LVM - Logical Volume Manager

Install logical volumes (LVM)

Prepare and install volumes

Prepare a single disk or a disk raid:

lvm
lvm> pvcreate <physvol>
lvm> vgcreate <physvol> // or several physical volumes: <physvol1> <physvol2> <physvol3>
lvm> lvcreate <volgroup> <physical>
lvm> quit
mkfs.ext4 /dev/mapper/<volgroup-logvol>
mount /dev/mapper/<volgroup-logvol> <dir>

Auto-mount during boot:

vim /etc/fstab
/dev/mapper/<volgroup-logvol> <dir> ext3 noatime,user_xattr 0 0

Replace LVM disk to upgrade capacity

This example will copy /dev/sde1 mounted on /backup to /dev/sdf1 mounted on /replace and afterwards change mount point, logical volume and volume group names to the source names.

Prepare new disk

fdisk /dev/sdf
n to create a new partition, select 1 and accept all defaults
w to write the new partition table and quit
mkfs.ext3 /dev/sdf1
lvm
lvm > pvcreate /dev/sdf1
lvm > vgcreate vg_replace /dev/sdf1
lvm > vgdisplay vg_replace to check ### of free extents
lvm > lvcreate -l### vg_replace (do not use -LxxGB, use size with -l### which is number of extents)
lvm > lvrename vg_replace lvol0 lv_replace
lvm > vgcfgbackup to backup volume group configurations
lvm > exit
mkfs -t ext3 /dev/vg_replace/lv_replace
mkdir /replace
mount /dev/vg_replace/lv_replace /replace

Copy content and assign new disk

rsync -avH /backup/* /replace/

for an entire BackupPC structure

rsync -avxHAWX --info=progress2 /backup/ /new-backup/

verify the structure between source and destination
```
rsync -avcn /backup/ /new-backup/
```
umount /backup
umount /new-backup
lvm
lvm > lvrename vg_backup/lv_backup lv_backup1
lvm > lvchange vg_backup/lv_backup1 -an
lvm > vgrename vg_backup vg_backup1
lvm > lvchange vg_backup1/lv_backup1 -ay
lvm > lvrename vg_replace/lv_replace lv_backup
lvm > lvchange vg_replace/lv_backup -an
lvm > vgrename vg_replace vg_backup
lvm > lvchange vg_backup/lv_backup -ay
lvm > exit
mount /dev/vg_backup/lv_backup /backup -o noatime,user_xattr

Setting up LVM on top of a Linux Software Raid

prepare the new disks with a partition (see above), do not create the filesystem yet
mdadm –create –verbose /dev/md5 –level=1 –raid-devices=2 /dev/sde /dev/sdf
watch the progress with cat /proc/mdstat
when finished, restart the computer (shutdown -r now)

Merge 2 volume groups

Unmount and remove LV0 and LV1 from VG1 with umount/lvremove
Remove VG1 with vgremove
Unmount LV0 and LV1 from VG0 with umount
Extend VG0 with any available PVs if necessary
Mount LV0 and LV1 on VG0 with mount

lvcreate -l<extents> -n testlv testvg
mkfs.ext4 /dev/mapper/vg-lv

Increase size of physical volume and/or logical volume

Replace hard disks in a RAID1 array, or change from a RAID1 to a RAID5 array, the resulting array being MD127 (see mdadm - Software RAID)
Use pvresize to extend the physical volume to consume all available space
```
pvresize /dev/md127
```
Use lvresize to extend the logical volume to consume all available space
```
lvresize -l +100%FREE /dev/mapper/<vg name>-<lv name>
```
Or, unmount the volume, delete the logical volumes and the volume group first, then create a new logical volume consuming all available space
```
lvremove /dev/mapper/<vg name>-<lv name>
vgremove <vg name>
lvcreate -l 100%FREE -n <lv name>  <vg name>
```

Increase / decrease size of logical volume

First, check the status with:

# pvs
# vgs
# lvs

Table of Contents