I like tarsnap.com
Did you back up today user?
Chances of a no warning mechanical failure are super small.
Requires off-site backup. Which is gonna be le Cloud for personal use, something botnet noided Zig Forums users won't have.
Use a non-shit OS.
Other reasons
git or other version control
Legit good reason.
Either highly unlikely or use a non-shit OS.
So we basically have backups for virus or you are using a shit OS. And this is why I backup my windows 7 data once a month and my gnu/linux data once a week.
Use tarsnap, locally encrypted, open source, cheapish, etc.
Or use any other storage provider, because you can do local encryption either way.
A smart rabbit should have three burrows.
Btrfs. It doesn't track bad blocks and will try to reuse them.
The problem is in how hard disk firmware handles bad sectors. From what I understand is that there is a reserved zone on the platter that the firmware will remap bad sectors to when they fail.
A bad sector is detected on a read when the ECC fails to recover. The firmware reports the content as zeros and remaps that sector to that reserved area for the next read. Btrfs is actually very good in this case because it checksums all blocks and can recover during a scrub if you are using RAID1 (btrfs RAID1, not mdraid).
When the disk gets enough of these bad sectors that the reserved area is depleted then those bad blocks are not abstracted away and the file system is aware of them. This is where btrfs fails. Ext2+ filesystems, being designed before hard drives did this, has built in bad block remapping so that is skips over known bad blocks. Btrfs lacks this ability.
I read one suggestion to find the bad sector and make partitions before and after it for btrfs to span across.
The real solution is to buy a new disk when this happens.
not unless you kill her
trust nobody
I'll bet that's what people used to say about electricity two hundred years ago, or landline phones a century ago.
Everything we've ever built lasts forever, or until it's deprecated by newer technology.