How to recover an ext4 partition with a bad superblock on a dubious 500GB internal hard drive?

Question

I bought this hard drive 6 months ago and it shouldn't be failing like this. It's a 500 GB sata III internal seagate that I have being using on an old ECS G31T-M9 computer with archlinux installed. I bought this disk brand new along with another 1 tera western digital also brand new for storing my personal archive and proJects, so this situation couldn't be worse. Both disks started failing at the same time and I can't tell if it is the system that is bad configured or the motherboard that's malcfuntioning either by propper malfunction or for bad forward compatibity with SATA III protocol (It was designed for SATA II i guess). Btw the ages old ide hard drive where the system is installed (and where nothing important is stored) is overwhermly healthy, so it must be related with SATA disks.

This two brand new disks randomly unmounted and were recognized or not wether by the OS or the motherboard on startup. Unplugging and plugging them in betwen boot ups helped for motherboard recognition.

This last time I've done this last procedure. Then as one of my partitions won't mount and the other wasn't listed by lsblk I removed all of the failing disks from the ECS and intalled one of them on a BIOSTAR P4M89-M7B computer. This computer is also old but recognizes the 500 GB hard disk along with the partition but not the ext4 filesystem, so I ran fsck on it.

Output was`bad superblock number ... please try fsck -b ... or fsck -b 32768 /dev/sda1

So I ran sudo fsck -n -b 32768 /dev/sda1 and succesfully detected the ext4 partition but with a bunch of errors.

then I ran dmesg|tail and I read a bunch of I/O errors coming from /dev/sda1, so I decided that the best was to make a backup of the partition before doing any changes, so I'm doing ddrescue of the hard drive, it's 59% complete on the first pass and fortunatly with no errors.

I know may be that this kind of development is candidate to removing on this forum but I'm overwhelmed by the situation and I can't ask a single question, so

What could be bad with my hard drives being that they're brand new?

What can I do if an fsck repair goes wrong? Is any way to undo it?

How can I mount this partition with an alternative superblock and is it safe?

How can I be certain if the disks are damaged or not?

Edit: Now a second ddreascue pass is completed with 99.99% rescued and only one error of 512 bytes. What does it means?

Answer 1

What could be bad with my hard drives being that they're brand new?

Did you purchase your hard drives simultaneously? They may have fallen to the ground.

What can I do if an fsck repair goes wrong? Is any way to undo it?

You can use your (second backup). I am not aware of undo function for file system checks.

How can I mount this partition with an alternative superblock and is it safe?

Mounting is safe in "read only" mode.

How can I be certain if the disks are damaged or not?

Never. As with testing for computer viruses a negative result does not prove anything. You might consider using backups.

Edit: Now a second ddreascue pass is completed with 99.99% rescued and only one error of 512 bytes. What does it means?

That is a good thing because ddrescue failed only on one sector when duplicating your drive. What is suspicious is that the error size is 512 bytes. As modern drives typically use a internal sector size of 4096 bytes I would have expected an error size of 4096 bytes. That leads me to the conclusion that your brand new drive maybe old. Luckily you have successfully hidden your drive model so this issue cannot be solved.