3 hours during the day (I don’t during the night). Sometimes, it stops transmitting data before the 3 hours mark, so I have to go and physically reboot it. It always helps, while there are very rare occasions when this software reboot does not help.

I have no idea what’s going on. I’ve bought it cheap as a broken one, but re-flashing it to OpenWrt seems like solved all its issues. However, I’m not qualified to say there’s no issues with it. It’s just that from a user perspective, it works exceptionally well. I see no issues. Except this forced auto-reboot thing, but I think it could be me not understanding the networking properly, and doing something wrong / not optimal. It gets the signal wirelessly via 5 GHz band (for speed) and shares it via 2.4 GHz band (for the distance). I fixed some obvious mistakes with the help of a GPT, which seems to work better now. But I’m not really sure. Could be that it’s winter and it was cold in there, I have to see how it’ll behave during the summer.

Honestly, I even started thinking maybe it has no issues now, and I can remove that cron job. But I think I can live with being offline for a minute or two a few times a day, when I’m in that remote location.

Yeah, I mean. Tried to compliment your story with mine.

For the rebellion!

Here I am, with my Raspberry Pi 2B and a few Orange Pi Zeros. Plus one Intel Atom 230 (an obsolete thing from circa 2010s if not older, it has DDR2 memory) as a file server. Most of my servers are at Intel Atoms old motherboards with these integrated processors. They are decent for what I do.

Excuse me for not having too much to say besides everyone else in this thread. I just wanted to check-in and say something that I felt upvoting everyone isn’t enough. Wish you all the recovery you can get, and thank you for inspiring too. You’re doing great, at least your grammar is quite good for the situation. Cheers!

Why both Netbird and ZeroTier? Don’t they conflict with each other?

I’ve missed both projects. What were they? Are they like Jackett or Prowlarr?

Thanks! So, why does it matter? It’s a server, you can have it to do the job unattended. Or does it affect other services and you’re unable to use anything else before it finishes?

If not joking, what would you want a huge amount of ram for on a server?

Imagine buying one for cheap because it has some bad blocks and it’s unreliable to keep real valuable data on it! I have a 8 TB HDD bought for like less than a $100 a decade ago, from a friend though, as he had some bad blocks there. I host only media for the HTPC there, but it’s been a solid all these years. And when it dies, sad, but nothing valuable that I cannot redownload.

Ok, thanks. It really sounds like a simple solution to the problem. I think even if it does drain battery for some reason (e.g. a repository with a huge number of files), this could be automated, like the on/off switch to run the app to sync and be done with it.

On iPhone, I use sushi train, and it does automated sync via Shortcuts (a built-in app for light automations), via timers or other events like charging. It works perfectly fine for my use case. It syncs my notes multiple times a night, plus during the day while on charge or when I join trusted WiFi networks. I expect the same can be achieved on an Android. So, really, the CLI version might do the job plenty good, I believe.

How does it handle the battery life? Is it run all the time or do you just start it to sync when you need it?

Can you explain this DNS thing further, please?

I start with what I understand. DNS stands for domains name system, which means a huge database of domain names and their IP addresses. When I ask for a website, DNS tells my computer / browser which IP addresses to look for, to reach the website.

At home, I have Pi-Hole and Unbound. The first one censors DNS addresses by not including domains that serve advertisements. It can work with various DNS providers, including those from Google or Cloudflare. Unbound allows me to self-host DNS database, periodically fetching it from somewhere. That way my ISP may not see … here I’m not sure what, DNS lookups? It sees which IPs I reach, so I assume there’s no big difference, if they’d want to know which resources I reach for. Frankly, I don’t understand this solution entirely, perhaps unbound is for something different. I used Pi-Hole without it for years, only recently I added unbound, because it was quite easy to do with DietPi distro.

Cloudflare actively promotes their WARP service, for people to use their DNS servers. They have three options, four ones, three ones and two, three ones and three. My guess is they theoretically can analyse these DNS lookups for some reason. (E.g. by partnering with three letter agencies, doing some service for them.)

What is DNS in the context of my website being registered with them? When I reach to my website, or any other website registered with them, what would happen? Isn’t the record everywhere already? I cannot understand what this means in this (different, isn’t it?) context.

The rug pull scheme ‘now you pay us for DNS too!’ seems unlikely, for some reason. If it’s no different from what they provide as a free service. If it’s something else, I assume you can migrate to any other registrar, unless you’re too heavy into their ecosystem.

On a personal note, I’m not too heavy into their ecosystem, I hope. I have a couple of static websites hosted for free with Cloudflare Pages. Plus I have a bare metal file server with images which is shared to the internet with Cloudflare Tunnel. I’m nobody with a few readers, tens of posts and hundreds of images, and I chose this architecture because I don’t understand how to properly self-host my blog on a residential connection (meaning dynamic IP behind a CG-NAT or what it’s called). When I do, I may drop them in favour of a simpler architecture. But also I was curious how it works.

So, saying all this, I still don’t understand what this them being an authoritative registrar means in this context. Perhaps I lack some web dev skills to understand that properly. When I had my domain with Squarespace, they allowed more than Cloudflare, but I lack understanding to properly formulate that, to even understand what it was. I think I could host my top level domain with Cloudflare Pages only when they are my registrar, while having those Pages on a subdomain was trivial even with a different registrar. If I remember that correctly now, I might’ve been confusing some things here.

Thanks for your previous explanation, it was quite informative.

Thanks! I haven’t thought of com as being the real TLD, actually!

Thanks! It’s a bit more clear now.

To contribute to the discussion, I remembered that with Squarespace (my previous registrar), I had unlimited redirects, which I used heavily. I am not really sure about the unlimited part, perhaps that was hidden somewhere in the interface, and they have limits, and I just never saw them. But I remember Cloudflare communicated I have like 10, so I decided to not use it for nice-to-have but not really needed things. E.g. I used a subdomain for a blog, and created redirects for typical misprints in my name. Was handy, but not really needed. I should have document this, but I was too busy at the time, and now, almost a year later, I don’t really remember. There were differences with Cloudflare and Squarespace.

My first registrar was Google domains. As always, they killed the business. And sold it to Squarespace. I’ve been their customer for a year or two, nothing bad I can say, except the price was about 1.5 or even 2x of that from Cloudflare for com domain, so I migrated there. I have no deep understanding of the nuances, so I cannot say whether Cloudflare is a bad actor. At least I trust them to not elevate the price, as it’s not their primary business, sell domains.

I see that, but what does it mean in practice?

I have my domain with Cloudflare too, and at this point, I’m not aware of these DNS servers. Can someone explain it a bit? I know what DNS is, but I don’t understand what’s the use case for having them elsewhere. I’m not to argue, just didn’t know where to register a domain, so I went with them. I’m concerned with the future of the domain either, but don’t understand the issues at this early point.

It’s even worse than advertisement. As it’s useless.

Oh, thanks, I missed it. It’s a very long thread. I’ve read only the first 40 messages so far, so I cannot really comment on that. But here is a nice advice from there:

FWIW It is possible to run Syncthing via Termux — it’s not as integrated but it runs fine.

What is wrong with the fork from F-Droid? I use it. I see no difference with the original, I’d say it’s even better. If you don’t trust them for some reason, why discard Syncthing as a project? I assume it can be built then. But I have no idea how.

By the way, I’m happy to use Sushi Train on iPhone. Works very well, and is lovingly polished. Now Syncthing is a centrepiece of my workflow to sync my files.