Project home | Documentation index | IEEE references
Before changing storage, save the output of:
lsblk -e7 -o NAME,PATH,TYPE,SIZE,FSTYPE,FSVER,LABEL,UUID,PARTUUID,MOUNTPOINTS
findmnt -o TARGET,SOURCE,FSTYPE,VFS-OPTIONS,FS-OPTIONS
sudo blkid
sudo cryptsetup luksDump /dev/disk/by-id/<existing-luks-device>Also record firmware settings, boot mode, partition layout, LUKS UUIDs, enrollment method, filesystem UUIDs, mount points, swap or resume target, and the commands needed to rebuild the initramfs and boot entries.
Choose one complete stack from the decision matrix. Do not combine COW snapshots, external RAID, multiple encryption systems, and exotic tuning without assigning ownership for recovery at every layer.
Restore representative files and metadata to a separate location. Verify that the backup includes permissions, ACLs, extended attributes, sparse files, hard links, database-consistent exports, and encryption recovery material where applicable.
Use /dev/disk/by-id/ and verify model, serial, size, partitions, and active
mounts immediately before partitioning, luksFormat, or mkfs. Unmount and
close only the exact target. Do not use a broad wildcard.
Use the LUKS2 baseline and exactly one
filesystem profile. Do not add --force, -F, or
equivalent safety-bypass flags to generic instructions. After creation, obtain
identifiers with:
lsblk -f
sudo blkid
sudo cryptsetup luksUUID /dev/disk/by-id/<exact-device>-part2On a maintenance boot or otherwise safe window:
findmnt --verify --verbose
sudo mount --all --verbose
findmnt -o TARGET,SOURCE,FSTYPE,VFS-OPTIONS,FS-OPTIONS
systemctl --failedWrite, sync, read, and delete a non-secret test file. Confirm ownership and security flags as the intended user. If the volume is removable, test absence at boot. If hibernation is required, test resume after encryption unlock. If a token unlock is used, also test the recovery credential and rescue media.
Benchmark a disposable file on the mounted, encrypted filesystem, not the raw device. Use representative block sizes, queue depth, read/write mix, dataset fullness, snapshot count, and application durability settings. Monitor CPU, temperature, throttling, memory pressure, and latency percentiles. A raw sequential headline cannot validate the storage policy.
Continue with the operations and recovery schedule after deployment.