I got unreadble sectors on the SSD disk of the brand new netbook I
just bought (after after an identical one had been stolen, yeah).
Of course this happened just after I spent some days configuring all
the bells and whistles, saying "I'll backup right after I've
finished".
The partition was cleanly unmounted, and on reboot I got a Grub error
17 (it was mounted /)
After booting with I USB dongle dumped the partition on an external drive with
ddrescue -r5 /dev/sda1 /media/sda1dumpddrescue2 /media/sda1dumpddrescue2.log
the logfile of ddrescue reads:
# Rescue Logfile. Created by GNU ddrescue version 1.2
# pos size status
0x00000000 0x00001000 -
0x00001000 0xEF3F4200 +
So it looks like only the very beginning of the partition is
unreadable; from the logfile it would be 1000B, but actually ddrescue
says
Initial status (read from logfile)
rescued: 4013 MB, errsize: 4096 B, errors: 8
Anyway, I tried to repair the dump with xfs_repair -f, the log is attached.
The fs is successfully repaired, but find lost+found -type c | wc -l
count 54000+ regular files lost.
Is this normal and consistent with the hypothesis that just the first
1K or 4K of the partition be lost? Can I do something to
reconnectomore inodes?
I still have the unrepaired dump to trick with.
Thanks in advance,
Luca Niccoli
xfs_repair.log
Description: Binary data
|