| Current File : //proc/self/root/usr/local/jetapps/usr/share/rear/layout/recreate/default/150_wipe_disks.sh |
# Skip wiping disks when the user has explicitly specified to not wipe disks
# or when the user selected "Skip wiping disks" in layout/recreate/default/120_confirm_wipedisk_disks.sh
is_false "$DISKS_TO_BE_WIPED" && return 0
# The disks that will be wiped are those disks
# where in diskrestore.sh the create_disk_label function is called
# (the create_disk_label function calls "parted -s $disk mklabel $label")
# for those that exist as disks on the bare hardware for example like
# create_disk_label /dev/sda gpt
# create_disk_label /dev/sdb msdos
# create_disk_label /dev/md127 gpt
# so in this example by default DISKS_TO_BE_WIPED="/dev/sda /dev/sdb "
# or if the user has specified DISKS_TO_BE_WIPED use only the existing block devices therein
# cf. layout/recreate/default/120_confirm_wipedisk_disks.sh
# Log the currently existing block devices structure on the unchanged replacement hardware
# cf. how lsblk is called in layout/save/GNU/Linux/100_create_layout_file.sh
Log "Block devices structure on the unchanged replacement hardware before the disks $DISKS_TO_BE_WIPED will be wiped (lsblk):"
Log "$( lsblk -ipo NAME,KNAME,PKNAME,TRAN,TYPE,FSTYPE,SIZE,MOUNTPOINT || lsblk -io NAME,KNAME,FSTYPE,SIZE,MOUNTPOINT || lsblk -i || lsblk )"
local disk_to_be_wiped
# Wiping LVM storage objects:
# The only way to get rid of LVM stuff is to deconstruct the LVM stuff
# with LVM tools step by step from LVs to VGs to PVs.
# But deconstructing LVM stuff step by step from LVs to VGs to PVs gets impossible
# when there are LVs that belong to VGs that contain disks that should not be wiped,
# cf. https://github.com/rear/rear/pull/2514#issuecomment-729009274
# Currently only PVs are removed and the code to remove VGs and LVs is commented out for now
# because currently it is too complicated to autodetect to which disks VGs and LVs belong
# (VGs and LVs can be spread over various disks) to be able to avoid that VGs and LVs
# may get accidentally changed on disks that should not be touched at all,
# cf. https://github.com/rear/rear/pull/2514#issuecomment-743049696
# Plain wiping a disk with LVM storage objects leads to errors later when "parted -s $disk mklabel $label" is called
# in the diskrestore.sh script because then 'parted' fails (which lets diskrestore.sh fail which lets "rear recover" fail)
# with the following error message:
# Partitions ... have been written, but we have been unable to inform the kernel of the change,
# probably because it/they are in use. As a result, the old partition(s) will remain in use.
# You should probably reboot now before making further changes.
# In this case it works to reboot the recovery system and then a second "rear recover" usually succeeds
# because after plain wiping a disk with LVM storage objects the LVM metadata/signatures are no longer there
# so that LVM cannot be in use in the rebooted recovery system and "parted -s $disk mklabel $label" can succeed.
# But we like to do all in one single "rear recover" run so we need to clean up LVM storage objects properly.
# LVM storage objects must be properly detected and removed with LVM tools step by step from LVs to VGs to PVs:
local detected_LVs detected_VGs detected_PVs
# Detect LVs:
# The 'lvscan' output in the recovery system looks like:
# inactive '/dev/system/swap' [2.00 GiB] inherit
# inactive '/dev/system/home' [5.39 GiB] inherit
# inactive '/dev/system/root' [12.59 GiB] inherit
# cf. layout/recreate/default/README.wipe_disks
detected_LVs="$( lvscan -y | cut -s -d "'" -f2 | tr -s '[:space:]' ' ' )"
# Detect VGs:
# The 'vgscan' output in the recovery system looks like:
# Reading all physical volumes. This may take a while...
# Found volume group "system" using metadata type lvm2
# cf. layout/recreate/default/README.wipe_disks
detected_VGs="$( vgscan -y | grep 'Found volume group' |cut -s -d '"' -f2 | tr -s '[:space:]' ' ' )"
# Detect PVs:
# The 'pvscan' output in the recovery system looks like:
# PV /dev/mapper/luks1sda2 VG system lvm2 [19.98 GiB / 0 free]
# Total: 1 [19.98 GiB] / in use: 1 [19.98 GiB] / in no VG: 0 [0 ]
# cf. layout/recreate/default/README.wipe_disks
detected_PVs="$( pvscan -y | grep -o 'PV /dev/[^ ]*' | cut -s -d ' ' -f2 | tr -s '[:space:]' ' ' )"
# From the detected LVs VGs PVs remove only those that belong to one of the disks that will be wiped
# i.e. keep LVs VGs PVs on disks that will not be wiped (e.g. a separated storage disk with LVM on it).
local belongs_to_a_disk_to_be_wiped devices_belonging_to_disk
local detected_LV detected_VG detected_PV
# Remove detected LVs ('for' does nothing when $detected_LVs is empty):
#for detected_LV in $detected_LVs ; do
# belongs_to_a_disk_to_be_wiped="no"
# for disk_to_be_wiped in $DISKS_TO_BE_WIPED ; do
# # Get NAME (which could be a symlink like /dev/system/home for the LV 'home' in the 'system' VG)
# # and KNAME (which is the kernel device like /dev/dm-2 where the /dev/system/home symlink points to)
# # to be on the safe side with which name a detected LV is reported (usually the symlink is reported):
# devices_belonging_to_disk=( $( lsblk -nlpo NAME,KNAME $disk_to_be_wiped ) )
# # When NAME and KNAME are same it results duplicates in the devices_belonging_to_disk array
# # but duplicates do not matter for IsInArray (see the IsInArray implementation in lib/array-functions.sh):
# IsInArray $detected_LV "${devices_belonging_to_disk[@]}" && belongs_to_a_disk_to_be_wiped="yes"
# done
# if ! is_true $belongs_to_a_disk_to_be_wiped ; then
# DebugPrint "Skip removing LVM logical volume $detected_LV (not recognized as belonging to a disk to be wiped)"
# continue
# fi
# if lvremove -f -y $detected_LV ; then
# DebugPrint "Removed LVM logical volume $detected_LV"
# else
# LogPrintError "Failed to remove LVM logical volume $detected_LV ('lvremove $detected_LV' failed)"
# fi
#done
# Remove detected VGs ('for' does nothing when $detected_VGs is empty):
#for detected_VG in $detected_VGs ; do
# if vgremove -ff -y $detected_VG ; then
# DebugPrint "Removed LVM volume group $detected_VG"
# else
# LogPrintError "Failed to remove LVM volume group $detected_VG ('vgremove $detected_VG' failed)"
# fi
#done
# Remove detected PVs ('for' does nothing when $detected_PVs is empty):
for detected_PV in $detected_PVs ; do
belongs_to_a_disk_to_be_wiped="no"
for disk_to_be_wiped in $DISKS_TO_BE_WIPED ; do
# Get NAME (which could be a symlink e.g. /dev/mapper/luks1sda2 for a PV that is a LUKS volume)
# and KNAME (which is the kernel device like /dev/dm-0 where /dev/mapper/luks1sda2 points to)
# to be on the safe side with which name a detected PV is reported
# (usually the symlink is reported when such a symlink exists):
devices_belonging_to_disk=( $( lsblk -nlpo NAME,KNAME $disk_to_be_wiped ) )
# When NAME and KNAME are same it results duplicates in the devices_belonging_to_disk array
# but duplicates do not matter for IsInArray (see the IsInArray implementation in lib/array-functions.sh):
IsInArray $detected_PV "${devices_belonging_to_disk[@]}" && belongs_to_a_disk_to_be_wiped="yes"
done
if ! is_true $belongs_to_a_disk_to_be_wiped ; then
DebugPrint "Skip removing LVM physical volume $detected_PV (not recognized as belonging to a disk to be wiped)"
continue
fi
# Example how things fail when we try to cleanly remove a PV that is a LUKS volume
# in the ReaR recovery system and what worked in the end (at least on SLES 15 SP1):
# RESCUE # pvscan
# No matching physical volumes found
# RESCUE # cryptsetup luksOpen /dev/sda2 luks1sda2
# Enter passphrase for /dev/sda2: ...
# RESCUE # pvscan
# PV /dev/mapper/luks1sda2 VG system lvm2 [19.98 GiB / 0 free]
# Total: 1 [19.98 GiB] / in use: 1 [19.98 GiB] / in no VG: 0 [0 ]
# RESCUE # pvremove -f -y /dev/mapper/luks1sda2
# PV /dev/mapper/luks1sda2 is used by VG system so please use vgreduce first.
# (If you are certain you need pvremove, then confirm by using --force twice.)
# /dev/mapper/luks1sda2: physical volume label not removed.
# RESCUE # vgreduce -f system /dev/mapper/luks1sda2
# Physical volume "/dev/mapper/luks1sda2" still in use
# RESCUE # pvremove -f -f -y /dev/mapper/luks1sda2
# WARNING: PV /dev/mapper/luks1sda2 is used by VG system.
# WARNING: Wiping physical volume label from /dev/mapper/luks1sda2 of volume group "system".
# Labels on physical volume "/dev/mapper/luks1sda2" successfully wiped.
# RESCUE # pvscan
# No matching physical volumes found
# So all we can do is using force twice because only that works (with warnings):
if pvremove --force --force -y $detected_PV ; then
DebugPrint "Removed LVM physical volume $detected_PV"
else
LogPrintError "Failed to remove LVM physical volume $detected_PV ('pvremove $detected_PV' failed)"
fi
done
# Wipe RAID plus LVM plus LUKS metadata.
# To wipe RAID Superblocks it is sufficient to wipe 133 KiB at the beginning and at the end of the device.
# To wipe to wipe LVM metadata is should be sufficient to wipe 4 MiB at the beginning and at the end of the device.
# To wipe LUKS headers is should be sufficient to wipe 8 MiB at the beginning of the device.
# To wipe RAID superblocks plus LVM metadata plus LUKS headers it should be sufficient to
# wipe 8 MiB + 4 MiB + 1 MiB = 13 MiB at the beginning of the device and to
# wipe 4 MiB + 1 MiB = 5 MiB at the end of the device.
# To be future proof (perhaps LUKS may add a backup header at the end of the device)
# wiping 16 MiB at the beginning and at the end of the device should be sufficiently safe.
local children_to_be_wiped device_to_be_wiped device_to_be_wiped_size_bytes bytes_to_be_wiped dd_seek_byte
# 16 * 1024 * 1024 = 16777216
local bytes_of_16_MiB=16777216
for disk_to_be_wiped in $DISKS_TO_BE_WIPED ; do
# Get all child kernel devices of the disk_to_be_wiped in reverse ordering
# so that nested children/grandchildren are first, then children and the disk_to_be_wiped last
# which is the required ordering to wipe devices (from lowest nested up to the top level device).
# For example when "lsblk -ipo NAME,KNAME,PKNAME,TYPE,FSTYPE,SIZE,MOUNTPOINT /dev/sda" shows
# NAME KNAME PKNAME TYPE FSTYPE SIZE MOUNTPOINT
# /dev/sda /dev/sda disk 20G
# |-/dev/sda1 /dev/sda1 /dev/sda part 8M
# `-/dev/sda2 /dev/sda2 /dev/sda part crypto_LUKS 20G
# `-/dev/mapper/cr_ata-QEMU_HARDDISK_QM00001-part2 /dev/dm-0 /dev/sda2 crypt LVM2_member 20G
# |-/dev/mapper/system-swap /dev/dm-1 /dev/dm-0 lvm swap 2G [SWAP]
# |-/dev/mapper/system-root /dev/dm-2 /dev/dm-0 lvm btrfs 12.6G /
# `-/dev/mapper/system-home /dev/dm-3 /dev/dm-0 lvm xfs 5.4G /home
# then "lsblk -nlpo KNAME /dev/sda | tac" shows
# /dev/dm-3
# /dev/dm-2
# /dev/dm-1
# /dev/dm-0
# /dev/sda2
# /dev/sda1
# /dev/sda
children_to_be_wiped="$( lsblk -nlpo KNAME $disk_to_be_wiped | tac | tr -s '[:space:]' ' ' )"
# "lsblk -nlpo KNAME $disk_to_be_wiped" does not work on SLES11-SP4 which is no longer supported since ReaR 2.6
# but it works at least on SLES12-SP5 so we test that children_to_be_wiped is not empty to avoid issues on older systems:
if ! test "$children_to_be_wiped" ; then
LogPrintError "Skip wiping $disk_to_be_wiped (no output for 'lsblk -nlpo KNAME $disk_to_be_wiped' or failed)"
continue
fi
LogPrint "Wiping child devices of $disk_to_be_wiped in reverse ordering: $children_to_be_wiped"
for device_to_be_wiped in $children_to_be_wiped ; do
if ! test -b $device_to_be_wiped ; then
LogPrintError "Skip wiping $device_to_be_wiped (no block device)"
continue
fi
# Get the size of the device in bytes which could be smaller than 16 MiB.
# For example a 'bios_grub' partition could be smaller (e.g only 8 MiB on SUSE systems)
# cf. the "lsblk -ipo NAME,KNAME,PKNAME,TYPE,FSTYPE,SIZE,MOUNTPOINT /dev/sda" output above.
device_to_be_wiped_size_bytes="$( lsblk -dbnlpo SIZE $device_to_be_wiped )"
if ! test -b $device_to_be_wiped_size_byte ; then
LogPrintError "Skip wiping $device_to_be_wiped (no output for 'lsblk -dbnlpo SIZE $device_to_be_wiped' or failed)"
continue
fi
# By default wipe 16 MiB at the beginning and at the end of the device:
bytes_to_be_wiped=$bytes_of_16_MiB
# Wipe at most the size of the device in bytes:
test $device_to_be_wiped_size_bytes -lt $bytes_to_be_wiped && bytes_to_be_wiped=$device_to_be_wiped_size_bytes
# Wipe at the beginning of the device:
if ! dd bs=1M if=/dev/zero of=$device_to_be_wiped count=$bytes_to_be_wiped conv=notrunc,fsync iflag=count_bytes ; then
LogPrintError "Failed to wipe first $bytes_to_be_wiped bytes of $device_to_be_wiped ('dd if=/dev/zero of=$device_to_be_wiped count=$bytes_to_be_wiped iflag=count_bytes' failed)"
else
DebugPrint "Wiped first $bytes_to_be_wiped bytes of $device_to_be_wiped"
fi
# Wipe at the end of the device:
if ! test $device_to_be_wiped_size_bytes -gt $bytes_to_be_wiped ; then
DebugPrint "Skip wiping at the end of $device_to_be_wiped (dvice size $device_to_be_wiped_size_bytes not greater than the bytes that were wiped)"
continue
fi
# The byte whereto dd should seek to wipe to the end of the device from that point:
dd_seek_byte=$(( device_to_be_wiped_size_bytes - bytes_to_be_wiped ))
if ! test $dd_seek_byte -gt 0 ; then
DebugPrint "Skip wiping at the end of $device_to_be_wiped (dd seek byte would be $dd_seek_byte)"
continue
fi
if ! dd bs=1M if=/dev/zero of=$device_to_be_wiped count=$bytes_to_be_wiped seek=$dd_seek_byte conv=notrunc,fsync iflag=count_bytes oflag=seek_bytes ; then
LogPrintError "Failed to wipe last $bytes_to_be_wiped bytes of $device_to_be_wiped ('dd if=/dev/zero of=$device_to_be_wiped count=$bytes_to_be_wiped seek=$dd_seek_byte iflag=count_bytes oflag=seek_bytes' failed)"
else
DebugPrint "Wiped last $bytes_to_be_wiped bytes of $device_to_be_wiped"
fi
done
done
# Log the still existing block devices structure after the disks in DISKS_TO_BE_WIPED were wiped:
Log "Remaining block devices structure after the disks $DISKS_TO_BE_WIPED were wiped (lsblk):"
Log "$( lsblk -ipo NAME,KNAME,PKNAME,TRAN,TYPE,FSTYPE,SIZE,MOUNTPOINT || lsblk -io NAME,KNAME,FSTYPE,SIZE,MOUNTPOINT || lsblk -i || lsblk )"