Storage

Where servers' disks live on a node: file storages, LVM-thin pools, ZFS, shared NFS exports and Ceph RBD pools, several per node, with placement by storage tag.

A node keeps its servers' disks on one or more storages. Every node has the default storage, /var/lib/vmstack/images, which also holds the cached images and ISOs. You can add more storages per node on its Storage tab, made from empty disks, empty partitions, free space in a volume group, or a directory you mounted, and share NFS exports or Ceph pools between the nodes of a location.

Kind Each disk is Snapshots Shared between nodes
Files (XFS) a qcow2 or raw file when the file system clones files (XFS made by VMStack does) no
LVM-thin pool a thin logical volume, raw yes, thin snapshots no
ZFS a sparse ZFS volume, raw, compressed yes, ZFS clones no
NFS a file on the export when the export clones files (NFS 4.2 on XFS or Btrfs) yes
Ceph RBD (beta) a raw RBD image yes, RBD clones yes

Plain (thick) LVM, where each disk is a fully allocated logical volume, is not offered.

New servers are placed by storage tag: a plan with a tag goes only to storages with that tag, a plan without one to any storage marked Use for new servers. Among those, the node's Fill order decides. Administrators can choose a storage when creating a server or adding a disk, and move a disk between storages of the same node later.

  1. Local file storageThe default storage, turning empty disks into storage, and adding file storages from disks, partitions, volume groups or a mounted directory.
  2. LVM and LVM-thinFile storages on LVM, and LVM-thin pools with a thin volume per disk: instant creation and snapshots, over-provisioning and autoextend.
  3. ZFSZFS storages: install ZFS, make a pool or use one, sparse compressed volumes per disk, snapshots and clones, pool health and removal.
  4. NFS shared storageShare an NFS export between the nodes of a location: prepare it, add it once, attach it to nodes, and move servers without copying disks.
  5. Ceph RBD BetaUse a pool of an existing Ceph cluster as shared storage: prepare the pool and cephx user, install the client, add the pool, attach nodes.
  6. Moving disks between storagesMove a disk of a server to another storage of its node: the checks, what happens if it fails, and how moves between nodes send only data.