LFS-MIGRATE(1)

📦 version 2.17.58, in ubuntu stonking

✨ NAME ✨

okay so bestie, welcome to lfs-migrate(1) from the Lustre User Utilities 👑 — she migrates files or directories between MDTs or OSTs. That's it, that's the whole job, and she does it flawlessly. no cap 💅

🦋 SYNOPSIS 🦋

lfs migrate [-h] [-v] [-A [-K CAP] [-M MIN_FREE] [-X MAX_FREE ]]
            [SETSTRIPE_OPTIONS ... ] FILES_FROM
lfs migrate -m START_MDT_INDEX [-cdHv] DIRECTORY

🌸 DESCRIPTION 🌸

so basically: migrate OST objects between OSTs for the file(s) you passed in FILES_FROM, OR — plot twist — use -m to migrate DIRECTORY and every inode/subdirectory hanging out inside it over to other MDTs (and yes, it's recursive by default, she does the absolute most 💖).

💜 OPTIONS 💜

🍓 OST MIGRATE OPTIONS 🍓

The lfs migrate command can be used for moving files from one (or more) OSTs to other OSTs (e.g. for space balancing between OSTs, or to evacuate an OST for hardware reasons), to change the stripe count or other layout parameters of a file (e.g. to increase the bandwidth of a file by striping it over multiple OSTs), or to move the file between different classes of storage (e.g. SSD vs. HDD OSTs, or local vs. remote OSTs in different pools). versatile queen behavior ngl 🎀

FILES_FROM may be:

  • FILENAME [...] — File names are listed on command line. Multiple file names are separated by space char.
  • -0, --null — Read file names from stdin by default. Each file name is followed by a NUL char. Usually is used after a pipeline from lfs find --print0 command.
  • --files-from=LIST_FILE — Read file names from file LIST_FILE. One line for each file name. If LIST_FILE is -, read filenames from standard input. If --null is also specified, file names are separated by a NUL char in the input.

In OST object migration mode, the command supports the same SETSTRIPE_OPTIONS listed in lfs-setstripe(1) to specify the layout of the target file. The migrate command differs from lfs setstripe in that lfs migrate will copy the data from the existing file(s) using the new layout parameters to the new OST(s). In contrast, lfs setstripe is used for creating new (empty) files with the specified layout. For OST object migration, there are additional options available:

If setstripe is restricted to specified users, this also applies to lfs migrate. See lfs-setstripe(1) for details.

  • -A — Automatically determine the stripe count for the file, using the algorithm count = sqrt(filesize_in_GB) + 1. This option may not be specified at the same time as the -c or -R options.
  • -K CAP — When -A is set, limit the migrated file to use on each OST at most 1/CAP of the available space of the smallest OST. If this option is not set, a default value of 100 is used, limiting the object size to 1% of available space.
  • -b, --block — Block access to the file by other applications during data migration (default). This prevents other processes from accessing the file during migration, which prevents data writes to the old file objects from being lost. This should be used if an OST needs to be completely emptied prior to its removal, to ensure all requested files are migrated off the OST.
  • -D, --non-direct — Do not use O_DIRECT read and write operations when migrating a file. The O_DIRECT option avoids data copy from kernel buffers into userspace, which can impose CPU and memory overhead on the copy operation, but makes read and write operations synchronous. Using the --non-direct option uses buffered read/write operations, which may improve migration speed at the cost of more CPU and memory overhead.

This option cannot be used on encrypted files when the encryption key is not available. It will result in -ENOKEY.

  • -F, --fid — FID(s) are provided from command line/stdin/ LIST_FILE instead of file path(s). Option --lustre-dir=LUSTRE_MOUNT_POINT may be specified at the same time.
  • -h, --help — Print usage message.
  • --lustre-dir=LUSTRE_MOUNT_POINT — The mount point of Lustre file system on which FIDs are. If there is only one mounted Lustre file system, this option is not necessary. This option must be specified together with --fid.
  • -M, --min-freeMIN_FREE — When -A is set, only consider OSTs with free space greater than the MIN_FREE value to be available for migration. The value is specified in KB. If this option is not set, a default of 256MB is used.
  • -X, --max-freeMAX_FREE — When -A is set, MAX_FREE is the maximum amount of free space that can be considered available for the migration of the file on each OST. The value is specified in KB. This option is useful for testing, by simulating OSTs that are nearly full.
  • -n, --non-block — Abort migration if concurrent file access is detected. This can be used with OST space balancing migration to avoid interfering with file access by applications if there is not a requirement to migrate any particular file to the new layout.
  • --restripe — Restripe the file using the default directory striping layout. This option copies the default layout from the parent directory and applies it to the file being migrated. This option may not be specified at the same time as the -c or -S or -p or -A or -E options.
  • --stats — This option enables progress updates every 5 seconds, in YAML format.
  • --stats-interval=STATS_INTERVAL — This option enables progress updates every STATS_INTERVAL seconds, in YAML format.
  • -v, --verbose — Print each filename as it is migrated.
  • -W, --bandwidth-limit, --bandwidth BANDWIDTH_MB[MG] — Limit the read and write bandwidth a migrate job consumes to no more than BANDWIDTH_MB MiB/s. An optional suffix can be used to specify units in Mebibytes or Gibibytes per second.

🧚 MDT MIGRATE OPTIONS 🧚

  • -m, --mdt-index=MDT_INDEX[,MDT_INDEX,...] — The specified DIRECTORY and (by default) all subdirectories and inodes will be migrated to the MDT with the specified MDT_INDEX. This is useful if new MDTs have been added to a filesystem and existing user or project directories should be migrated off old MDTs to balance the space usage and future metadata workload. If MDT_INDEX is -1, the MDT index will be selected to balance by free space and inodes among available MDTs. If multiple MDT_INDEX values are specified in a comma-separated list, then all subdirectories will be striped across all of the specified MDT indices as if an equivalent -cN option were given. This is typically NOT what you want to do. 💀
  • -c, --mdt-count=COUNT — All directories and subdirectories in the tree will be striped across COUNT MDTs, always using MDT_INDEX as the primary MDT for the directory. If MDT_INDEX is -1 then COUNT directory stripes will be chosen from MDTs proportional to the amount of free inodes and space on each MDT. If multiple MDT_INDEX values are specified in a comma-separated list, then the number of specified MDT_INDEX values must match COUNT.
  • -C, --mdt-overcount=COUNT — The number of directory stripes to create, creating > 1 stripe per MDT if COUNT exceeds the number of MDTs in the file system. 0 means to use the filesystem-wide default stripe count (default 1), and range -1 to -5 means to stripe over all available MDTs in multiple of MDT count. For example, -1 means one stripe per MDT, -2 means two stripes per MDT, and so on. The maximum is 5 stripes per MDT. This allows a single directory to more fully use the capability of each MDT.
  • -d, --directory — Only migrate the specified DIRECTORY and the non-directory inodes that are directly located within it. Similar to ls -d and lfs getstripe -d.
  • --clear-fixed — Do not set LMV_HASH_FLAG_FIXED upon migrating the directory. It allows a striped dir with only one stripe to be converted to a plain directory.
  • -H, --mdt-hash=HASH_TYPE — Use HASH_TYPE for the new directory layout.
    • all_char (type 1) — Sum of ASCII characters modulo number of MDTs. This provides weak hashing of the filename, and is suitable for only testing or when the input is known to have perfectly uniform distribution (e.g. sequential numbers).
    • fnv_1a_64 (type 2) — Fowler-Noll-Vo (FNV-1a) hash algorithm. This provides reasonably uniform, but not cryptographically strong, hashing of the filename. (default)
    • crush (type 3) — CRUSH hash algorithm. This is a consistent hash algorithm, so minimum sub files need to relocate during directory restripe.
    • crush2 (type 4) — CRUSH hash algorithm (as above), with optimizations that allow temporary names created by the mktemp(3) family of functions to have both temporary and final names hash to the same MDT index. This allows renaming the temporary name to the final name without creating a remote directory entry or doing a cross-MDT rename (e.g. test_foo.XXXXXX or .test_foo.XXXXXXXXXX is on the same MDT as test_foo).

⭐ NOTES ⭐

Only the root user can migrate directories. Files that have been archived by HSM or are currently opened will fail to migrate, user can run the same migrate command again to finish migration when files are ready. Both inode and directory entry will be migrated. During migration directory and sub files can be accessed like normal ones, but the migration itself cannot be interrupted. 🫶

It is not currently possible to migrate files with an mdt component (Data-on-MDT, DoM). If it is necessary to migrate such files off a particular MDT, they must first be migrated to have a non-DoM file layout and then the inodes migrated separately. See EXAMPLES for details on how to migrate DoM files between MDTs.

🔥 WARNING 🔥

Each migrated file or directory will have a new FID, and hence a new inode number. As a consequence, files archived by Lustre HSM that depend on the FID as the identifier in the HSM archive cannot currently be migrated. Having a new inode number may also cause backup tools to consider the migrated file(s) to be a new, and cause them to be backed up again.

like… be so serious about that one, bestie, it will bite you 😭

✨ EXAMPLES ✨

ok so here's where it gets FUN bestie, real commands, real vibes 🫶

one: two stripes, one file, zero stress 💅

this migrates the data in file1 into a new layout with 2 stripes:

# lfs migrate -c 2 /mnt/lustre/file1

two: let the tool do the math ✨

this migrates file2 using automatic striping (the stripe count gets calculated based on file size, she's doing the thinking so you don't have to):

# lfs migrate -A /mnt/lustre/file2

three: auto striping but make it custom 🎀

this migrates file3 using automatic striping with custom capacity limits (5% of OST space):

# lfs migrate -A -K 20 /mnt/lustre/file3

four: composite layout, three components, main character energy 👑

this migrates the data in file2 into a three component composite layout (number of stripes depends on file size):

# lfs migrate -E 256M -c 1 -E 16G -c 4 -E eof -c 40 /mnt/lustre/file2

five: moving a whole directory tree across MDTs 🦋

recursively move the subdirectories and inodes contained in directory remotedir from its current MDT to a 2-striped directory on MDT0000 and MDT0002. the testremote directory and all of its subdirectories will be striped across both MDTs:

# lfs migrate -m 0,2 testremote

six: same vibe, but ONLY the first level 🙈

move ./testremote and the first level of sub files from their current MDT to the MDT with index 0 and 2. different from the above case, the layout of subdirectories under ./testremote won't be changed:

# lfs migrate -m 0,2 -d ./testremote

seven: the big combo move, DoM edition 🔥

set a default PFL layout (without any DoM component) on the directory topdir:

# lfs setstripe -E 256M -c 1 -E 16G -c 4 -E eof -c 40 topdir

then find and migrate all regular files that have an mdt component to copy the default layout from the specified topdir:

# lfs find dir -type f -L mdt -0 | lfs migrate -0 --copy topdir

and finally migrate the directory topdir and all files and subdirectories in that tree to MDT0002. this allows migrating files with DoM components off an MDT:

# lfs migrate -m 2 topdir

🌈 AVAILABILITY 🌈

the lfs migrate command is part of the lustre(7) filesystem package since release 2.4.0 💖 like, she's been around, she's established, she's iconic.

🌸 SEE ALSO 🌸

and that's the whole page bestie, you made it to the end, I'm so proud 🥹💜 go forth and migrate responsibly, and remember: check your flags, don't yolo a PFL layout, and always read the caveats. this was the Lustre manual page for LFS-MIGRATE(1), dated 2025-12-20 — star it, bookmark it, keep it in the group chat. periodt. 🎀✨

📦 where this page comes from