Skip to content

lvm_logical_volume

An LVM logical volume, grown to its size.

Use lvm_logical_volume to make a logical volume in a volume group, and to grow it. Each step is one volume, at /dev/<vg>/<name>, ready for a filesystem step. The host needs the LVM tools.

A 20 GiB volume for an app, with a file system that grows with it:

step "app_lv" {
lvm_logical_volume {
vg = "vg0"
name = "app"
size = "20g"
}
}
step "app_fs" {
filesystem {
device = "/dev/vg0/app"
type = "xfs"
resize = true
}
}

Give a volume whatever space the group has left when it’s made:

step "logs_lv" {
lvm_logical_volume {
vg = "vg0"
name = "logs"
size = "100%FREE"
}
}
Name Type Required Description
force bool No Shrink the volume when it’s bigger than size, destroying whatever is past the new end, file system included.
name string Yes The volume’s name. Its device is /dev//.
size string Yes Its size: 512m, 20g or 1.5t (binary units; a bare number is MiB), a share of the group such as 50%VG, or a share of its free space such as 100%FREE, which sizes the volume only when it’s made.
vg string Yes The volume group it’s in.

Conn3x sets these; reference them as conn3x_lvm_logical_volume.<circuit>__<step>["<host>"].<name>.

Name Type Description
in_sync bool Whether the volume exists at its size.
revision string Changes each time Conn3x changes this on the host. Handlers watch it.

A step lowers to the conn3x_lvm_logical_volume resource, one instance per host, and connects with its circuit’s ssh block.

Reading. On every plan, Conn3x asks LVM for the volume’s size and the group’s extent size. in_sync is true when the volume exists at its size, rounded up to whole extents, as LVM rounds it.

Sizes. size is one of:

  • An amount: 512m, 20g, 1.5t, in binary units, so 1g is 1 GiB. A bare number is MiB, as for lvcreate -L.
  • A share of the group, such as 50%VG. Conn3x keeps the volume at that share of the group’s size.
  • A share of the group’s free space, such as 100%FREE. It sizes the volume only when it’s made, since the free space shrinks once it is. After that, the volume is in sync at whatever size it has.

Making and growing it. If the volume doesn’t exist, Conn3x runs lvcreate, wiping any stale signature in the new space. If it’s smaller than size, Conn3x runs lvextend. Grow the file system on it with a filesystem step that sets resize = true.

What it won’t shrink. A volume bigger than size stops the apply, unless force = true. Then Conn3x runs lvreduce, which destroys whatever is past the new end, and shrinks no file system first. With force on, every plan warns that the step can destroy data.

Rules. vg and name are LVM names: letters, digits and _ . + -, not starting with -. A percentage is from 1 to 100.

Changes. Changing size grows the volume, or shrinks it with force. Changing vg or name replaces the step.

When you delete a lvm_logical_volume step, the volume stays on the host, with its data. There’s no on_remove.

conn3x convert ansible turns community.general.lvol tasks with state: present into lvm_logical_volume steps, with lv as name. force carries over, with a TODO to take it off. resizefs leaves a TODO to add a filesystem step with resize = true. A size relative to the current one, such as +100%FREE, or a share of %PVS, stays a TODO, as do snapshots and thin pools.