sj_zero on Nostr: [admin mode] So I went to my data center, and swapped out the ssds for the new ssds ...
[admin mode] So I went to my data center, and swapped out the ssds for the new ssds that I'm replacing them with, and then I decided to go in and spend some time working on the main server that serves everything.
After changing a couple settings, I noticed that I was stuck in a reboot loop on startup. I figured that the cause was the USB hard drive I use for Mass storage, so I fiddled around with that for a bit but eventually I came to realize the issue was that I turned on hyperthreading.
I started playing with the idea that maybe the CPU is getting starved of power by having the extra load. And you know what? Once I disabled turbo, I was able to complete a task that previously caused the server to crash (systemwide rsync backup). I guess the 13-year-old processor just doesn't want a turbo like it used to. Thanks Obama!
There's good news for everyone but me. The good news is that the site is stable again, and the othe good news is that my servers are both set up for proxmox and I'm starting to work on clustering and the like so fbxl social is rock solid stable. And the downside is that I'll have to set up matrix again which is just a pain in the butt.
With everything working properly again for now, I can slow down and make sure I do this right with proper separation of roles and functions to ensure I'm using better practices as well as doing High Availability properly. Pretty interested in seeing how QUIC support plays out.
I'll keep everyone posted either way.
After changing a couple settings, I noticed that I was stuck in a reboot loop on startup. I figured that the cause was the USB hard drive I use for Mass storage, so I fiddled around with that for a bit but eventually I came to realize the issue was that I turned on hyperthreading.
I started playing with the idea that maybe the CPU is getting starved of power by having the extra load. And you know what? Once I disabled turbo, I was able to complete a task that previously caused the server to crash (systemwide rsync backup). I guess the 13-year-old processor just doesn't want a turbo like it used to. Thanks Obama!
There's good news for everyone but me. The good news is that the site is stable again, and the othe good news is that my servers are both set up for proxmox and I'm starting to work on clustering and the like so fbxl social is rock solid stable. And the downside is that I'll have to set up matrix again which is just a pain in the butt.
With everything working properly again for now, I can slow down and make sure I do this right with proper separation of roles and functions to ensure I'm using better practices as well as doing High Availability properly. Pretty interested in seeing how QUIC support plays out.
I'll keep everyone posted either way.