I joined Lemmy back in 2020 and have been using it as qaz@lemmy.ml until somewhere in 2023 when I switched to lemmy.world. I’m interested in Linux, FOSS, and several other subjects.

  • 75 Posts
  • 536 Comments
Joined 1 year ago
cake
Cake day: June 10th, 2023

help-circle





  • qaz@lemmy.worldtoSelfhosted@lemmy.worldPost your setup. no matter how uggo
    link
    fedilink
    English
    arrow-up
    13
    arrow-down
    1
    ·
    edit-2
    12 hours ago

    Old setup:

    Lenovo ThinkCentre M900 that I bought refurbished for ~€130

    • i5-6500T (Passmark score 4792)
    • 8GB RAM
    • 512GB SATA SSD + 128GB SATA SSD (completely used for swap)
    • Buffalo DriveStation™ HD-WLU3 that I bought second hand for €10
    • 2 × 2TB SATA HDD’s in RAID 1
    • ~20W

    Old setup

    New setup:

    Custom build

    • ASUS Prime N100I-D D4 (Passmark score 5501) (~€100)
    • 16GB RAM - Crucial CT16G4SFRA32A (€28)
    • 512GB SATA SSD
    • 4 × 4TB SATA HDD’s in RAID 5 using mdadm (€160)
    • M.2 NVME to SATA 6x (ASM1116 for C-states) (€17)
    • 17.8W

    New setup

    (Not the Proliant Microserver Gen8 on top, the device below)

    The antennas are from a Sonoff Zigbee dongle and a bluetooth dongle for Home Assistant.

    I’ve mostly focused on power usage, price, and reliability since I’m a student and don’t want to spend a month’s worth of income on a “home lab”.

    It’s running the following:

    • Forgejo
    • Grafana
    • Home Assistant
    • Jellyfin
    • Kopia
    • Nginx-proxy-manager
    • Paperless NGX
    • Photoprism
    • Syncthing
    • TimescaleDB
    • Uptime-kuma
    • Vaultwarden: As backup
    • Watch Your LAN
    • Arr stack (currently disabled)
    • Homebox: Still up for testing, like it has been for the past couple months. It’s a great concept but the execution ain’t great (does anyone happen to know an alternative?)

    It’s using about 10% CPU and is running below 40°.








  • An N100 would be fine, I use it for my own server. Despite it being about as fast as an i5-6500T with a general benchmark, quicksync makes a big difference when encoding video with e.g. Jellyfin. I “upgraded” from a i5-6500T to a custom built N100 server and the performance improved a lot. However, if you plan on hosting game servers it probably won’t be enough.










  • It consisted of tensors weights, datasets (which can reach several gigabytes), images, 3d models, and roughly 250+ programming projects with binaries, git without LFS and also a lot build files.

    Nextcloud was able to sync it all, but syncing was getting so slow that I had to keep my new laptop running for almost an entire day to get all synced to it. It also wasn’t that great at excluding certain folders (like build cache folders or NPM package files), you would have to set up exclusions on each device separately. Another problem with Nextcloud sync was that it would sometimes duplicate projects after had been moved in a subfolder.