]>
Commit | Line | Data |
---|---|---|
80c0adcb | 1 | [[chapter_pveceph]] |
0840a663 | 2 | ifdef::manvolnum[] |
b2f242ab DM |
3 | pveceph(1) |
4 | ========== | |
404a158e | 5 | :pve-toplevel: |
0840a663 DM |
6 | |
7 | NAME | |
8 | ---- | |
9 | ||
21394e70 | 10 | pveceph - Manage Ceph Services on Proxmox VE Nodes |
0840a663 | 11 | |
49a5e11c | 12 | SYNOPSIS |
0840a663 DM |
13 | -------- |
14 | ||
15 | include::pveceph.1-synopsis.adoc[] | |
16 | ||
17 | DESCRIPTION | |
18 | ----------- | |
19 | endif::manvolnum[] | |
0840a663 | 20 | ifndef::manvolnum[] |
fe93f133 DM |
21 | Manage Ceph Services on Proxmox VE Nodes |
22 | ======================================== | |
49d3ad91 | 23 | :pve-toplevel: |
0840a663 DM |
24 | endif::manvolnum[] |
25 | ||
8997dd6e DM |
26 | [thumbnail="gui-ceph-status.png"] |
27 | ||
a474ca1f AA |
28 | {pve} unifies your compute and storage systems, i.e. you can use the same |
29 | physical nodes within a cluster for both computing (processing VMs and | |
30 | containers) and replicated storage. The traditional silos of compute and | |
31 | storage resources can be wrapped up into a single hyper-converged appliance. | |
32 | Separate storage networks (SANs) and connections via network attached storages | |
33 | (NAS) disappear. With the integration of Ceph, an open source software-defined | |
34 | storage platform, {pve} has the ability to run and manage Ceph storage directly | |
35 | on the hypervisor nodes. | |
c994e4e5 DM |
36 | |
37 | Ceph is a distributed object store and file system designed to provide | |
1d54c3b4 AA |
38 | excellent performance, reliability and scalability. |
39 | ||
04ba9b24 TL |
40 | .Some advantages of Ceph on {pve} are: |
41 | - Easy setup and management with CLI and GUI support | |
a474ca1f AA |
42 | - Thin provisioning |
43 | - Snapshots support | |
44 | - Self healing | |
a474ca1f AA |
45 | - Scalable to the exabyte level |
46 | - Setup pools with different performance and redundancy characteristics | |
47 | - Data is replicated, making it fault tolerant | |
48 | - Runs on economical commodity hardware | |
49 | - No need for hardware RAID controllers | |
a474ca1f AA |
50 | - Open source |
51 | ||
1d54c3b4 AA |
52 | For small to mid sized deployments, it is possible to install a Ceph server for |
53 | RADOS Block Devices (RBD) directly on your {pve} cluster nodes, see | |
c994e4e5 DM |
54 | xref:ceph_rados_block_devices[Ceph RADOS Block Devices (RBD)]. Recent |
55 | hardware has plenty of CPU power and RAM, so running storage services | |
56 | and VMs on the same node is possible. | |
21394e70 DM |
57 | |
58 | To simplify management, we provide 'pveceph' - a tool to install and | |
59 | manage {ceph} services on {pve} nodes. | |
60 | ||
a474ca1f | 61 | .Ceph consists of a couple of Daemons footnote:[Ceph intro http://docs.ceph.com/docs/master/start/intro/], for use as a RBD storage: |
1d54c3b4 AA |
62 | - Ceph Monitor (ceph-mon) |
63 | - Ceph Manager (ceph-mgr) | |
64 | - Ceph OSD (ceph-osd; Object Storage Daemon) | |
65 | ||
66 | TIP: We recommend to get familiar with the Ceph vocabulary. | |
67 | footnote:[Ceph glossary http://docs.ceph.com/docs/luminous/glossary] | |
68 | ||
21394e70 DM |
69 | |
70 | Precondition | |
71 | ------------ | |
72 | ||
c994e4e5 DM |
73 | To build a Proxmox Ceph Cluster there should be at least three (preferably) |
74 | identical servers for the setup. | |
21394e70 | 75 | |
a474ca1f AA |
76 | A 10Gb network, exclusively used for Ceph, is recommended. A meshed network |
77 | setup is also an option if there are no 10Gb switches available, see our wiki | |
78 | article footnote:[Full Mesh Network for Ceph {webwiki-url}Full_Mesh_Network_for_Ceph_Server] . | |
21394e70 DM |
79 | |
80 | Check also the recommendations from | |
1d54c3b4 | 81 | http://docs.ceph.com/docs/luminous/start/hardware-recommendations/[Ceph's website]. |
21394e70 | 82 | |
a474ca1f AA |
83 | .Avoid RAID |
84 | While RAID controller are build for storage virtualisation, to combine | |
85 | independent disks to form one or more logical units. Their caching methods, | |
86 | algorithms (RAID modes; incl. JBOD), disk or write/read optimisations are | |
87 | targeted towards aforementioned logical units and not to Ceph. | |
88 | ||
89 | WARNING: Avoid RAID controller, use host bus adapter (HBA) instead. | |
90 | ||
21394e70 DM |
91 | |
92 | Installation of Ceph Packages | |
93 | ----------------------------- | |
94 | ||
95 | On each node run the installation script as follows: | |
96 | ||
97 | [source,bash] | |
98 | ---- | |
19920184 | 99 | pveceph install |
21394e70 DM |
100 | ---- |
101 | ||
102 | This sets up an `apt` package repository in | |
103 | `/etc/apt/sources.list.d/ceph.list` and installs the required software. | |
104 | ||
105 | ||
106 | Creating initial Ceph configuration | |
107 | ----------------------------------- | |
108 | ||
8997dd6e DM |
109 | [thumbnail="gui-ceph-config.png"] |
110 | ||
21394e70 DM |
111 | After installation of packages, you need to create an initial Ceph |
112 | configuration on just one node, based on your network (`10.10.10.0/24` | |
113 | in the following example) dedicated for Ceph: | |
114 | ||
115 | [source,bash] | |
116 | ---- | |
117 | pveceph init --network 10.10.10.0/24 | |
118 | ---- | |
119 | ||
a474ca1f | 120 | This creates an initial configuration at `/etc/pve/ceph.conf`. That file is |
c994e4e5 | 121 | automatically distributed to all {pve} nodes by using |
21394e70 DM |
122 | xref:chapter_pmxcfs[pmxcfs]. The command also creates a symbolic link |
123 | from `/etc/ceph/ceph.conf` pointing to that file. So you can simply run | |
124 | Ceph commands without the need to specify a configuration file. | |
125 | ||
126 | ||
d9a27ee1 | 127 | [[pve_ceph_monitors]] |
21394e70 DM |
128 | Creating Ceph Monitors |
129 | ---------------------- | |
130 | ||
8997dd6e DM |
131 | [thumbnail="gui-ceph-monitor.png"] |
132 | ||
1d54c3b4 AA |
133 | The Ceph Monitor (MON) |
134 | footnote:[Ceph Monitor http://docs.ceph.com/docs/luminous/start/intro/] | |
a474ca1f AA |
135 | maintains a master copy of the cluster map. For high availability you need to |
136 | have at least 3 monitors. | |
1d54c3b4 AA |
137 | |
138 | On each node where you want to place a monitor (three monitors are recommended), | |
139 | create it by using the 'Ceph -> Monitor' tab in the GUI or run. | |
21394e70 DM |
140 | |
141 | ||
142 | [source,bash] | |
143 | ---- | |
144 | pveceph createmon | |
145 | ---- | |
146 | ||
1d54c3b4 AA |
147 | This will also install the needed Ceph Manager ('ceph-mgr') by default. If you |
148 | do not want to install a manager, specify the '-exclude-manager' option. | |
149 | ||
150 | ||
151 | [[pve_ceph_manager]] | |
152 | Creating Ceph Manager | |
153 | ---------------------- | |
154 | ||
a474ca1f | 155 | The Manager daemon runs alongside the monitors, providing an interface for |
1d54c3b4 AA |
156 | monitoring the cluster. Since the Ceph luminous release the |
157 | ceph-mgr footnote:[Ceph Manager http://docs.ceph.com/docs/luminous/mgr/] daemon | |
158 | is required. During monitor installation the ceph manager will be installed as | |
159 | well. | |
160 | ||
161 | NOTE: It is recommended to install the Ceph Manager on the monitor nodes. For | |
162 | high availability install more then one manager. | |
163 | ||
164 | [source,bash] | |
165 | ---- | |
166 | pveceph createmgr | |
167 | ---- | |
168 | ||
21394e70 | 169 | |
d9a27ee1 | 170 | [[pve_ceph_osds]] |
21394e70 DM |
171 | Creating Ceph OSDs |
172 | ------------------ | |
173 | ||
8997dd6e DM |
174 | [thumbnail="gui-ceph-osd-status.png"] |
175 | ||
21394e70 DM |
176 | via GUI or via CLI as follows: |
177 | ||
178 | [source,bash] | |
179 | ---- | |
180 | pveceph createosd /dev/sd[X] | |
181 | ---- | |
182 | ||
1d54c3b4 AA |
183 | TIP: We recommend a Ceph cluster size, starting with 12 OSDs, distributed evenly |
184 | among your, at least three nodes (4 OSDs on each node). | |
185 | ||
a474ca1f AA |
186 | If the disk was used before (eg. ZFS/RAID/OSD), to remove partition table, boot |
187 | sector and any OSD leftover the following commands should be sufficient. | |
188 | ||
189 | [source,bash] | |
190 | ---- | |
191 | dd if=/dev/zero of=/dev/sd[X] bs=1M count=200 | |
192 | ceph-disk zap /dev/sd[X] | |
193 | ---- | |
194 | ||
195 | WARNING: The above commands will destroy data on the disk! | |
1d54c3b4 AA |
196 | |
197 | Ceph Bluestore | |
198 | ~~~~~~~~~~~~~~ | |
21394e70 | 199 | |
1d54c3b4 AA |
200 | Starting with the Ceph Kraken release, a new Ceph OSD storage type was |
201 | introduced, the so called Bluestore | |
a474ca1f AA |
202 | footnote:[Ceph Bluestore http://ceph.com/community/new-luminous-bluestore/]. |
203 | This is the default when creating OSDs in Ceph luminous. | |
21394e70 DM |
204 | |
205 | [source,bash] | |
206 | ---- | |
1d54c3b4 AA |
207 | pveceph createosd /dev/sd[X] |
208 | ---- | |
209 | ||
210 | NOTE: In order to select a disk in the GUI, to be more failsafe, the disk needs | |
a474ca1f AA |
211 | to have a GPT footnoteref:[GPT, GPT partition table |
212 | https://en.wikipedia.org/wiki/GUID_Partition_Table] partition table. You can | |
213 | create this with `gdisk /dev/sd(x)`. If there is no GPT, you cannot select the | |
214 | disk as DB/WAL. | |
1d54c3b4 AA |
215 | |
216 | If you want to use a separate DB/WAL device for your OSDs, you can specify it | |
a474ca1f AA |
217 | through the '-journal_dev' option. The WAL is placed with the DB, if not |
218 | specified separately. | |
1d54c3b4 AA |
219 | |
220 | [source,bash] | |
221 | ---- | |
a474ca1f | 222 | pveceph createosd /dev/sd[X] -journal_dev /dev/sd[Y] |
1d54c3b4 AA |
223 | ---- |
224 | ||
225 | NOTE: The DB stores BlueStore’s internal metadata and the WAL is BlueStore’s | |
226 | internal journal or write-ahead log. It is recommended to use a fast SSDs or | |
227 | NVRAM for better performance. | |
228 | ||
229 | ||
230 | Ceph Filestore | |
231 | ~~~~~~~~~~~~~ | |
232 | Till Ceph luminous, Filestore was used as storage type for Ceph OSDs. It can | |
233 | still be used and might give better performance in small setups, when backed by | |
234 | a NVMe SSD or similar. | |
235 | ||
236 | [source,bash] | |
237 | ---- | |
238 | pveceph createosd /dev/sd[X] -bluestore 0 | |
239 | ---- | |
240 | ||
241 | NOTE: In order to select a disk in the GUI, the disk needs to have a | |
242 | GPT footnoteref:[GPT] partition table. You can | |
243 | create this with `gdisk /dev/sd(x)`. If there is no GPT, you cannot select the | |
244 | disk as journal. Currently the journal size is fixed to 5 GB. | |
245 | ||
246 | If you want to use a dedicated SSD journal disk: | |
247 | ||
248 | [source,bash] | |
249 | ---- | |
e677b344 | 250 | pveceph createosd /dev/sd[X] -journal_dev /dev/sd[Y] -bluestore 0 |
21394e70 DM |
251 | ---- |
252 | ||
253 | Example: Use /dev/sdf as data disk (4TB) and /dev/sdb is the dedicated SSD | |
254 | journal disk. | |
255 | ||
256 | [source,bash] | |
257 | ---- | |
e677b344 | 258 | pveceph createosd /dev/sdf -journal_dev /dev/sdb -bluestore 0 |
21394e70 DM |
259 | ---- |
260 | ||
261 | This partitions the disk (data and journal partition), creates | |
262 | filesystems and starts the OSD, afterwards it is running and fully | |
1d54c3b4 | 263 | functional. |
21394e70 | 264 | |
1d54c3b4 AA |
265 | NOTE: This command refuses to initialize disk when it detects existing data. So |
266 | if you want to overwrite a disk you should remove existing data first. You can | |
267 | do that using: 'ceph-disk zap /dev/sd[X]' | |
21394e70 DM |
268 | |
269 | You can create OSDs containing both journal and data partitions or you | |
270 | can place the journal on a dedicated SSD. Using a SSD journal disk is | |
1d54c3b4 | 271 | highly recommended to achieve good performance. |
21394e70 DM |
272 | |
273 | ||
07fef357 | 274 | [[pve_ceph_pools]] |
1d54c3b4 AA |
275 | Creating Ceph Pools |
276 | ------------------- | |
21394e70 | 277 | |
8997dd6e DM |
278 | [thumbnail="gui-ceph-pools.png"] |
279 | ||
1d54c3b4 AA |
280 | A pool is a logical group for storing objects. It holds **P**lacement |
281 | **G**roups (PG), a collection of objects. | |
282 | ||
283 | When no options are given, we set a | |
284 | default of **64 PGs**, a **size of 3 replicas** and a **min_size of 2 replicas** | |
285 | for serving objects in a degraded state. | |
286 | ||
287 | NOTE: The default number of PGs works for 2-6 disks. Ceph throws a | |
288 | "HEALTH_WARNING" if you have too few or too many PGs in your cluster. | |
289 | ||
290 | It is advised to calculate the PG number depending on your setup, you can find | |
a474ca1f AA |
291 | the formula and the PG calculator footnote:[PG calculator |
292 | http://ceph.com/pgcalc/] online. While PGs can be increased later on, they can | |
293 | never be decreased. | |
1d54c3b4 AA |
294 | |
295 | ||
296 | You can create pools through command line or on the GUI on each PVE host under | |
297 | **Ceph -> Pools**. | |
298 | ||
299 | [source,bash] | |
300 | ---- | |
301 | pveceph createpool <name> | |
302 | ---- | |
303 | ||
304 | If you would like to automatically get also a storage definition for your pool, | |
305 | active the checkbox "Add storages" on the GUI or use the command line option | |
306 | '--add_storages' on pool creation. | |
21394e70 | 307 | |
1d54c3b4 AA |
308 | Further information on Ceph pool handling can be found in the Ceph pool |
309 | operation footnote:[Ceph pool operation | |
310 | http://docs.ceph.com/docs/luminous/rados/operations/pools/] | |
311 | manual. | |
21394e70 | 312 | |
9fad507d AA |
313 | Ceph CRUSH & device classes |
314 | --------------------------- | |
315 | The foundation of Ceph is its algorithm, **C**ontrolled **R**eplication | |
316 | **U**nder **S**calable **H**ashing | |
317 | (CRUSH footnote:[CRUSH https://ceph.com/wp-content/uploads/2016/08/weil-crush-sc06.pdf]). | |
318 | ||
319 | CRUSH calculates where to store to and retrieve data from, this has the | |
320 | advantage that no central index service is needed. CRUSH works with a map of | |
321 | OSDs, buckets (device locations) and rulesets (data replication) for pools. | |
322 | ||
323 | NOTE: Further information can be found in the Ceph documentation, under the | |
324 | section CRUSH map footnote:[CRUSH map http://docs.ceph.com/docs/luminous/rados/operations/crush-map/]. | |
325 | ||
326 | This map can be altered to reflect different replication hierarchies. The object | |
327 | replicas can be separated (eg. failure domains), while maintaining the desired | |
328 | distribution. | |
329 | ||
330 | A common use case is to use different classes of disks for different Ceph pools. | |
331 | For this reason, Ceph introduced the device classes with luminous, to | |
332 | accommodate the need for easy ruleset generation. | |
333 | ||
334 | The device classes can be seen in the 'ceph osd tree' output. These classes | |
335 | represent their own root bucket, which can be seen with the below command. | |
336 | ||
337 | [source, bash] | |
338 | ---- | |
339 | ceph osd crush tree --show-shadow | |
340 | ---- | |
341 | ||
342 | Example output form the above command: | |
343 | ||
344 | [source, bash] | |
345 | ---- | |
346 | ID CLASS WEIGHT TYPE NAME | |
347 | -16 nvme 2.18307 root default~nvme | |
348 | -13 nvme 0.72769 host sumi1~nvme | |
349 | 12 nvme 0.72769 osd.12 | |
350 | -14 nvme 0.72769 host sumi2~nvme | |
351 | 13 nvme 0.72769 osd.13 | |
352 | -15 nvme 0.72769 host sumi3~nvme | |
353 | 14 nvme 0.72769 osd.14 | |
354 | -1 7.70544 root default | |
355 | -3 2.56848 host sumi1 | |
356 | 12 nvme 0.72769 osd.12 | |
357 | -5 2.56848 host sumi2 | |
358 | 13 nvme 0.72769 osd.13 | |
359 | -7 2.56848 host sumi3 | |
360 | 14 nvme 0.72769 osd.14 | |
361 | ---- | |
362 | ||
363 | To let a pool distribute its objects only on a specific device class, you need | |
364 | to create a ruleset with the specific class first. | |
365 | ||
366 | [source, bash] | |
367 | ---- | |
368 | ceph osd crush rule create-replicated <rule-name> <root> <failure-domain> <class> | |
369 | ---- | |
370 | ||
371 | [frame="none",grid="none", align="left", cols="30%,70%"] | |
372 | |=== | |
373 | |<rule-name>|name of the rule, to connect with a pool (seen in GUI & CLI) | |
374 | |<root>|which crush root it should belong to (default ceph root "default") | |
375 | |<failure-domain>|at which failure-domain the objects should be distributed (usually host) | |
376 | |<class>|what type of OSD backing store to use (eg. nvme, ssd, hdd) | |
377 | |=== | |
378 | ||
379 | Once the rule is in the CRUSH map, you can tell a pool to use the ruleset. | |
380 | ||
381 | [source, bash] | |
382 | ---- | |
383 | ceph osd pool set <pool-name> crush_rule <rule-name> | |
384 | ---- | |
385 | ||
386 | TIP: If the pool already contains objects, all of these have to be moved | |
387 | accordingly. Depending on your setup this may introduce a big performance hit on | |
388 | your cluster. As an alternative, you can create a new pool and move disks | |
389 | separately. | |
390 | ||
391 | ||
21394e70 DM |
392 | Ceph Client |
393 | ----------- | |
394 | ||
8997dd6e DM |
395 | [thumbnail="gui-ceph-log.png"] |
396 | ||
21394e70 DM |
397 | You can then configure {pve} to use such pools to store VM or |
398 | Container images. Simply use the GUI too add a new `RBD` storage (see | |
399 | section xref:ceph_rados_block_devices[Ceph RADOS Block Devices (RBD)]). | |
400 | ||
1d54c3b4 AA |
401 | You also need to copy the keyring to a predefined location for a external Ceph |
402 | cluster. If Ceph is installed on the Proxmox nodes itself, then this will be | |
403 | done automatically. | |
21394e70 DM |
404 | |
405 | NOTE: The file name needs to be `<storage_id> + `.keyring` - `<storage_id>` is | |
406 | the expression after 'rbd:' in `/etc/pve/storage.cfg` which is | |
407 | `my-ceph-storage` in the following example: | |
408 | ||
409 | [source,bash] | |
410 | ---- | |
411 | mkdir /etc/pve/priv/ceph | |
412 | cp /etc/ceph/ceph.client.admin.keyring /etc/pve/priv/ceph/my-ceph-storage.keyring | |
413 | ---- | |
0840a663 DM |
414 | |
415 | ||
416 | ifdef::manvolnum[] | |
417 | include::pve-copyright.adoc[] | |
418 | endif::manvolnum[] |