| 2 Apr 2024 |
 K900 | Which one of them is lying? | 07:17:47 |
| K900 | I don't know | 07:17:48 |
 ElvishJerricco | wait so literally nothing exhibits a problem except an EIO on fsync? | 07:18:04 |
 K900 broke the homeserver | Yes | 07:18:09 |
| ElvishJerricco | ouch | 07:18:14 |
| K900 broke the homeserver | At least as far as I can see | 07:18:16 |
| ElvishJerricco | I don't feel safe | 07:18:18 |
| K900 broke the homeserver | Exactly | 07:18:30 |
 hexa | I tell people how I feel about btrfs, but they also tend not to listen | 07:18:46 |
| K900 broke the homeserver | I am starting to doubt my "btrfs is generally OK" position very hard rn | 07:19:09 |
| hexa | look, there is certainly a custom tool somewhere on github to debug your exact issue | 07:19:24 |
| hexa | join #btrfs on libera and ask for it 😛 | 07:19:31 |
| K900 broke the homeserver | Failing in UNCLEAR ways is the worst thing a filesystem can do | 07:19:32 |
| ElvishJerricco | this isn't actually important but I don't like that "K900 broke the homeserver" appeared with the same text color as my user right after I asked a question :P | 07:19:36 |
| K900 broke the homeserver | I'm taking down the 0upti.me box so I can try and offline check it | 07:20:04 |
| K900 broke the homeserver | btrfs check is clean | 07:23:05 |
| K900 broke the homeserver | I DO NOT LIKE THIS. | 07:23:13 |
| K900 broke the homeserver | So uh is bcachefs migrate still broken | 07:23:53 |
| ElvishJerricco | yea, the fact that nix thinks everything is fine indicates a more innocuous FS bug in btrfs | 07:24:13 |
 raitobezarius | In reply to @hexa:lossy.network
I tell people how I feel about btrfs, but they also tend not to listen As we all do | 07:24:15 |
| ElvishJerricco | still scary | 07:24:19 |
| ElvishJerricco | but I bet it's fine | 07:24:22 |
| ElvishJerricco | (it's not my data; I am obviously less incentivized to care about it) | 07:25:15 |
| K900 broke the homeserver | It's not even that important data | 07:25:27 |
| K900 broke the homeserver | Realistically | 07:25:49 |
| raitobezarius | An EIO should have minimally generated a trace in the kernel log | 07:26:02 |
| raitobezarius | It's really disconcerting | 07:26:06 |
| hexa | agreed | 07:26:11 |
| K900 broke the homeserver | Running another scrub from the rescue image | 07:26:44 |
| K900 broke the homeserver | Just to make sure I'm not going insane | 07:26:47 |
