1 =============================================================================
2 Management of storage types and disk templates, incl. storage space reporting
3 =============================================================================
5 .. contents:: :depth: 4
10 Currently, there is no consistent management of different variants of storage
11 in Ganeti. One direct consequence is that storage space reporting is currently
12 broken for all storage that is not based on lvm technolgy. This design looks at
13 the root causes and proposes a way to fix it.
18 We propose to streamline handling of different storage types and disk templates.
19 Currently, there is no consistent implementation for dis/enabling of disk
20 templates and/or storage types.
22 Our idea is to introduce a list of enabled disk templates, which can be
23 used by instances in the cluster. Based on this list, we want to provide
24 storage reporting mechanisms for the available disk templates. Since some
25 disk templates share the same underlying storage technology (for example
26 ``drbd`` and ``plain`` are based on ``lvm``), we map disk templates to storage
27 types and implement storage space reporting for each storage type.
32 Add a new attribute "enabled_disk_templates" (type: list of strings) to the
33 cluster config which holds disk templates, for example, "drbd", "file",
34 or "ext". This attribute represents the list of disk templates that are enabled
35 cluster-wide for usage by the instances. It will not be possible to create
36 instances with a disk template that is not enabled, as well as it will not be
37 possible to remove a disk template from the list if there are still instances
40 The list of enabled disk templates can contain any non-empty subset of
41 the currently implemented disk templates: ``blockdev``, ``diskless``, ``drbd``,
42 ``ext``, ``file``, ``plain``, ``rbd``, and ``sharedfile``. See
43 ``DISK_TEMPLATES`` in ``constants.py``.
45 Note that the abovementioned list of enabled disk types is just a "mechanism"
46 parameter that defines which disk templates the cluster can use. Further
47 filtering about what's allowed can go in the ipolicy, which is not covered in
48 this design doc. Note that it is possible to force an instance to use a disk
49 template that is not allowed by the ipolicy. This is not possible if the
50 template is not enabled by the cluster.
52 FIXME: In what way should verification between the enabled disk templates in
53 the cluster and in the ipolicy take place?
55 We consider the first disk template in the list to be the default template for
56 instance creation and storage reporting. This will remove the need to specify
57 the disk template with ``-t`` on instance creation.
59 Currently, cluster-wide dis/enabling of disk templates is not implemented
60 consistently. ``lvm`` based disk templates are enabled by specifying a volume
61 group name on cluster initialization and can only be disabled by explicitly
62 using the option ``--no-lvm-storage``. This will be replaced by adding/removing
63 ``drbd`` and ``plain`` from the set of enabled disk templates.
65 Up till now, file storage and shared file storage could be dis/enabled at
66 ``./configure`` time. This will also be replaced by adding/removing the
67 respective disk templates from the set of enabled disk templates.
69 There is currently no possibility to dis/enable the disk templates
70 ``diskless``, ``blockdev``, ``ext``, and ``rdb``. By introducing the set of
71 enabled disk templates, we will require these disk templates to be explicitely
72 enabled in order to be used. The idea is that the administrator of the cluster
73 can tailor the cluster configuration to what is actually needed in the cluster.
74 There is hope that this will lead to cleaner code, better performance and fewer
77 When upgrading the configuration from a version that did not have the list
78 of enabled disk templates, we have to decide which disk templates are enabled
79 based on the current configuration of the cluster. We propose the following
80 update logic to be implemented in the online update of the config in
81 the ``Cluster`` class in ``objects.py``:
82 - If a ``volume_group_name`` is existing, then enable ``drbd`` and ``plain``.
83 (TODO: can we narrow that down further?)
84 - If ``file`` or ``sharedfile`` was enabled at configure time, add the
85 respective disk template to the list of enabled disk templates.
86 - For disk templates ``diskless``, ``blockdev``, ``ext``, and ``rbd``, we
87 inspect the current cluster configuration regarding whether or not there
88 are instances that use one of those disk templates. We will add only those
89 that are currently in use.
90 The order in which the list of enabled disk templates is built up will be
91 determined by a preference order based on when in the history of Ganeti the
92 disk templates were introduced (thus being a heuristic for which are used
95 The list of enabled disk templates can be specified on cluster initialization
96 with ``gnt-cluster init`` using the optional parameter
97 ``--enabled-disk-templates``. If it is not set, it will be set to a default
98 set of enabled disk templates, which includes the following disk templates:
99 ``drbd`` and ``plain``. The list can be shrunk or extended by
100 ``gnt-cluster modify`` using the same parameter.
105 The storage reporting in ``gnt-node list`` will be the first user of the
106 newly introduced list of enabled disk templates. Currently, storage reporting
107 works only for lvm-based storage. We want to extend that and report storage
108 for the enabled disk templates. The default of ``gnt-node list`` will only
109 report on storage of the default disk template (the first in the list of enabled
110 disk templates). One can explicitly ask for storage reporting on the other
111 enabled disk templates with the ``-o`` option.
113 Some of the currently implemented disk templates share the same base storage
114 technology. Since the storage reporting is based on the underlying technology
115 rather than on the user-facing disk templates, we introduce storage types to
116 represent the underlying technology. There will be a mapping from disk templates
117 to storage types, which will be used by the storage reporting backend to pick
118 the right method for estimating the storage for the different disk templates.
120 The proposed storage types are ``blockdev``, ``diskless``, ``ext``, ``file``,
121 ``lvm-pv``, ``lvm-vg``, ``rados``.
123 The mapping from disk templates to storage types will be: ``drbd`` and ``plain``
124 to ``lvm-vg``, ``file`` and ``sharedfile`` to ``file``, and all others to their
125 obvious counterparts.
127 Note that there is no disk template mapping to ``lvm-pv``, because this storage
128 type is currently only used to enable the user to mark it as (un)allocatable.
129 (See ``man gnt-node``.) It is not possible to create an instance on a storage
130 unit that is of type ``lvm-pv`` directly, therefore it is not included in the
133 The storage reporting for file storage will report space on the file storage
134 dir, which is currently limited to one directory. In the future, if we'll have
135 support for more directories, or for per-nodegroup directories this can be
138 For now, we will implement only the storage reporting for non-shared storage,
139 that is disk templates ``file``, ``lvm``, and ``drbd``. For disk template
140 ``diskless``, there is obviously nothing to report about. When implementing
141 storage reporting for file, we can also use it for ``sharedfile``, since it
142 uses the same file system mechanisms to determine the free space. In the
143 future, we can optimize storage reporting for shared storage by not querying
144 all nodes that use a common shared file for the same space information.
146 In the future, we extend storage reporting for shared storage types like
147 ``rados`` and ``ext``. Note that it will not make sense to query each node for
148 storage reporting on a storage unit that is used by several nodes.
150 We will not implement storage reporting for the ``blockdev`` disk template,
151 because block devices are always adopted after being provided by the system
152 administrator, thus coming from outside Ganeti. There is no point in storage
153 reporting for block devices, because Ganeti will never try to allocate storage
154 inside a block device.
159 The noded RPC call that reports node storage space will be changed to
160 accept a list of <disktemplate>,<key> string tuples. For each of them, it will
161 report the free amount of storage space found on storage <key> as known
162 by the requested disk template. Depending on the disk template, the key would
163 be a volume group name, in case of lvm-based disk templates, a directory name
164 for the file and shared file storage, and a rados pool name for rados storage.
166 Masterd will know through the mapping of disk templates to storage types which
167 storage type uses which mechanism for storage calculation and invoke only the
170 Note that for file and sharedfile the node knows which directories are allowed
171 and won't allow any other directory to be queried for security reasons. The
172 actual path still needs to be passed to distinguish the two, as the type will
173 be the same for both.
175 These calculations will be implemented in the node storage system
176 (currently lib/storage.py) but querying will still happen through the
177 ``node info`` call, to avoid requiring an extra RPC each time.
182 `gnt-node list`` can be queried for the different disk templates, if they
183 are enabled. By default, it will just report information about the default
184 disk template. Examples::
187 Node DTotal DFree MTotal MNode MFree Pinst Sinst
188 mynode1 3.6T 3.6T 64.0G 1023M 62.2G 1 0
189 mynode2 3.6T 3.6T 64.0G 1023M 62.0G 2 1
190 mynode3 3.6T 3.6T 64.0G 1023M 62.3G 0 2
192 > gnt-node list -o dtotal/drbd,dfree/file
193 Node DTotal (drbd, myvg) DFree (file, mydir)
197 Note that for drbd, we only report the space of the vg and only if it was not
198 renamed to something different than the default volume group name. With this
199 design, there is also no possibility to ask about the meta volume group. We
200 restrict the design here to make the transition to storage pools easier (as it
201 is an interim state only). It is the administrator's responsibility to ensure
202 that there is enough space for the meta volume group.
204 When storage pools are implemented, we switch from referencing the disk template
205 to referencing the storage pool name. For that, of course, the pool names need
206 to be unique over all storage types. For drbd, we will use the default 'drbd'
207 storage pool and possibly a second lvm-based storage pool for the metavg. It
208 will be possible to rename storage pools (thus also the default lvm storage
209 pool). There will be new functionality to ask about what storage pools are
210 available and of what type. Storage pools will have a storage pool type which is
211 one of the disk templates. There can be more than one storage pool based on the
212 same disk template, therefore we will then start referencing the storage pool
213 name instead of the disk template.
215 ``gnt-cluster info`` will report which disk templates are enabled, i.e.
216 which ones are supported according to the cluster configuration. Example
223 - enabled disk templates: plain, drbd, sharedfile, rados
226 ``gnt-node list-storage`` will not be affected by any changes, since this design
227 is restricted only to free storage reporting for non-shared storage types.
232 The iallocator protocol doesn't need to change: since we know which
233 disk template an instance has, we'll pass only the "free" value for that
234 disk template to the iallocator, when asking for an allocation to be
235 made. Note that for DRBD nowadays we ignore the case when vg and metavg
236 are different, and we only consider the main volume group. Fixing this is
237 outside the scope of this design.
239 With this design, we ensure forward-compatibility with respect to storage
240 pools. For now, we'll report space for all available disk templates that
241 are based on non-shared storage types, in the future, for all available
247 Hbal will not need changes, as it handles it already. We don't forecast
248 any changes needed to it.
250 Space reporting changes
251 -----------------------
253 Hspace will by default report by assuming the allocation will happen on
254 the default disk template for the cluster/nodegroup. An option will be added
255 to manually specify a different storage.
257 Interactions with Partitioned Ganeti
258 ------------------------------------
260 Also the design for :doc:`Partitioned Ganeti <design-partitioned>` deals
261 with reporting free space. Partitioned Ganeti has a different way to
262 report free space for LVM on nodes where the ``exclusive_storage`` flag
263 is set. That doesn't interact directly with this design, as the specifics
264 of how the free space is computed is not in the scope of this design.
265 But the ``node info`` call contains the value of the
266 ``exclusive_storage`` flag, which is currently only meaningful for the
267 LVM storage type. Additional flags like the ``exclusive_storage`` flag
268 for lvm might be useful for other disk templates / storage types as well.
269 We therefore extend the RPC call with <disktemplate>,<key> to
270 <disktemplate>,<key>,<params> to include any disk-template-specific
271 (or storage-type specific) parameters in the RPC call.
273 The reporting of free spindles, also part of Partitioned Ganeti, is not
274 concerned with this design doc, as those are seen as a separate resource.
276 .. vim: set textwidth=72 :
projects / ganeti-local / blob