local-lvm.adoc

   1 [[chapter_lvm]]
   2 Logical Volume Manager (LVM)
   3 ----------------------------
   4 ifdef::wiki[]
   5 :pve-toplevel:
   6 endif::wiki[]
   7
   8 Most people install {pve} directly on a local disk. The {pve}
   9 installation CD offers several options for local disk management, and
  10 the current default setup uses LVM. The installer lets you select a
  11 single disk for such setup, and uses that disk as physical volume for
  12 the **V**olume **G**roup (VG) `pve`. The following output is from a
  13 test installation using a small 8GB disk:
  14
  15 ----
  16 # pvs
  17   PV         VG   Fmt  Attr PSize PFree
  18   /dev/sda3  pve  lvm2 a--  7.87g 876.00m
  19
  20 # vgs
  21   VG   #PV #LV #SN Attr   VSize VFree
  22   pve    1   3   0 wz--n- 7.87g 876.00m
  23 ----
  24
  25 The installer allocates three **L**ogical **V**olumes (LV) inside this
  26 VG:
  27
  28 ----
  29 # lvs
  30   LV   VG   Attr       LSize   Pool Origin Data%  Meta%
  31   data pve  twi-a-tz--   4.38g             0.00   0.63
  32   root pve  -wi-ao----   1.75g
  33   swap pve  -wi-ao---- 896.00m
  34 ----
  35
  36 root:: Formatted as `ext4`, and contains the operating system.
  37
  38 swap:: Swap partition
  39
  40 data:: This volume uses LVM-thin, and is used to store VM
  41 images. LVM-thin is preferable for this task, because it offers
  42 efficient support for snapshots and clones.
  43
  44 For {pve} versions up to 4.1, the installer creates a standard logical
  45 volume called ``data'', which is mounted at `/var/lib/vz`.
  46
  47 Starting from version 4.2, the logical volume ``data'' is a LVM-thin pool,
  48 used to store block based guest images, and `/var/lib/vz` is simply a
  49 directory on the root file system.
  50
  51 Hardware
  52 ~~~~~~~~
  53
  54 We highly recommend to use a hardware RAID controller (with BBU) for
  55 such setups. This increases performance, provides redundancy, and make
  56 disk replacements easier (hot-pluggable).
  57
  58 LVM itself does not need any special hardware, and memory requirements
  59 are very low.
  60
  61
  62 Bootloader
  63 ~~~~~~~~~~
  64
  65 We install two boot loaders by default. The first partition contains
  66 the standard GRUB boot loader. The second partition is an **E**FI **S**ystem
  67 **P**artition (ESP), which makes it possible to boot on EFI systems and to
  68 apply xref:sysadmin_firmware_persistent[persistent firmware updates] from the
  69 user space.
  70
  71
  72 Creating a Volume Group
  73 ~~~~~~~~~~~~~~~~~~~~~~~
  74
  75 Let's assume we have an empty disk `/dev/sdb`, onto which we want to
  76 create a volume group named ``vmdata''.
  77
  78 CAUTION: Please note that the following commands will destroy all
  79 existing data on `/dev/sdb`.
  80
  81 First create a partition.
  82
  83  # sgdisk -N 1 /dev/sdb
  84
  85
  86 Create a **P**hysical **V**olume (PV) without confirmation and 250K
  87 metadatasize.
  88
  89  # pvcreate --metadatasize 250k -y -ff /dev/sdb1
  90
  91
  92 Create a volume group named ``vmdata'' on `/dev/sdb1`
  93
  94  # vgcreate vmdata /dev/sdb1
  95
  96
  97 Creating an extra LV for `/var/lib/vz`
  98 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
  99
 100 This can be easily done by creating a new thin LV.
 101
 102  # lvcreate -n <Name> -V <Size[M,G,T]> <VG>/<LVThin_pool>
 103
 104 A real world example:
 105
 106  # lvcreate -n vz -V 10G pve/data
 107
 108 Now a filesystem must be created on the LV.
 109
 110  # mkfs.ext4 /dev/pve/vz
 111
 112 At last this has to be mounted.
 113
 114 WARNING: be sure that `/var/lib/vz` is empty. On a default
 115 installation it's not.
 116
 117 To make it always accessible add the following line in `/etc/fstab`.
 118
 119  # echo '/dev/pve/vz /var/lib/vz ext4 defaults 0 2' >> /etc/fstab
 120
 121
 122 Resizing the thin pool
 123 ~~~~~~~~~~~~~~~~~~~~~~
 124
 125 Resize the LV and the metadata pool with the following command:
 126
 127  # lvresize --size +<size[\M,G,T]> --poolmetadatasize +<size[\M,G]> <VG>/<LVThin_pool>
 128
 129 NOTE: When extending the data pool, the metadata pool must also be
 130 extended.
 131
 132
 133 Create a LVM-thin pool
 134 ~~~~~~~~~~~~~~~~~~~~~~
 135
 136 A thin pool has to be created on top of a volume group.
 137 How to create a volume group see Section LVM.
 138
 139  # lvcreate -L 80G -T -n vmstore vmdata