Very odd scary thing happened just now… I was doing some stuff on my computer and it suddenly hung in the way that I know is just classic case of the disk being disconnected or failing. So, uh oh. SysRq confirms the disk likely failed (after killing all tasks, the sync and emergency umount still failed). The system automatically warm restarted (since SysRq b reboots) and just says "no boot device found", and chainloading into the UEFI diagnostics said it can't see the hard drive. So yeah.
But then, fully shutting down, the diagnostics could see the hard drive again and it passed all the long-ass self-tests, so hmm.
I booted a live USB and the disk was visible fine, no issues could be found (cryptsetup repair & fsck on everything, mounting and chrooting in to check key files), but I noticed that efibootmgr claimed there were no boot entries for the internal drive other than the external ones (USB, etc.). So I booted Linux manually with an EFI shell and everything seemed to go fine.
After booting again, all installed files match their expected checksums according to paccheck, and the files in my home directory and /etc all match hashes from backups, so doesn't seem to be any disk corruption or other signs of failure. Re-adding the UEFI entries and rebooting multiple times has worked fine.
I think it seems like either the PCI controller or the hard drive's controller got into a bugged state and just needed a power cycle? (I was doing a lot of disk I/O with multiple programs at the time, so could've been overheating, or just hitting a firmware bug) And then the UEFI "helpfully" cleared all the temporarily invalid boot entries which made it look like complete drive failure
nytpu
npub138…r6wax
2026-01-29 01:34:30 UTC
Author Public Key
npub138lrje5y0ucetrpzg2mkuavq4e7esrln90tqsrszgy3hk9xncuequr6waxSeen on
wss://relay.momostr.pinkPublished at
2026-01-29 01:34:30 UTCEvent JSON
{
"id": "50b8557e6099bf1c1a769b8af6bbf631057901ebeb8805e596fe6eb6c6fe21c5",
"pubkey": "89fe3966847f31958c2242b76e7580ae7d980ff32bd6080e0241237b14d3c732",
"created_at": 1769650470,
"kind": 1,
"tags": [
[
"proxy",
"https://tilde.zone/@nytpu/115975813260665594",
"web"
],
[
"content-warning",
"Long"
],
[
"proxy",
"https://tilde.zone/users/nytpu/statuses/115975813260665594",
"activitypub"
],
[
"L",
"pink.momostr"
],
[
"l",
"pink.momostr.activitypub:https://tilde.zone/users/nytpu/statuses/115975813260665594",
"pink.momostr"
],
[
"-"
]
],
"content": "Very odd scary thing happened just now… I was doing some stuff on my computer and it suddenly hung in the way that I know is just classic case of the disk being disconnected or failing. So, uh oh. SysRq confirms the disk likely failed (after killing all tasks, the sync and emergency umount still failed). The system automatically warm restarted (since SysRq b reboots) and just says \"no boot device found\", and chainloading into the UEFI diagnostics said it can't see the hard drive. So yeah.\n\nBut then, fully shutting down, the diagnostics could see the hard drive again and it passed all the long-ass self-tests, so hmm.\n\nI booted a live USB and the disk was visible fine, no issues could be found (cryptsetup repair \u0026 fsck on everything, mounting and chrooting in to check key files), but I noticed that efibootmgr claimed there were no boot entries for the internal drive other than the external ones (USB, etc.). So I booted Linux manually with an EFI shell and everything seemed to go fine.\n\nAfter booting again, all installed files match their expected checksums according to paccheck, and the files in my home directory and /etc all match hashes from backups, so doesn't seem to be any disk corruption or other signs of failure. Re-adding the UEFI entries and rebooting multiple times has worked fine.\n\nI think it seems like either the PCI controller or the hard drive's controller got into a bugged state and just needed a power cycle? (I was doing a lot of disk I/O with multiple programs at the time, so could've been overheating, or just hitting a firmware bug) And then the UEFI \"helpfully\" cleared all the temporarily invalid boot entries which made it look like complete drive failure",
"sig": "371a1d294c29105fd229a4e4737b1ffd77f6e1e03f0ec10dac1a71a790639ee1663340dd7579f2194e7a1c77dfeb9fcd3fb9e465abf22fceaf576ce7602d4a8d"
}