LCTL-POOL_ADD(8)
LCTL-POOL_ADD(8) π
okay bestie, welcome to lctl-pool_add β section 8 of the Lustre Configuration Utilities, aka the girlboss move where you add OSTs to a named pool. let's get into it π«Ά
β¨ NAME β¨
lctl-pool_add β add OSTs to a named pool π
πΈ SYNOPSIS πΈ
lctl pool_add [--nowait|-n] FSNAME.POOLNAME OST_INDEX1 [OST_INDEX2...]
lctl pool_add [--nowait|-n] FSNAME.POOLNAME OST_RANGE1 [OST_RANGE2...]
π¦ DESCRIPTION π¦
So what we're doing here is adding one or more OSTs to the pool named POOLNAME inside the filesystem named FSNAME. Iconic. There are a few formats you can use to name those OSTs, and ngl the options are giving versatility.
The simplest one? Just hand over a list of OST_INDEX numbers or OST names. Easy. But you can also drop an OST_RANGE expression shaped like [START-END] or [START-END/STEP], which adds a whole bunch of OSTs with index numbers from START to END inclusive, and if you add the STEP part it skips every STEP index values. lowkey powerful.
NOTE: after the MGS configuration gets updated, this command will wait up to 12s and check if pools are updated on a client β unless you used the--nowaitoption. And heads up: if the MGS is on a separate node from the MDS, a Lustre client must be mounted on the MGS node while thelctlcommands are being run for this to work. Otherwise, the client check is just skipped. no drama, just facts. π
This named list of OSTs is what lfs-setstripe(1) can use to specify which OSTs new files get created on. BUT β super important β the OSTs that make up a specific pool may change over time, and what's actually stored on each file is the poolname used at creation time, not necessarily the OSTs sitting in the current pool. periodt. π
The lfs-find(1) command can locate files that were initially created on the specified poolname. Meanwhile lfs-df(1) can show you just the free space or inodes in a named pool. And lfs-quota(1) plus lfs-setquota(1) can use a pool name to get/set a quota limit for OSTs in the pool. the whole squad is eating. π
π OPTIONS π
-n,--nowaitβ do not wait and check if the pool is updated on a client. this is the move when you're calling a lot oflctl pool_*commands in a row, so you don't sit there waiting for each one to finish. π
π EXAMPLES π
Create a pool named flash in the testfs filesystem:
# lfs pool_new testfs.flash
Add OSTs numbered 8, 10, and 12 through 14 to the testfs.flash pool:
# lfs pool_add testfs.flash 8 10 12-14
π AVAILABILITY π
lctl pool_add is part of the lustre(7) filesystem package since release 1.7.0 β¨
π«Ά SEE ALSO π«Ά
lfs-df(1), lfs-find(1), lfs-setstripe(1), lustre(7), lctl(8), lctl-pool_new(8)
and that's the whole page, bestie π this was Lustre, dated 2025-05-01, and yeah the manual page title is literally LCTL-POOL_ADD(8) β go forth and pool responsibly π§πΈ
π¦ where this page comes from
- ubuntu stonking: lustre-client-utils 2.17.58+dfsg-0ubuntu2 Β· GPL-2.0-only
- π girlified by deepseek-ai/DeepSeek-V4.1-Flash on : a language model's rewrite of the original, under its terms