add warning about over-provisioning
[pve-docs.git] / pvesm.adoc
CommitLineData
aa039b0f
DM
1[[chapter-storage]]
2ifdef::manvolnum[]
3PVE({manvolnum})
4================
38fd0958 5include::attributes.txt[]
aa039b0f
DM
6
7NAME
8----
9
10pvesm - Proxmox VE Storage Manager
11
12
13SYNOPSYS
14--------
15
16include::pvesm.1-synopsis.adoc[]
17
18DESCRIPTION
19-----------
20endif::manvolnum[]
21
22ifndef::manvolnum[]
23{pve} Storage
24=============
38fd0958 25include::attributes.txt[]
aa039b0f
DM
26endif::manvolnum[]
27
28The {pve} storage model is very flexible. Virtual machine images
29can either be stored on one or several local storages, or on shared
30storage like NFS or iSCSI (NAS, SAN). There are no limits, and you may
31configure as many storage pools as you like. You can use all
32storage technologies available for Debian Linux.
33
34One major benefit of storing VMs on shared storage is the ability to
35live-migrate running machines without any downtime, as all nodes in
36the cluster have direct access to VM disk images. There is no need to
37copy VM image data, so live migration is very fast in that case.
38
39The storage library (package 'libpve-storage-perl') uses a flexible
40plugin system to provide a common interface to all storage types. This
41can be easily adopted to include further storage types in future.
42
43
44Storage Types
45-------------
46
47There are basically two different classes of storage types:
48
49Block level storage::
50
51Allows to store large 'raw' images. It is usually not possible to store
52other files (ISO, backups, ..) on such storage types. Most modern
53block level storage implementations support snapshots and clones.
54RADOS, Sheepdog and DRBD are distributed systems, replicating storage
55data to different nodes.
56
57File level storage::
58
59They allow access to a full featured (POSIX) file system. They are
60more flexible, and allows you to store any content type. ZFS is
61probably the most advanced system, and it has full support for
62snapshots and clones.
63
64
65.Available storage types
66[width="100%",cols="<d,1*m,4*d",options="header"]
67|===========================================================
68|Description |PVE type |Level |Shared|Snapshots|Stable
69|ZFS (local) |zfspool |file |no |yes |yes
70|Directory |dir |file |no |no |yes
71|NFS |nfs |file |yes |no |yes
72|GlusterFS |glusterfs |file |yes |no |yes
73|LVM |lvm |block |no |no |yes
9801e1c3 74|LVM-thin |lvmthin |block |no |yes |yes
aa039b0f
DM
75|iSCSI/kernel |iscsi |block |yes |no |yes
76|iSCSI/libiscsi |iscsidirect |block |yes |no |yes
77|Ceph/RBD |rbd |block |yes |yes |yes
78|Sheepdog |sheepdog |block |yes |yes |beta
79|DRBD9 |drbd |block |yes |yes |beta
80|ZFS over iSCSI |zfs |block |yes |yes |yes
81|=========================================================
82
83TIP: It is possible to use LVM on top of an iSCSI storage. That way
84you get a 'shared' LVM storage.
85
ebc15cbc 86Thin provisioning
2afe468c 87~~~~~~~~~~~~~~~~~
ebc15cbc 88
2afe468c
DM
89A number of storages, and the Qemu image format `qcow2`, support _thin
90provisioning_. With thin provisioning activated, only the blocks that
91the guest system actually use will be written to the storage.
ebc15cbc 92
2afe468c
DM
93Say for instance you create a VM with a 32GB hard disk, and after
94installing the guest system OS, the root filesystem of the VM contains
953 GB of data. In that case only 3GB are written to the storage, even
96if the guest VM sees a 32GB hard drive. In this way thin provisioning
97allows you to create disk images which are larger than the currently
98available storage blocks. You can create large disk images for your
99VMs, and when the need arises, add more disks to your storage without
100resizing the VMs filesystems.
101
102All storage types which have the 'Snapshots' feature also support thin
103provisioning.
ebc15cbc 104
ba1d96fd
DM
105CAUTION: If a storage runs full, all guests using volumes on that
106storage receives IO error. This can cause file system inconsistencies
107and may corrupt your data. So it is advisable to avoid
108over-provisioning of your storage resources, or carefully observe
109free space to avoid such conditions.
ebc15cbc 110
aa039b0f
DM
111Storage Configuration
112---------------------
113
114All {pve} related storage configuration is stored within a single text
115file at '/etc/pve/storage.cfg'. As this file is within '/etc/pve/', it
116gets automatically distributed to all cluster nodes. So all nodes
117share the same storage configuration.
118
119Sharing storage configuration make perfect sense for shared storage,
120because the same 'shared' storage is accessible from all nodes. But is
121also useful for local storage types. In this case such local storage
122is available on all nodes, but it is physically different and can have
123totally different content.
124
125Storage Pools
126~~~~~~~~~~~~~
127
128Each storage pool has a `<type>`, and is uniquely identified by its `<STORAGE_ID>`. A pool configuration looks like this:
129
130----
131<type>: <STORAGE_ID>
132 <property> <value>
133 <property> <value>
134 ...
135----
136
137NOTE: There is one special local storage pool named `local`. It refers to
871e1fd6 138the directory '/var/lib/vz' and is automatically generated at installation
aa039b0f
DM
139time.
140
141The `<type>: <STORAGE_ID>` line starts the pool definition, which is then
871e1fd6
FG
142followed by a list of properties. Most properties have values, but some of
143them come with reasonable default. In that case you can omit the value.
aa039b0f
DM
144
145.Default storage configuration ('/etc/pve/storage.cfg')
9801e1c3
DM
146----
147dir: local
aa039b0f 148 path /var/lib/vz
9801e1c3
DM
149 content iso,vztmpl,backup
150
151lvmthin: local-lvm
152 thinpool data
153 vgname pve
154 content rootdir,images
155----
aa039b0f
DM
156
157Common Storage Properties
158~~~~~~~~~~~~~~~~~~~~~~~~~
159
871e1fd6 160A few storage properties are common among different storage types.
aa039b0f
DM
161
162nodes::
163
164List of cluster node names where this storage is
165usable/accessible. One can use this property to restrict storage
166access to a limited set of nodes.
167
168content::
169
170A storage can support several content types, for example virtual disk
171images, cdrom iso images, container templates or container root
871e1fd6 172directories. Not all storage types support all content types. One can set
aa039b0f
DM
173this property to select for what this storage is used for.
174
175images:::
176
177KVM-Qemu VM images.
178
179rootdir:::
180
871e1fd6 181Allow to store container data.
aa039b0f
DM
182
183vztmpl:::
184
185Container templates.
186
187backup:::
188
189Backup files ('vzdump').
190
191iso:::
192
193ISO images
194
195shared::
196
197Mark storage as shared.
198
199disable::
200
201You can use this flag to disable the storage completely.
202
203maxfiles::
204
205Maximal number of backup files per VM. Use `0` for unlimted.
206
207format::
208
209Default image format (`raw|qcow2|vmdk`)
210
211
212WARNING: It is not advisable to use the same storage pool on different
871e1fd6 213{pve} clusters. Some storage operation need exclusive access to the
aa039b0f 214storage, so proper locking is required. While this is implemented
871e1fd6 215within a cluster, it does not work between different clusters.
aa039b0f
DM
216
217
218Volumes
219-------
220
221We use a special notation to address storage data. When you allocate
871e1fd6 222data from a storage pool, it returns such a volume identifier. A volume
aa039b0f
DM
223is identified by the `<STORAGE_ID>`, followed by a storage type
224dependent volume name, separated by colon. A valid `<VOLUME_ID>` looks
225like:
226
227 local:230/example-image.raw
228
229 local:iso/debian-501-amd64-netinst.iso
230
231 local:vztmpl/debian-5.0-joomla_1.5.9-1_i386.tar.gz
232
233 iscsi-storage:0.0.2.scsi-14f504e46494c4500494b5042546d2d646744372d31616d61
234
235To get the filesystem path for a `<VOLUME_ID>` use:
236
237 pvesm path <VOLUME_ID>
238
239Volume Ownership
240~~~~~~~~~~~~~~~~
241
242There exists an ownership relation for 'image' type volumes. Each such
243volume is owned by a VM or Container. For example volume
244`local:230/example-image.raw` is owned by VM 230. Most storage
245backends encodes this ownership information into the volume name.
246
871e1fd6 247When you remove a VM or Container, the system also removes all
aa039b0f
DM
248associated volumes which are owned by that VM or Container.
249
250
251Using the Command Line Interface
252--------------------------------
253
871e1fd6
FG
254It is recommended to familiarize yourself with the concept behind storage
255pools and volume identifiers, but in real life, you are not forced to do any
aa039b0f
DM
256of those low level operations on the command line. Normally,
257allocation and removal of volumes is done by the VM and Container
258management tools.
259
260Nevertheless, there is a command line tool called 'pvesm' ({pve}
261storage manager), which is able to perform common storage management
262tasks.
263
264
265Examples
266~~~~~~~~
267
268Add storage pools
269
270 pvesm add <TYPE> <STORAGE_ID> <OPTIONS>
271 pvesm add dir <STORAGE_ID> --path <PATH>
272 pvesm add nfs <STORAGE_ID> --path <PATH> --server <SERVER> --export <EXPORT>
273 pvesm add lvm <STORAGE_ID> --vgname <VGNAME>
274 pvesm add iscsi <STORAGE_ID> --portal <HOST[:PORT]> --target <TARGET>
275
276Disable storage pools
277
278 pvesm set <STORAGE_ID> --disable 1
279
280Enable storage pools
281
282 pvesm set <STORAGE_ID> --disable 0
283
284Change/set storage options
285
286 pvesm set <STORAGE_ID> <OPTIONS>
287 pvesm set <STORAGE_ID> --shared 1
288 pvesm set local --format qcow2
289 pvesm set <STORAGE_ID> --content iso
290
291Remove storage pools. This does not delete any data, and does not
292disconnect or unmount anything. It just removes the storage
293configuration.
294
295 pvesm remove <STORAGE_ID>
296
297Allocate volumes
298
299 pvesm alloc <STORAGE_ID> <VMID> <name> <size> [--format <raw|qcow2>]
300
301Allocate a 4G volume in local storage. The name is auto-generated if
302you pass an empty string as `<name>`
303
304 pvesm alloc local <VMID> '' 4G
305
306Free volumes
307
308 pvesm free <VOLUME_ID>
309
310WARNING: This really destroys all volume data.
311
312List storage status
313
314 pvesm status
315
316List storage contents
317
318 pvesm list <STORAGE_ID> [--vmid <VMID>]
319
320List volumes allocated by VMID
321
322 pvesm list <STORAGE_ID> --vmid <VMID>
323
324List iso images
325
326 pvesm list <STORAGE_ID> --iso
327
328List container templates
329
330 pvesm list <STORAGE_ID> --vztmpl
331
332Show filesystem path for a volume
333
334 pvesm path <VOLUME_ID>
335
deb4673f
DM
336ifdef::wiki[]
337
338See Also
339--------
340
341* link:/index.php/Storage:_Directory[Storage: Directory]
342
343* link:/index.php/Storage:_GlusterFS[Storage: GlusterFS]
344
345* link:/index.php/Storage:_User_Mode_iSCSI[Storage: User Mode iSCSI]
346
347* link:/index.php/Storage:_iSCSI[Storage: iSCSI]
348
349* link:/index.php/Storage:_LVM[Storage: LVM]
350
351* link:/index.php/Storage:_LVM_Thin[Storage: LVM Thin]
352
353* link:/index.php/Storage:_NFS[Storage: NFS]
354
355* link:/index.php/Storage:_RBD[Storage: RBD]
356
357* link:/index.php/Storage:_ZFS[Storage: ZFS]
358
359
360endif::wiki[]
361
251666be
DM
362ifndef::wiki[]
363
aa039b0f
DM
364// backend documentation
365
366include::pve-storage-dir.adoc[]
367
368include::pve-storage-nfs.adoc[]
369
370include::pve-storage-glusterfs.adoc[]
371
372include::pve-storage-zfspool.adoc[]
373
374include::pve-storage-lvm.adoc[]
375
9801e1c3
DM
376include::pve-storage-lvmthin.adoc[]
377
aa039b0f
DM
378include::pve-storage-iscsi.adoc[]
379
380include::pve-storage-iscsidirect.adoc[]
381
382include::pve-storage-rbd.adoc[]
383
384
251666be 385
aa039b0f
DM
386ifdef::manvolnum[]
387include::pve-copyright.adoc[]
388endif::manvolnum[]
389
251666be
DM
390endif::wiki[]
391