On 1/18/24 03:47, gene heskett wrote:
On 1/18/24 03:57, David Christensen wrote:
The old /home RAID10 still has its metadata on disk. I would install
the "mdadm" package, edit /etc/fstab, copy and rework the old /home
line (new mount point, add option "ro"), create the mount point, and
mount.
I believe mdadm is already installed. At least enough to collect and
mount this raid10 and use it for /home for the last nearly 2 years.
I made the suggestion to install the "mdadm" package because I thought
you were going to do a fresh install of Debian.
Now after all this folderall, all 4 of the SSD's are reporting read
errors at very high lba's.
all 4 drives are reporting the same poh, 21027 hours for the occurence
of the error, that sounds like it could be just one crash or dirty power
down. In which case it s/b repairable
Do we have a repair utility that will force the drive to reallocate a
spare sector and fix those?
I have issued a smartctl -tlong on all 4 drives, results in about 3 hours.
A SMART long test should find and fix any read errors.
When deploying an SSD into a new role, I like to do a "secure erase"
followed by a SMART long test.
David