I want a server running nextcloud, immich and others.

I have a N100 mini server with a 2TB external HDD. I want to secure the system against data loss. Hence, I want a backup and redundancy.

  1. Most important question: How do I build everything? Is this a NAS? My naive approach is to buy 3 external HDDs and connect them to the N100 with a USB hub. I assume this is not “the right way” but to use/build a NAS. Do I have to build a separate NAS computer? When I lookup NAS buying, it is a computer with a case for 4 drives, excluding the drives and costs 400 bucks. I am confused because this is incredibly expensive compared to what I already have. What is the additional benefit compared to my setup? Am I cheap?

  2. Regarding redundancy, is RAID still the way to go? At 2 TB, using RAID 5 with 3 drives sounds good. I’d have 4 TB of usable space, much more than I intend to use in the next years, and adding a drive increases the storage by 2 TB, effectively increasing space by 50%.

  3. I have 4 TB usable space, but I won’t reach 2 TB in the next one or two years. I’d use a 2 TB HDD for a local backup via borg. Once my hot storage needs to increase, I replace the backup drive with a larger one and use it to increase the RAID storage. Is one backup sufficient? Or should I keeping multiple versions of the data. Daily, weekly, monthly backups? What is your experience with it?

  4. Another 2 TB HDD for an offsite backup, LUKS encrypted, backed up once a year (that’s the goal for now).

Does that sound good?

  • Onomatopoeia@lemmy.cafeEnglish
    18·
    3 days ago

    RAID isn’t data redundancy, it’s an array of drives combined to form a single logical storage pool. It solves the problem of needing a single storage pool larger than the available drives. As such, it’s very sensitive to loss of a single drive.

    At your storage size requirements (2 TB), RAID is unnecessary today.

    Edit: Let me say it again for you downvoters-RAID is NOT data redundancy.

    There is only ONE copy of your data in RAID (excepting mirroring). It’s why RAID now has double parity and hot spare drive capability.

    RAID is for creating a single pool that’s larger than available drive size.

    Go ahead and downvote in ignorance, and learn about data redundancy when your RAID fails.

    RAID is NOT data redundancy - it’s DRIVE redundancy.

    Take it from the source https://www2.eecs.berkeley.edu/Pubs/TechRpts/1987/CSD-87-391.html

    • talkingpumpkin@lemmy.worldEnglish
      15·
      4 days ago

      RAID (except RAID0) is data redundancy, it just isn’t backup (ie. it doesn’t help if you accidentally delete stuff, or if some bug corrupts it, or if you drop the computer while moving it).

    • ExcessShiv@lemmy.dbzer0.comEnglish
      9·
      3 days ago

      You’re using redundancy and backup synonymously, but they’re not. Raid 1+ absolutely provides redundancy, you are 100% wrong in saying that it doesn’t, because it provides a failover system that prevents operational interruption if a drive fails.

      • Onomatopoeia@lemmy.cafeEnglish
        23·
        3 days ago

        That’s not data redundancy - there’s still only one copy of your data.

        Those are mitigations against loss of data due to loss of parity.

        There’s still only ONE copy of your data.

        • Courant d'air 🍃@jlai.luEnglish
          5·
          3 days ago

          I think you’re confusing backup and redundancy. While I totally agree RAID1+ should not be considered backup, it absolutely is redundancy, as the same data is present on at least two disks (either on the form of the exact same data or something that can be used to rebuild the missing data).