[gentoo-user] backup horror story (happy ending)

[ralfconn <mentadent47@yahoo.com>]

I have a 4Tb hard disk half full of videos and photos my daughter took 
with her cell phone over the years, shared with Win11 in a dual boot box 
so it is NTFS-formatted. The disk is backed up on a different EXT4 disk 
and the backup is performed by (ana)cron via an rsync bash script.

Last evening there was a power outage. When I rebooted in linux the NTFS 
disk would not mount. OK, just e2fsck the disk and it will fix it, I 
thought, forgetting that it was an NTFS not EXT. e2fsck -y starts 
finding and fixing hundreds of issues on the disk, till I get bored, I 
stop it and reboot into Win11, which chkdsk's it and mounts with no 
problem in less than 10s.

Finally I realize the huge mistake I had made, allowing e2fsck to delete 
thousands of otherwise fine clusters/nodes/whatever on a filesystem it 
does not understand.

But I have a backup, no problem... till I realize the cron job had 
already run so it had overwritten the old files with the new, corrupt 
versions.

Fortunately rsync uses the file access date to quickly find potential 
differences and since the e2fsck did not touch those the backup was 
still fine.

Later I found that the disk did not mount in linux due to mount not 
finding anymore the NTFS's UUID that I had in fstab, but it did mount 
fine with /dev/sdx.

I felt like a triple-idiot.

raf