What is the best way to save my linux (open media vault / debian) install from failing disk?

[jonossaseuraava@eliitin-some.fi]

@fd93
Booting from the source disk isnt an issue anyways. Ive got another computer with one of those usb sata cradles.

I'll give clonezilla a try.

[eal@shinonome.technology]

@jonossaseuraava @yonle set up etckeeper with automatic push on your next install (won't rescue these configs but you'll sleep better)

[phnt@fluffytail.org]

@cell @jonossaseuraava Clonezilla works in this case. You can create an image that is compressed and at maximum the size of used disk space. Only limitation is, on restoration the disk has the be at least the same size as the original.

https://clonezilla.org/

[phnt@fluffytail.org]

@cell @jonossaseuraava Oh also, for failing hard drives an advanced option called "rescue" has to be set, the "beginner" mode does not have it I think.

[skalli@dresden.network]

@jonossaseuraava

I would say rwsd and try (g)ddrescue

Its like dd but first safes the good sectors and later tries multiple times the bad ones...

It will probably try to do a full bir backup of the drive so try to get hand on another drive which can handle all stuff on the source Drive.

If the source drive has multiple partitions, try to safe them one by one.

Good luck.

[peteriskrisjanis@toot.lv]

@jonossaseuraava does data partition can be found and mounted? Sounds pretty dire

[jonossaseuraava@eliitin-some.fi]

@peteriskrisjanis
Yeah it can be found. Reading works most of the time - until a bad sector is hit or something.

[jonossaseuraava@eliitin-some.fi]

@yonle
If i knew enough to be spesific i wouldnt be asking for help

[jonossaseuraava@eliitin-some.fi]

@phnt
Too bad the target disk is smaller.

[phnt@fluffytail.org]

@jonossaseuraava Then you would have to make the partition smaller prior to the backup with something like gparted, but that is risky on failing disks and might involve reading and writing large amounts of data to the disk.

Do you have some intermediary disk where you could do this after making a backup first and restoring it to the intermediary before going on the target?

[phnt@fluffytail.org]

@jonossaseuraava
>Do you have some intermediary disk where you could do this after making a backup first and restoring it to the intermediary before going on the target?
You could technically do this in a VM since VM disks are usually only as big as the data used itself, but that is getting into voodoo territory.

[jonossaseuraava@eliitin-some.fi]

@eal @yonle

Nah, i need real backups. Users, symliks, installed aoftware and all.

I'll hooefully manage that during the lifetime of the new disk.

[hosnoopy@m.efg-ober-ramstadt.de]

@jonossaseuraava if dd doesnt work, the hdd is broken. Which filesystem is on the broken disk? btrfs? ext4?

[jonossaseuraava@eliitin-some.fi]

@HoSnoopy
ext4

[mkj@social.mkj.earth]

@jonossaseuraava Would definitely second (third, fourth) ddrescue if the target disk wasn't smaller. That would be exactly the use case it's designed for.

Best you can do is probably copy all files to the new disk. Recreate partitions, mount as the same hierarchy under e.g. /mnt, then something like

# rsync -aAHXSv --exclude=/sys --exclude=/proc --exclude=/dev --exclude=/mnt /./ /mnt/./

You will need to reinstall the boot loader on the new disk for bootability.

[cybercow@social.tchncs.de]

@jonossaseuraava My suggestion is to only copy /home/username anyway. and from there not everything. If you are lucky, the bad sector is somewhere else and you can copy your data without problems.

[hosnoopy@m.efg-ober-ramstadt.de]

@jonossaseuraava huff... I dont know. When hdd is broken... I think is impossible.

[keithduthie@mastodon.nz]

@fd93 @jonossaseuraava With advanced mode and the rescue option selected, because otherwise it'll still choke on the I/O errors.

[jonossaseuraava@eliitin-some.fi]

@cybercow
Homes are on separate raid array ans have almost nothing of value on them . luckily.