tl;dr Surprise maintenance! It looks like it is once again time for some fresh hardware. Within the next hours (perhaps also tomorrow, depending on time among other things), you might experience a downtime of up to one to two hours due to Lemmy being moved to a new machine.

This is needed for more future proofing as well as making sure the current hardware is not about to die and forcing us to apply a backup.

This post will be updated as soon as there are any updates.

Old specs:

  • i7-7700 CPU @ 3.60GHz
  • 2x512 GB KXG50ZNV512G TOSHIBA Gen 3 in Software Raid
  • 4x RAM 16384 MB DDR4

Expected new specs:

  • likely some AMD 7
  • about 2x1 TB Gen 4 in Software Raid
  • 64 or 128 GB RAM

If in your budget and if you want to support this instance, feel welcome to donate at tchncs.de/donate :)

Right now. both of the servers drives report 255% used. I have already received this machine in the unluckly state of equally used NVMe’s… and there is some promising iowait going on, making me worry.

It’s not just the very unfortunate and (as i had to learn by actually losing drives upon reboot) dangerous state of the drives, but also the fact that the i7-7700 seems a bit overwhelmed with the load from Lemmy and Firefish.

Another thing I am not (yet) sure about is whether the 64G memory are actually sufficient…

Total virtual memory: 61.74% - 38.6 GiB of 62.6 GiB
RAM: 61.74% - 38.6 GiB of 62.6 GiB
Committed: 65.29% - 40.9 GiB of 62.6 GiB virtual memory
Commit Limit: 50.00% - 31.3 GiB of 62.6 GiB virtual memory
Shared memory: 17.05% - 10.7 GiB of 62.6 GiB RAM
Page tables: 3.15% - 1.97 GiB of 62.6 GiB RAM
Disk Writeback: <0.01% - 2.42 MiB of 62.6 GiB RAM
RAM available: 20.10% free - 12.6 GiB of 62.6 GiB
Hardware Corrupted: 0% - 0 B of 62.6 GiB RAM

This is fine’ish but not sure for how long. Maybe for now I wait with upgrading the memory… – on the other hand… why not give some room for even more database tweaking perhaps?