• Shurimal
    link
    fedilink
    986 months ago

    Probably it doesn’t quite count as a gadget, but repurposing my old PC as a home server. Firstly it makes a great mass storage solution making all my media accessible from any device, no matter what architecture it is and what apps it can run. I also self-host Home Assistant, Syncthing, Radicale, Navidrome, Jellyfin and UrBackup. The ten years old 2 core Pentium with 8GB of RAM can do it all, it’s much cheaper to run than half a dozen subscription services and I have total control over my data and privacy.

    • @[email protected]
      link
      fedilink
      26 months ago

      If you don’t mind, which processor do you have? I’ve been thinking of setting up a Jellyfin server too, but I have a G4500 and I’ve always been worried that it can’t handle the load…

      • Shurimal
        link
        fedilink
        36 months ago

        Mine is the venerable G3258—the budget overclocking champ in the 4th generation Core family. Runs at 4,4 GHz and handles OpenMediaVault and 19 Docker containers just fine. I think G4500 would be fine, too.

        • bruhduh
          link
          16 months ago

          deleted by creator

          • Shurimal
            link
            fedilink
            16 months ago

            True, the old server parts are dirt cheap if you can source them (theyr’e not really available in my country and importing from China, UK or US would more than double the cost). But they’re also quite power hungry and energy cost has gone up crazy over the past few years. My current setup consumes around 127 W total (overclocked CPU, 3 HDD-s) and it costs me around 20€/month or half my energy bill (small one bedroom apartment with full LED lighting). If I upgrade it’ll probably be cheap current PC hardware which tends to be much more power efficient.

            My dream setup is a 16 core ARM CPU with something like 64…128 GB of LPDDR🫠

      • @ilinamorato
        link
        16 months ago

        As I understand it, media streaming isn’t actually that taxing because your server doesn’t actually have to render all of that data, just transfer it; so as long as it can handle a copy operation faster than one second per second, and you’re only watching from one device at a time, it’ll still work.

        I haven’t done it, though, so I’m not sure how much overhead Plex/Jellyfin add by way of transcoding.

    • @[email protected]
      link
      fedilink
      26 months ago

      wow that’s amazing. so it’s connected to all other PCs in the house? did you have to buy a lot of new storage?

      • Shurimal
        link
        fedilink
        16 months ago

        I actually bought just one new 6TB HDD and repurposed an older 3TB one as a redundancy drive for mirroring most critical data using a simple rsync cron job (no need for realtime mirroring of media files that are write-once), plus another old 1 TB drive just because. I haven’t run out of storage yet and I have automated download/sharing for OpenStreetMap and some Linux distros which takes up half a TB or so, but I plan on expanding the array using MergerFS and SnapRAID when the need arises.

        The rest is just SMB shares, Navidrome, Jellyfin, DLNA and FTP. Remote access from outside my local network is done via Tailscale VPN.

        • @[email protected]
          link
          fedilink
          16 months ago

          What benefits do you see in navidrome compared to having your music in Jellyfin? I’m just starting out with jellyfin and added some music to it. I listen to it with findroid on my phone and so far it seems to work okay.

          • Shurimal
            link
            fedilink
            16 months ago

            Navidrome just seems to be faster and more responsive. But the main reason of using both is that I just like to try things out and tinker. I also use Foobar2000, Kodi, MPC-HD, AIMP and other media players.

    • @[email protected]
      link
      fedilink
      16 months ago

      I recently picked up a 13 year old dell inspiron to run my instance of home assistant and Plex. It was an upgrade from a shitty old Linux laptop that was literally falling apart. All I had to do was add ram (it only had 6gb and it wasn’t stable, so I maxed it out with 16gb) and I swapped the old slow HDD for a crucial SATA SSD and it’s been perfect. It probably pulls more wattage than necessary but it’s exactly what I need for now.

    • @protput
      link
      16 months ago

      I bought a cheap low power minipc. Don’t know the numbers but having a 10yo desktop powered 27/7 can’t be that great for your power consumption.

      The one I bought is an Intel Alder Lake N100 Quad Core up to 3.4GHZ, 16GB DDR5 512GB for €160.

    • @[email protected]
      link
      fedilink
      16 months ago

      What benefits do you see in navidrome compared to having your music in Jellyfin? I’m just starting out with jellyfin and added some music to it. I listen to it with findroid on my phone and so far it seems to work okay.

      • @[email protected]
        link
        fedilink
        16 months ago

        I’m not the guy you replied to.

        I originally stored my music in Plex and used Plexamp. I have a large playlist downloaded from youtube which caused horrible performance issues in Plexamp. Navidrome is pretty much a read-only service. It can only read metadata from the files, not add any or manage them. For me this feels safer to expose to the internet since my docker container only has read-only access to all of my files. Even if someone broke into the service for some reason, they couldn’t do anything to my files.

        I don’t know if jellyfin has similar performance issues with large playlists since I already had navidrome set up by then.

        • @[email protected]
          link
          fedilink
          16 months ago

          Thanks! I don’t have too much music on it yet, I guess, so not sure on the performance. I do like that read only approach, though. Currently I’m running just the regular jellyfin app on my Mac. What made you use it in docker? It sounds like in Linux it’s a safeguard to prevent dependency issues but I don’t think that’s really a factor on mac

          • @[email protected]
            link
            fedilink
            16 months ago

            Mostly ease of management. I have a server on which I run multiple applications. If I don’t need something anymore, I can just purge the container. The directories used by that container are clearly listed in my docker-compose file so I never have to wonder whether I purged everything that is now unnecessary.

            It also makes it very easy to deploy a new service.