LVM mainly offers the following benefits which a classical MBR or GPT partition layout cannot offer due to its nature of having to reside in a partition table with physical sector boundaries.
increased abstraction, flexibility, and control. Logical volumes can have meaningful names likedatabasesorroot-backup. Volumes can be resized dynamically as space requirements change and migrated between physical devices within the pool on a running system or exported easily. LVM also offers advanced features like snapshotting, striping, and mirroring.1
| ready drive for LVM | pvcreate |
|---|---|
| create a new VG | vgcreate NAME PVNAME |
| show logical volumes | lvdisplay |
| enlarge LOGICALVOL by 10GB | lvresize --size +10G LOGICALVOL |
| create logical volume with 10GB in vg0 | lvcreate -L 10G vg0 -n NAME |
| PV | Physical Volume (e.g. a disk; can contain a LV) |
|---|---|
| LV | Logical Volume |
| LVM | Logical Volume Manager |
| LE | Logical Extent |
| PE | Physical Extent |
| VG | Volume Group (grouping of one or more PVs) |
| VGDA | Volume Group Descriptor Area |
This requires temporarily having 2x the size of your LVM volume. You need to create a mirror of your data, with the new leg of the mirror striped over the target disks, then drop the old leg of the mirror that was not striped.
If you want to stripe over ALL of your disks (including the one that was already used), you also need to specify --alloc anywhere otherwise the mirror code will refuse to use any disk twice.
# convert to a mirror (-m1), with new leg striped over 4 disks: /dev/sdb, /dev/sdc, /dev/sdd, /dev/sde # --mirrorlog core - use in-memory status during the conversion # --interval 1: print status every second lvconvert --interval 1 -m1 $myvg/$mylv --mirrorlog core --type mirror --stripes 4 /dev/sd{b,c,d,e} # drop the old leg, /dev/sda lvconvert --interval 1 -m0 $myvg/$mylv /dev/sda
mdadm --create /dev/md0 --verbose --level=1 --raid-devices=2 missing /dev/sdb1