Conversation

Charlotte lotteheartplural/Cinny cinny_heart_plural thetadelta ursaminor treblesand

root cause analysis of server failure: bit 30 is dead around address 0x3_0xxx_xxxx

1
1
2

i am now looping a memtest overnight to see if that range is the only broken one

1
0
0

given that i did not find anything in like 4h outside of the affected region i am guessing this is going to be stable: https://git.chir.rs/darkkirb/nixos-config/pulls/881/files

1
0
0

more failure analysis: one of the 3 disks is mcfuckingdead

2
0
0

raid masked the failure because it was that some data was dead uncorrectably on one drive but raid5 means that it can recover that no issue

0
0
0

by mcfuckingdead i mean some dead sectors

good news: seagate covers dead sectors under warranty
bad news: drive warranty expired exactly 10 days ago

1
0
1

Mx Rey 💜 🏳️‍⚧️

@charlotte Mr. Seagate is looking in your window and cackling with glee

0
0
0