Viewing: osd-zfs.fallocate_zero_blocks.4
.TH OSD-ZFS.FALLOCATE_ZERO_BLOCKS 4 2026-07-11 "Lustre" "Lustre Tunable"
.SH NAME
osd-zfs.fallocate_zero_blocks \- control ZFS OSD fallocate
.SH SYNOPSIS
.SY "lctl set_param"
.BI osd-zfs. FSNAME -*.fallocate_zero_blocks= MODE
.YS
.SY "lctl get_param"
.BI osd-zfs. FSNAME -*.fallocate_zero_blocks
.YS
.SS PROPERTIES
.TP
.B Access Permissions
.br
.BR 644 | -rw-r--r--
.TP
.B Scope
.br
Per-MDT and OST device.
.TP
.B Config
.br
.B osd-zfs.*.fallocate_zero_blocks
is always present for MDT and OST devices after 2.18.0
.TP
.B Default
.br
.RB fallocate_zero_blocks=0
.TP
.B Valid Range
.br
.RB fallocate_zero_blocks= -1
.br
.RB fallocate_zero_blocks= 2
.br
.SH DESCRIPTION
.B fallocate_zero_blocks
controls fallocate operation for ZFS OSD.
.PP
.I MODE
can be -1, 0, 1 or 2. Where -1 means that all fallocate operation is
disabled. 0 and 1 is related to generic fallocate alloc operation and for now is
not supported at all. Setting value of 0 or 1 will always return -EOPNOTSUPP for
all fallocate operations even though the actual mode passed to fallocate() is
set to 3. 2 means punch operation is supported.
.SH EXAMPLES
.PP
To disable fallocate under osd-zfs
.EX
.RB oss# " lctl set_param osd-zfs.*.fallocate_zero_blocks=-1"
.EE
.PP
To enable fallocate(PUNCH_HOLE) under osd-zfs:
.EX
.RB oss# " lctl set_param osd-zfs.testfs-OST0000.fallocate_zero_blocks=2"
.SH AVAILABILITY
The
.B fallocate_zero_blocks
parameter is part of the
.BR lustre (7)
filesystem package since release 2.18.0.
.\" Added in commit 2.17.56
.SH SEE ALSO
.BR zfs (4),
.BR lustre (7),
.BR lctl (8),
.BR lctl-get_param (8),
.BR lctl-set_param (8),
.BR zpool (8)
