I just used Debian Testing to recover data from a set of failing SSDs.

With the added difficulty, the new one had to go back in its CentOS 6 host.

It turns out that ext4 defaults had been updated since the release of CentOS 6, and the kernel doesn't support ext4 with 64bit and R10(still not sure what this is).

Downloading the latest e2fsprog source, compiling it, and using it to drop the extra features and the journal, then with the system's
e2fsprogs add the journaling back and a short forced fsck.ext4 later all 500G where readily accessible again.

Show thread

wget e2fsprogs/github/release/v1.45.tar.gz
tar xzf v1.45.*.tar.gz
cd e2fsprogs*
./misc/tune2fs -O ^64bit /dev/mapper/vg-lv
./e2fsck/e2fsck -f /dev/mapper/vg-lv
./misc/resize -s /dev/mapper/vg-lv
./misc/tune2fs -O ^has_journal,^FEATURE_R10 /dev/mapper/vg-lv
tune2fs -j /dev/mapper/vg-lv
fsck.ext4 -f /dev/mapper/vg-lv

Show thread
Sign in to participate in the conversation

Linux fueled mayhem & madness with a side of news, reviews, and whatever the Hell-Elks™ we come up with.