Skip to content

Latest commit

 

History

History
69 lines (49 loc) · 2.28 KB

DISK.md

File metadata and controls

69 lines (49 loc) · 2.28 KB
Raw

Disk Recovery Procedures

Prerequisites

dnf install smartmontools hdparm sysstat testdisk

Phase 0 - Analyze disk

Connect drive on SATA/SCSI directly (do not use USB adapters).

smartctl -a /dev/sdX

Phase 1 - Recover gross data as fast as possible

We disable standby, acoustics and APM. Then we enable fast error recovery (7 sec; only Enterprise and NAS drives support this).

hdparm -S 0 -M 254 -B 255 /dev/sdy
smartctl -l scterc,70,70 /dev/sdx
fdisk -xu /dev/sdx > part_sdx.log
ddrescue -f -n /dev/sdx /dev/sdy rescue_sdx.map

Phase 2 - Recover bad sectors if possible

We disable fast error recovery to try to read as much of the data as possible.

smartctl -l scterc,0,0 /dev/sdx
ddrescue -f -d r3 -W /dev/sdx /dev/sdy rescue_sdx.map

We leave it try some more.

ddrescue -f -d r100 -W /dev/sdx /dev/sdy rescue_sdx.map

If data is important, then manually test with -i.

Phase 3 - Sanitize

  1. Check capability for sanitize hdparm --sanitize-status /dev/sdx && hdparm -I /dev/sdX
  2. If disk supports it (Enterprise), then hdparm --sanitize-crypto-scramble /dev/sdX. Go to step 9.
  3. Otherwise, if SSD, then hdparm --sanitize-block-erase /dev/sdX
  4. Otherwise, if HDD, then hdparm --sanitize-overwrite --sanitize-overwrite-passes 1 hex:DEADBEEF /dev/sdX
  5. If santize is not supported at all, then unlock the drive (if frozen only; suspend computer or enable Hotplug in BIOS and remove ATA cable).
  6. Enter high security mode hdparm --user-master u --security-set-pass YOURPASS /dev/sdX
  7. If drive supports it, use enhanced security erase hdparm --user-master u --security-erase-enhanced YOURPASS /dev/sdX
  8. Otherwise, use normal security erase hdparm --user-master u --security-erase YOURPASS /dev/sdX
  9. For HDD, cryptsetup -c aes-xts-plain64 open /dev/sdX --type plain -d /dev/urandom --sector-size=4096 to_be_wiped. Otherwise go to 12.
  10. Run wipe dd if=/dev/zero of=/dev/mapper/to_be_wiped status=progress bs=1M conv=sync,noerror oflag=direct.
  11. Finish up cryptsetup close to_be_wiped.
  12. For SSD, blkdiscard /dev/sdX.

Phase 4 - Confirm if any deterioration or improvement.

smartctl -a /dev/sdX

Pending sector count increase/decrease. Relocated sector count increase/decrease. Offline irrecoverable errors increase/decrease.