Those are true, but "existing system" doesn't mean the same partitions already used without data being wiped. You have to backup the data, setup the LVM, restore the data.
Or
you can have a running system on 1 disk (A), add another disk (B), partition "B", create LVM PVs into 1 or more partitions on "B" and use it that way. All without touching "A". For example,
Code:
sdd 7.3T disk
└─sdd1 7.3T part LVM2_member
├─istar--8TB-istar--back3--a 3.7T lvm ext4 /misc/b-D3
└─istar--8TB-istar--back3--b 3.6T lvm ext4 /misc/b-D4
$ sudo vgs
VG #PV #LV #SN Attr VSize VFree
istar-8TB 1 2 0 wz--n- 7.28t 14.03g
...
$ sudo lvs
LV VG Attr LSize Pool Origin Data%
istar-back3-a istar-8TB -wi-ao---- 3.65t
istar-back3-b istar-8TB -wi-ao---- 3.62t
sdd is an 8TB disk.
sdd1 is a partition that fills the entire disk. It is also a PV.
istar-8TB is the VG that uses the total PV.
istar-back3-a and istar-back3-b are the LVs that take storage from the VG, just a little less than everything available. This leaves room for an emergency increase where it is needed just before storage is really all used up.
Any LVM isn't a "disk" thing. It is a "Partition" thing. Windows people seem to get those confused, because MSFT calls partitions "drives." This is factually incorrect.
Bookmarks