Here is the thing, I have 4 RPi’s of different generations (all the way from Zero W to 4B 4GB) that I use to host services at home for personal use.

Lately, I have realized I am running out of RAM to host more services, not to mention not enough switch ports to connect to.

Now I know the obvious solution is to get a more powerful setup (maybe a thin client) but electricity isn’t cheap and I am not particularly in the best shape financially speaking to shell out $300+ on a decent client to host my services.

Any suggestions?

  • admin@lemmy.mohammadodeh.comOPEnglish
    1·
    11 months ago

    Can you please explain to me the difference? How does a swapdisk compare to RAM? I don’t mind googling it but I highly doubt I’ll get a straightforward ELI5 style answer from there.

    I would really appreciate it if you can elaborate, if you have the time that is.

    Thank you.

    • ryannathansEnglish
      2·
      11 months ago

      It just compresses your ram contents so you can fit more stuff in memory at once

    • bastion@feddit.nlEnglish
      2·
      11 months ago

      There are two types of computer memory that fundamentally matter on the consumer level:

      • memory, as in RAM, which is like short-term memory in people, or how much you can think about at once. It is blanked when the computer restarts. It is very fast.
      • drive storage, as in hdd or ssd space. This is like long-term memory or overall knowledge capacity, and persists across boots. It is comparatively slow.

      Solid state disk storage, and in particular some SD cards, can be vulnerable to excessive writes.

      Ram, however, is not impacted by the number of uses.

      A swap file works like this: When memory gets full, you move the least-used parts onto the swap file.

      A normal swapfile is on-disk. When memory gets close to full, the system moves some onto the (much much slower, like 10-1000x) on-disk swapfile.

      Zram swap creates a compressed swapfile out of your free memory. A file in linux does not have to be on a hard disk/ssd, it just has to look and quack like a file. When memory gets close to full, the system copies some onto the in-memory compressed file. This is very fast, but uses some cpu. It doesn’t touch your drive storage.