back online

the fix has been applied, verified and the nixos system flake no longer hard-depends on md0 coming up; this should no longer happen with disastrous recovery setiup needed in the future

server will be replaced in a cupboard and come back online soon

a reverted uuid change in the nixos hardware configuration has only shown itself to be erroneous due to the rebooting of the system during a power surge

this uuid will be updated, and the system will be rebuilt from another live environment

nixos, the server's linux distro, is unable to find the bulk data raid array at boot

however, it is visible and r/w-able in fedora thanks to livecd auto-assembling

even if a disk has died, the 3x8TB hdds are set up in a RAID 10 mdadm array, meaning one drive failure will result in data survival

🚨 server is down, and i know, and am fixing it

affected services: all
forgejo @ forge.joshwel.co
affected users: n/a
objectstore @ datashower.joshwel.co
affected users: me and ryan
immich @ i.scorchthinks.dev
affected users: ryan-side + cmm album r/w
jellyfin @ watch.scorchthinks.dev
affected users: me and ryan

hope to be back up by 1940 earliest, no upper bound on latest