Ok… sure. But what physical devices would I use, and what software would they run?
Ok… sure. But what physical devices would I use, and what software would they run?
Are there any “open” solutions to mesh networking that can compare to TP-Link Omada? I don’t think any open source hardware or software can come close, especially not for the newer Wi-Fi standards.
I haven’t bought them yet, but I’m seriously thinking about some Omadas. I imagine I can prevent them from phoning home, and the management software can run locally in a Docker container. Running it like that would be good enough for me even though they’re not “open.”
I’m planning a rework of my home Wi-Fi, and my current plan is an OPNsense box from Protectli, and a few EAP772’s:
https://www.tp-link.com/us/business-networking/omada-wifi-ceiling-mount/eap772/
If there’s something comparable/better that’s more of an open ecosystem, you definitely have my attention while I’m shopping around for different options.
Definitely recommend Motrix:
If the Google download link supports it, it should be fairly resistant to interruptions. If it doesn’t, this might not help much, but you should still use this instead of just a browser.
I haven’t tried to download a Google takeout, so you might need to get clever with how you add the download link to it.
If you just can’t get it to work, you can try getting the browser extension to automatically send all downloads to Motrix. There is some setup required, though:
https://github.com/gautamkrishnar/motrix-webextension
Good luck!
Before it got enshittified with an update a few years ago, I used the RealVNC Android app to connect to a few of my own VNC servers. Wasn’t interested in any of the fancy features, I just wanted a good VNC app.
Now I use AVNC. It’s solid, performs better than RealVNC used to, and it’s open source! You can get it on FDroid.
It should still work!
I only go back and make changes to LED if something breaks with a major Lemmy update, but Lemmy hasn’t had a major update since January. Lemmy v0.19.4 isn’t released yet, but when it is, I’ll make sure the deployment is up to date.
Note that it does not have any advanced features that a major instance might want, such as storing images on S3, exporting data, or image moderation. If you intend for your instance to grow for 100+ users, this isn’t for you. This is only intended for beginners who are overwhelmed by the other Lemmy hosting options, and want an easy way to host a small single-user or small-user instance.
You also can’t open two spreadsheets that have the same filename. I’m sure that’s led to a helpdesk call or two.
I’m scratching my head to think what Vultr could do better in this case
There was substantial room for improvement in the way they spoke publicly about this issue. See my comment above.
I still don’t like how flippant they’ve been in every public communication. I read the ToS. It’s short for a ToS, everyone should read it. They claim it was taken “out of context,” but there wasn’t much context to take it out of. The ToS didn’t make this distinction they’re claiming, there was no separation of Vultr forum data from cloud service data. It was just a bad, poorly written ToS, plain and simple.
They haven’t taken an ounce of responsibility for that, and have instead placed the blame on “a Reddit post” (when this was being discussed in way more detail on other tech forums, Vultr even chimed in on LowEndTalk).
As for this:
Section 12.1(a) of our ToS, which was added in 2021, ends with “for purposes of providing the Services to you.” This is intended to make it clear that any rights referenced are solely for the purposes of providing the Services to you.
This means nothing. A simple “we are enhancing your user experience by mining your data and giving you a better quality service” would have covered them on this.
We only got an explanation behind the ToS ransom dialog after their CMO whined in a CRN article. That information should have been right in the dialog on the website.
In both places, they’ve actively done vague things to cause confusion, and are offended when people interpret it incorrectly.
You are giving it the -d
flag. -d
means “detached.” There are logs, you are just preventing yourself from seeing them.
Replace the -d
with an -i
(for interactive) and try again.
Have you completed the podman rootless setup in order to be able to use it? You may need to edit /etc/subuid
and /etc/subgid
to get containers to run:
More than likely, this might have something to do with podman being unprivileged, and this wanting to bind to port 80
in the container (a privileged port). You may need to specify a --userns
flag to podman.
Running in interactive mode will give you the logs you want and will hopefully point you in the right direction.
What’s the logo that looks exactly like Arch but with a stylized X over it? I can’t find any distro with that logo.
Hi :)
If you’re already running an instance, you’re not going to have a good time of this on the same server unfortunately. The webserver config I ship assumes a single instance, and all of the handling assumes only one domain. You would have to basically modify my entire script to support something like this.
You can take a look at my advanced configuration page to figure out what files you can edit, but this would be a very manual process for what you want to do.
Apologies, but you would be better off deploying a new server.
I use Backblaze B2, but stored in an encrypted Restic container, set up using this guide:
https://helgeklein.com/blog/restic-encrypted-offsite-backup-with-ransomware-protection-for-your-homeserver/
Restic has been great for automating backups, and even letting me mount the encrypted storage to grab individual files. I like doing it this way since I don’t have to trust Backblaze isn’t reading my data - I know for sure that they can’t.
Performance of storage that is both remote and encrypted is about what you would expect, but I don’t need access to the data unless something bad happens.