• 0 Posts
  • 49 Comments
Joined 9 months ago
cake
Cake day: December 14th, 2023

help-circle


  • I’ve done a backup swap with friends a couple times. Security wasn’t much of a worry since we connected to each other’s boxes over ssh or wireguard or similar and used tools that allowed encryption. The biggest challenge for us was that in my selfhosting friend group we all prefer different protocols so we had to figure out what each of us wanted to use to connect and access filesystems and set that up. The second challenge was ensuring uptime and that the remote access we set up for each other stayed up - and that’s what killed the project as we all eventually stopped maintaining the remote access and nobody seemed to care - so if I were to do it again I would make sure all participants have alerts monitoring their shared endpoint.





  • To add to the other reply, client isolation is about controlling whether an ap, switch, or router willingly sends traffic between clients. Because of that, it doesn’t kick in if you listen to packets over the air before they’ve been received by an AP. For that kind of security you need a wifi specific security measure - which I think “enhanced open” is what you’d be interested in. It allows you to have an open passwordless wifi but it generates temporary encryption keys for each connected client, then the rest is as if it was using WPA, so that you don’t need to enter a password but your traffic gets encrypted and protected from anyone else listening in on the WiFi.

    If you combine both then you should have a network where each device is isolated both over the air and from a routing perspective so that each device only sees an Internet connection and no other devices.


  • The same way filebot and any other tool does - the file needs to have some label, either an absolute episode number or a season + episode number. I’m not aware of any tool that is able to look at the contents of the video to figure out which episode it is visually without any information from the filename - but I’d be happy to be proven wrong because I would be impressed.

    Sonarr/radarr does analyze the content somewhat but that’s just for gathering resolution, codec, HDR, audio languages, and subtitle information, which can all be added to the filename format for inclusion during renaming.


  • I second using sonarr/radarr, once imported it detects episodes and lets you one click rename to a specific format and folder organization.

    If you don’t want any of the other features of sonarr/radarr (like having a way to filter and manage your collection to see what’s in what quality or from what release group, searching multiple indexers with a single search, being able to send a specific search result to a downloader and have it automatically imported and organized when complete, or have auto downloading based on requests using scoring rules that you set), then there’s also filebot which a lot of people seem to like and seems to be just for matching with online metadata and renaming.

    But I haven’t tried filebot since I like the extra features and capabilities of sonarr/radarr. It makes it easy to manage several library folders like an archive for anything that’s been reviewed, is complete, and in a quality/codec that I’m satisfied with, and keeping track of currently airing shows in my active folder which is where I also keep auto downloaded stuff I haven’t reviewed.


  • I use a nuc10i7fnkn and since transcoding is almost entirely done using the dedicated quicksync hardware in the CPU you don’t end up actually using the CPU much. So I’m sure it would work on an older generation or the i5 version. I don’t know much about the N100 but it looks like it would be very capable. Supposedly it boosts to 3+GHz and it’s a 10nm node compared to my NUCs 14nm. But the GPU has the same number of execution units so I’m not sure if the quicksync transcoding performance is that different. I saw someone mention 3 simultaneous 4K transcodes and I think I got about that much on mine. Generally for quick sync performance you just compare the Intel hd or uhd graphics model (like 630, 730, uhd, etc) and the number of execution units and that should correlate to the performance. Also check the Wikipedia page for quicksync for codec compatibility (under the Hardware decoding and encoding section), but anything recent will handle most stuff you’d need: https://en.m.wikipedia.org/wiki/Intel_Quick_Sync_Video


  • I actually run my arrstack on a Synology, it has official support for docker and docker-compose. Granted I do have a higher powered model (the DS1621xs+) but most of the arrstack is fairly low power friendly.

    You can also get away with running Plex on a nas but I would only do it if 1. Your nas has a quick sync supported CPU and you get that enabled properly or 2. You go the direct streaming only / no transcoding setup - which means checking the codec support for all client devices and either only downloading exactly the supported codecs or pre-transcoding everything.

    What I do is actually run Plex/JF on a separate nuc and point it at the nas using a network mount. Just don’t use a network mount for the Plex app database (maybe same applies to JF too), just mount the media files itself. Running Plex and having it access the DB over a network mount is a big no no for various reasons.


  • I use a Synology nas which has official support for docker / docker-compose to run my arrstack and has n+2 btrfs redundancy. Then for running Plex and jellyfin I use an Intel nuc10i7 with quick sync with the nas media folder mounted over the network but using a direct gigabit link between the 2 so that the traffic stays off my switch.

    I could have gotten away with doing it all on the nas if I forewent ECC in favor of quick sync, but my first priority with my nas is keeping personal artifacts safe so I went with ECC.


  • This is super neat even though it’s basically just running the main board plugged into peripherals - I already knew the main board was pretty small but it’s still surprising seeing it sitting on a table.

    Makes me wonder what you could do with a case the size of a wii (or even a mini wii). Plenty of space for cooling, a 2.5" SSD bay, and I’m sure you could fit lots of other goodies in it too.


  • I’m on unrooted lineage with mindthegapps / Google play services with my Google Services Framework ID registered with Google, but I still have to make 3 attempts to log in to my bank with the first 2 attempts always giving a vague error like “we’re not sure why we couldn’t connect”, similar with fidelity. Using a password manager so I’m entering the same credentials every time.

    (Edit: in the case of fidelity, instead of faking a connection issue it tells me my account is blocked and to call support to unblock it - that’s also fake because I called once and they said my account wasn’t locked and trying to log in a second time always works)

    My understanding is that it’s impossible to pass strong integrity unless you’re using the stock unmodified rom with the bootloader locked.

    I changed banks last week and the new bank (Aspiration) logs in fine the first time every time.

    It sounds like the situation is better with graphene but I find it a lot easier to switch banks than roms.



  • What kind of TV service / set top box did you have at the time? I remember a lot of talk about providers pushing set top boxes both because it lets them use newer broadcast tech with customers using old TV tuners, but crucially it allows them to have their own software running on the box that you use to switch channels, which let them use out of band communication over the cable network to report what channels you watched, when, and how long, which I don’t doubt gets sold and aggregated by ad targeting firms.

    It’s pretty common for smart TVs to do a similar thing to collect streaming app watch data when using the TVs built in apps.