]>
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 | ||
1ff5e4e8 | 26 | [thumbnail="screenshot/gui-ceph-status.png"] |
8997dd6e | 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 | ||
127ca409 | 61 | .Ceph consists of a couple of Daemons footnote:[Ceph intro http://docs.ceph.com/docs/luminous/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 | ||
477fbcfb AA |
66 | TIP: We highly recommend to get familiar with Ceph's architecture |
67 | footnote:[Ceph architecture http://docs.ceph.com/docs/luminous/architecture/] | |
68 | and vocabulary | |
69 | footnote:[Ceph glossary http://docs.ceph.com/docs/luminous/glossary]. | |
1d54c3b4 | 70 | |
21394e70 DM |
71 | |
72 | Precondition | |
73 | ------------ | |
74 | ||
76f6eca4 AA |
75 | To build a hyper-converged Proxmox + Ceph Cluster there should be at least |
76 | three (preferably) identical servers for the setup. | |
21394e70 DM |
77 | |
78 | Check also the recommendations from | |
1d54c3b4 | 79 | http://docs.ceph.com/docs/luminous/start/hardware-recommendations/[Ceph's website]. |
21394e70 | 80 | |
76f6eca4 | 81 | .CPU |
2f19a6b0 TL |
82 | Higher CPU core frequency reduce latency and should be preferred. As a simple |
83 | rule of thumb, you should assign a CPU core (or thread) to each Ceph service to | |
84 | provide enough resources for stable and durable Ceph performance. | |
76f6eca4 AA |
85 | |
86 | .Memory | |
87 | Especially in a hyper-converged setup, the memory consumption needs to be | |
2f19a6b0 TL |
88 | carefully monitored. In addition to the intended workload from virtual machines |
89 | and container, Ceph needs enough memory available to provide good and stable | |
90 | performance. As a rule of thumb, for roughly 1 TiB of data, 1 GiB of memory | |
91 | will be used by an OSD. OSD caching will use additional memory. | |
76f6eca4 AA |
92 | |
93 | .Network | |
94 | We recommend a network bandwidth of at least 10 GbE or more, which is used | |
95 | exclusively for Ceph. A meshed network setup | |
96 | footnote:[Full Mesh Network for Ceph {webwiki-url}Full_Mesh_Network_for_Ceph_Server] | |
97 | is also an option if there are no 10 GbE switches available. | |
98 | ||
2f19a6b0 TL |
99 | The volume of traffic, especially during recovery, will interfere with other |
100 | services on the same network and may even break the {pve} cluster stack. | |
76f6eca4 AA |
101 | |
102 | Further, estimate your bandwidth needs. While one HDD might not saturate a 1 Gb | |
2f19a6b0 TL |
103 | link, multiple HDD OSDs per node can, and modern NVMe SSDs will even saturate |
104 | 10 Gbps of bandwidth quickly. Deploying a network capable of even more bandwith | |
105 | will ensure that it isn't your bottleneck and won't be anytime soon, 25, 40 or | |
106 | even 100 GBps are possible. | |
76f6eca4 AA |
107 | |
108 | .Disks | |
109 | When planning the size of your Ceph cluster, it is important to take the | |
110 | recovery time into consideration. Especially with small clusters, the recovery | |
111 | might take long. It is recommended that you use SSDs instead of HDDs in small | |
112 | setups to reduce recovery time, minimizing the likelihood of a subsequent | |
113 | failure event during recovery. | |
114 | ||
2f19a6b0 | 115 | In general SSDs will provide more IOPs than spinning disks. This fact and the |
76f6eca4 | 116 | higher cost may make a xref:pve_ceph_device_classes[class based] separation of |
2f19a6b0 | 117 | pools appealing. Another possibility to speedup OSDs is to use a faster disk |
76f6eca4 AA |
118 | as journal or DB/WAL device, see xref:pve_ceph_osds[creating Ceph OSDs]. If a |
119 | faster disk is used for multiple OSDs, a proper balance between OSD and WAL / | |
120 | DB (or journal) disk must be selected, otherwise the faster disk becomes the | |
121 | bottleneck for all linked OSDs. | |
122 | ||
123 | Aside from the disk type, Ceph best performs with an even sized and distributed | |
2f19a6b0 TL |
124 | amount of disks per node. For example, 4 x 500 GB disks with in each node is |
125 | better than a mixed setup with a single 1 TB and three 250 GB disk. | |
126 | ||
127 | One also need to balance OSD count and single OSD capacity. More capacity | |
128 | allows to increase storage density, but it also means that a single OSD | |
129 | failure forces ceph to recover more data at once. | |
76f6eca4 | 130 | |
a474ca1f | 131 | .Avoid RAID |
86be506d | 132 | As Ceph handles data object redundancy and multiple parallel writes to disks |
c78756be | 133 | (OSDs) on its own, using a RAID controller normally doesn’t improve |
86be506d TL |
134 | performance or availability. On the contrary, Ceph is designed to handle whole |
135 | disks on it's own, without any abstraction in between. RAID controller are not | |
136 | designed for the Ceph use case and may complicate things and sometimes even | |
137 | reduce performance, as their write and caching algorithms may interfere with | |
138 | the ones from Ceph. | |
a474ca1f AA |
139 | |
140 | WARNING: Avoid RAID controller, use host bus adapter (HBA) instead. | |
141 | ||
76f6eca4 | 142 | NOTE: Above recommendations should be seen as a rough guidance for choosing |
2f19a6b0 TL |
143 | hardware. Therefore, it is still essential to adapt it to your specific needs, |
144 | test your setup and monitor health and performance continuously. | |
76f6eca4 | 145 | |
2394c306 TM |
146 | [[pve_ceph_install_wizard]] |
147 | Initial Ceph installation & configuration | |
148 | ----------------------------------------- | |
149 | ||
150 | [thumbnail="screenshot/gui-node-ceph-install.png"] | |
151 | ||
152 | With {pve} you have the benefit of an easy to use installation wizard | |
153 | for Ceph. Click on one of your cluster nodes and navigate to the Ceph | |
154 | section in the menu tree. If Ceph is not installed already you will be | |
155 | offered to do this now. | |
156 | ||
157 | The wizard is divided into different sections, where each needs to be | |
158 | done successfully in order to use Ceph. After starting the installation | |
159 | the wizard will load and install all required packages. | |
160 | ||
161 | After finishing the first step, you will need to create a configuration. | |
162 | This step is only needed on the first run of the wizard, because the | |
163 | configuration is cluster wide and therefore distributed automatically | |
164 | to all remaining cluster members - see xref:chapter_pmxcfs[cluster file system (pmxcfs)] section. | |
165 | ||
166 | The configuration step includes the following settings: | |
167 | ||
168 | * *Public Network:* You should setup a dedicated network for Ceph, this | |
169 | setting is required. Separating your Ceph traffic is highly recommended, | |
170 | because it could lead to troubles with other latency dependent services | |
171 | e.g. cluster communication. | |
172 | ||
173 | [thumbnail="screenshot/gui-node-ceph-install-wizard-step2.png"] | |
174 | ||
175 | * *Cluster Network:* As an optional step you can go even further and | |
176 | separate the xref:pve_ceph_osds[OSD] replication & heartbeat traffic | |
177 | as well. This will relieve the public network and could lead to | |
178 | significant performance improvements especially in big clusters. | |
179 | ||
180 | You have two more options which are considered advanced and therefore | |
181 | should only changed if you are an expert. | |
182 | ||
183 | * *Number of replicas*: Defines the how often a object is replicated | |
184 | * *Minimum replicas*: Defines the minimum number of required replicas | |
185 | for I/O. | |
186 | ||
187 | Additionally you need to choose a monitor node, this is required. | |
188 | ||
189 | That's it, you should see a success page as the last step with further | |
190 | instructions on how to go on. You are now prepared to start using Ceph, | |
191 | even though you will need to create additional xref:pve_ceph_monitors[monitors], | |
192 | create some xref:pve_ceph_osds[OSDs] and at least one xref:pve_ceph_pools[pool]. | |
193 | ||
194 | The rest of this chapter will guide you on how to get the most out of | |
195 | your {pve} based Ceph setup, this will include aforementioned and | |
196 | more like xref:pveceph_fs[CephFS] which is a very handy addition to your | |
197 | new Ceph cluster. | |
21394e70 | 198 | |
58f95dd7 | 199 | [[pve_ceph_install]] |
21394e70 DM |
200 | Installation of Ceph Packages |
201 | ----------------------------- | |
2394c306 TM |
202 | Use {pve} Ceph installation wizard (recommended) or run the following |
203 | command on each node: | |
21394e70 DM |
204 | |
205 | [source,bash] | |
206 | ---- | |
19920184 | 207 | pveceph install |
21394e70 DM |
208 | ---- |
209 | ||
210 | This sets up an `apt` package repository in | |
211 | `/etc/apt/sources.list.d/ceph.list` and installs the required software. | |
212 | ||
213 | ||
214 | Creating initial Ceph configuration | |
215 | ----------------------------------- | |
216 | ||
1ff5e4e8 | 217 | [thumbnail="screenshot/gui-ceph-config.png"] |
8997dd6e | 218 | |
2394c306 TM |
219 | Use the {pve} Ceph installation wizard (recommended) or run the |
220 | following command on one node: | |
21394e70 DM |
221 | |
222 | [source,bash] | |
223 | ---- | |
224 | pveceph init --network 10.10.10.0/24 | |
225 | ---- | |
226 | ||
2394c306 TM |
227 | This creates an initial configuration at `/etc/pve/ceph.conf` with a |
228 | dedicated network for ceph. That file is automatically distributed to | |
229 | all {pve} nodes by using xref:chapter_pmxcfs[pmxcfs]. The command also | |
230 | creates a symbolic link from `/etc/ceph/ceph.conf` pointing to that file. | |
231 | So you can simply run Ceph commands without the need to specify a | |
232 | configuration file. | |
21394e70 DM |
233 | |
234 | ||
d9a27ee1 | 235 | [[pve_ceph_monitors]] |
21394e70 DM |
236 | Creating Ceph Monitors |
237 | ---------------------- | |
238 | ||
1ff5e4e8 | 239 | [thumbnail="screenshot/gui-ceph-monitor.png"] |
8997dd6e | 240 | |
1d54c3b4 AA |
241 | The Ceph Monitor (MON) |
242 | footnote:[Ceph Monitor http://docs.ceph.com/docs/luminous/start/intro/] | |
a474ca1f | 243 | maintains a master copy of the cluster map. For high availability you need to |
2394c306 TM |
244 | have at least 3 monitors. One monitor will already be installed if you |
245 | used the installation wizard. You wont need more than 3 monitors as long | |
246 | as your cluster is small to midsize, only really large clusters will | |
247 | need more than that. | |
1d54c3b4 AA |
248 | |
249 | On each node where you want to place a monitor (three monitors are recommended), | |
250 | create it by using the 'Ceph -> Monitor' tab in the GUI or run. | |
21394e70 DM |
251 | |
252 | ||
253 | [source,bash] | |
254 | ---- | |
255 | pveceph createmon | |
256 | ---- | |
257 | ||
1d54c3b4 AA |
258 | This will also install the needed Ceph Manager ('ceph-mgr') by default. If you |
259 | do not want to install a manager, specify the '-exclude-manager' option. | |
260 | ||
261 | ||
262 | [[pve_ceph_manager]] | |
263 | Creating Ceph Manager | |
264 | ---------------------- | |
265 | ||
a474ca1f | 266 | The Manager daemon runs alongside the monitors, providing an interface for |
1d54c3b4 AA |
267 | monitoring the cluster. Since the Ceph luminous release the |
268 | ceph-mgr footnote:[Ceph Manager http://docs.ceph.com/docs/luminous/mgr/] daemon | |
269 | is required. During monitor installation the ceph manager will be installed as | |
270 | well. | |
271 | ||
272 | NOTE: It is recommended to install the Ceph Manager on the monitor nodes. For | |
273 | high availability install more then one manager. | |
274 | ||
275 | [source,bash] | |
276 | ---- | |
277 | pveceph createmgr | |
278 | ---- | |
279 | ||
21394e70 | 280 | |
d9a27ee1 | 281 | [[pve_ceph_osds]] |
21394e70 DM |
282 | Creating Ceph OSDs |
283 | ------------------ | |
284 | ||
1ff5e4e8 | 285 | [thumbnail="screenshot/gui-ceph-osd-status.png"] |
8997dd6e | 286 | |
21394e70 DM |
287 | via GUI or via CLI as follows: |
288 | ||
289 | [source,bash] | |
290 | ---- | |
291 | pveceph createosd /dev/sd[X] | |
292 | ---- | |
293 | ||
1d54c3b4 AA |
294 | TIP: We recommend a Ceph cluster size, starting with 12 OSDs, distributed evenly |
295 | among your, at least three nodes (4 OSDs on each node). | |
296 | ||
a474ca1f AA |
297 | If the disk was used before (eg. ZFS/RAID/OSD), to remove partition table, boot |
298 | sector and any OSD leftover the following commands should be sufficient. | |
299 | ||
300 | [source,bash] | |
301 | ---- | |
302 | dd if=/dev/zero of=/dev/sd[X] bs=1M count=200 | |
303 | ceph-disk zap /dev/sd[X] | |
304 | ---- | |
305 | ||
306 | WARNING: The above commands will destroy data on the disk! | |
1d54c3b4 AA |
307 | |
308 | Ceph Bluestore | |
309 | ~~~~~~~~~~~~~~ | |
21394e70 | 310 | |
1d54c3b4 AA |
311 | Starting with the Ceph Kraken release, a new Ceph OSD storage type was |
312 | introduced, the so called Bluestore | |
a474ca1f AA |
313 | footnote:[Ceph Bluestore http://ceph.com/community/new-luminous-bluestore/]. |
314 | This is the default when creating OSDs in Ceph luminous. | |
21394e70 DM |
315 | |
316 | [source,bash] | |
317 | ---- | |
1d54c3b4 AA |
318 | pveceph createosd /dev/sd[X] |
319 | ---- | |
320 | ||
ee4a0e96 | 321 | NOTE: In order to select a disk in the GUI, to be more fail-safe, the disk needs |
a474ca1f AA |
322 | to have a GPT footnoteref:[GPT, GPT partition table |
323 | https://en.wikipedia.org/wiki/GUID_Partition_Table] partition table. You can | |
324 | create this with `gdisk /dev/sd(x)`. If there is no GPT, you cannot select the | |
325 | disk as DB/WAL. | |
1d54c3b4 AA |
326 | |
327 | If you want to use a separate DB/WAL device for your OSDs, you can specify it | |
a474ca1f AA |
328 | through the '-journal_dev' option. The WAL is placed with the DB, if not |
329 | specified separately. | |
1d54c3b4 AA |
330 | |
331 | [source,bash] | |
332 | ---- | |
a474ca1f | 333 | pveceph createosd /dev/sd[X] -journal_dev /dev/sd[Y] |
1d54c3b4 AA |
334 | ---- |
335 | ||
336 | NOTE: The DB stores BlueStore’s internal metadata and the WAL is BlueStore’s | |
ee4a0e96 | 337 | internal journal or write-ahead log. It is recommended to use a fast SSD or |
1d54c3b4 AA |
338 | NVRAM for better performance. |
339 | ||
340 | ||
341 | Ceph Filestore | |
342 | ~~~~~~~~~~~~~ | |
343 | Till Ceph luminous, Filestore was used as storage type for Ceph OSDs. It can | |
344 | still be used and might give better performance in small setups, when backed by | |
ee4a0e96 | 345 | an NVMe SSD or similar. |
1d54c3b4 AA |
346 | |
347 | [source,bash] | |
348 | ---- | |
349 | pveceph createosd /dev/sd[X] -bluestore 0 | |
350 | ---- | |
351 | ||
352 | NOTE: In order to select a disk in the GUI, the disk needs to have a | |
353 | GPT footnoteref:[GPT] partition table. You can | |
354 | create this with `gdisk /dev/sd(x)`. If there is no GPT, you cannot select the | |
355 | disk as journal. Currently the journal size is fixed to 5 GB. | |
356 | ||
357 | If you want to use a dedicated SSD journal disk: | |
358 | ||
359 | [source,bash] | |
360 | ---- | |
e677b344 | 361 | pveceph createosd /dev/sd[X] -journal_dev /dev/sd[Y] -bluestore 0 |
21394e70 DM |
362 | ---- |
363 | ||
364 | Example: Use /dev/sdf as data disk (4TB) and /dev/sdb is the dedicated SSD | |
365 | journal disk. | |
366 | ||
367 | [source,bash] | |
368 | ---- | |
e677b344 | 369 | pveceph createosd /dev/sdf -journal_dev /dev/sdb -bluestore 0 |
21394e70 DM |
370 | ---- |
371 | ||
372 | This partitions the disk (data and journal partition), creates | |
373 | filesystems and starts the OSD, afterwards it is running and fully | |
1d54c3b4 | 374 | functional. |
21394e70 | 375 | |
1d54c3b4 AA |
376 | NOTE: This command refuses to initialize disk when it detects existing data. So |
377 | if you want to overwrite a disk you should remove existing data first. You can | |
378 | do that using: 'ceph-disk zap /dev/sd[X]' | |
21394e70 DM |
379 | |
380 | You can create OSDs containing both journal and data partitions or you | |
381 | can place the journal on a dedicated SSD. Using a SSD journal disk is | |
1d54c3b4 | 382 | highly recommended to achieve good performance. |
21394e70 DM |
383 | |
384 | ||
07fef357 | 385 | [[pve_ceph_pools]] |
1d54c3b4 AA |
386 | Creating Ceph Pools |
387 | ------------------- | |
21394e70 | 388 | |
1ff5e4e8 | 389 | [thumbnail="screenshot/gui-ceph-pools.png"] |
8997dd6e | 390 | |
1d54c3b4 | 391 | A pool is a logical group for storing objects. It holds **P**lacement |
90682f35 | 392 | **G**roups (`PG`, `pg_num`), a collection of objects. |
1d54c3b4 | 393 | |
90682f35 TL |
394 | When no options are given, we set a default of **128 PGs**, a **size of 3 |
395 | replicas** and a **min_size of 2 replicas** for serving objects in a degraded | |
396 | state. | |
1d54c3b4 | 397 | |
5a54ef44 | 398 | NOTE: The default number of PGs works for 2-5 disks. Ceph throws a |
90682f35 | 399 | 'HEALTH_WARNING' if you have too few or too many PGs in your cluster. |
1d54c3b4 AA |
400 | |
401 | It is advised to calculate the PG number depending on your setup, you can find | |
a474ca1f AA |
402 | the formula and the PG calculator footnote:[PG calculator |
403 | http://ceph.com/pgcalc/] online. While PGs can be increased later on, they can | |
404 | never be decreased. | |
1d54c3b4 AA |
405 | |
406 | ||
407 | You can create pools through command line or on the GUI on each PVE host under | |
408 | **Ceph -> Pools**. | |
409 | ||
410 | [source,bash] | |
411 | ---- | |
412 | pveceph createpool <name> | |
413 | ---- | |
414 | ||
415 | If you would like to automatically get also a storage definition for your pool, | |
416 | active the checkbox "Add storages" on the GUI or use the command line option | |
417 | '--add_storages' on pool creation. | |
21394e70 | 418 | |
1d54c3b4 AA |
419 | Further information on Ceph pool handling can be found in the Ceph pool |
420 | operation footnote:[Ceph pool operation | |
421 | http://docs.ceph.com/docs/luminous/rados/operations/pools/] | |
422 | manual. | |
21394e70 | 423 | |
76f6eca4 | 424 | [[pve_ceph_device_classes]] |
9fad507d AA |
425 | Ceph CRUSH & device classes |
426 | --------------------------- | |
427 | The foundation of Ceph is its algorithm, **C**ontrolled **R**eplication | |
428 | **U**nder **S**calable **H**ashing | |
429 | (CRUSH footnote:[CRUSH https://ceph.com/wp-content/uploads/2016/08/weil-crush-sc06.pdf]). | |
430 | ||
431 | CRUSH calculates where to store to and retrieve data from, this has the | |
432 | advantage that no central index service is needed. CRUSH works with a map of | |
433 | OSDs, buckets (device locations) and rulesets (data replication) for pools. | |
434 | ||
435 | NOTE: Further information can be found in the Ceph documentation, under the | |
436 | section CRUSH map footnote:[CRUSH map http://docs.ceph.com/docs/luminous/rados/operations/crush-map/]. | |
437 | ||
438 | This map can be altered to reflect different replication hierarchies. The object | |
439 | replicas can be separated (eg. failure domains), while maintaining the desired | |
440 | distribution. | |
441 | ||
442 | A common use case is to use different classes of disks for different Ceph pools. | |
443 | For this reason, Ceph introduced the device classes with luminous, to | |
444 | accommodate the need for easy ruleset generation. | |
445 | ||
446 | The device classes can be seen in the 'ceph osd tree' output. These classes | |
447 | represent their own root bucket, which can be seen with the below command. | |
448 | ||
449 | [source, bash] | |
450 | ---- | |
451 | ceph osd crush tree --show-shadow | |
452 | ---- | |
453 | ||
454 | Example output form the above command: | |
455 | ||
456 | [source, bash] | |
457 | ---- | |
458 | ID CLASS WEIGHT TYPE NAME | |
459 | -16 nvme 2.18307 root default~nvme | |
460 | -13 nvme 0.72769 host sumi1~nvme | |
461 | 12 nvme 0.72769 osd.12 | |
462 | -14 nvme 0.72769 host sumi2~nvme | |
463 | 13 nvme 0.72769 osd.13 | |
464 | -15 nvme 0.72769 host sumi3~nvme | |
465 | 14 nvme 0.72769 osd.14 | |
466 | -1 7.70544 root default | |
467 | -3 2.56848 host sumi1 | |
468 | 12 nvme 0.72769 osd.12 | |
469 | -5 2.56848 host sumi2 | |
470 | 13 nvme 0.72769 osd.13 | |
471 | -7 2.56848 host sumi3 | |
472 | 14 nvme 0.72769 osd.14 | |
473 | ---- | |
474 | ||
475 | To let a pool distribute its objects only on a specific device class, you need | |
476 | to create a ruleset with the specific class first. | |
477 | ||
478 | [source, bash] | |
479 | ---- | |
480 | ceph osd crush rule create-replicated <rule-name> <root> <failure-domain> <class> | |
481 | ---- | |
482 | ||
483 | [frame="none",grid="none", align="left", cols="30%,70%"] | |
484 | |=== | |
485 | |<rule-name>|name of the rule, to connect with a pool (seen in GUI & CLI) | |
486 | |<root>|which crush root it should belong to (default ceph root "default") | |
487 | |<failure-domain>|at which failure-domain the objects should be distributed (usually host) | |
488 | |<class>|what type of OSD backing store to use (eg. nvme, ssd, hdd) | |
489 | |=== | |
490 | ||
491 | Once the rule is in the CRUSH map, you can tell a pool to use the ruleset. | |
492 | ||
493 | [source, bash] | |
494 | ---- | |
495 | ceph osd pool set <pool-name> crush_rule <rule-name> | |
496 | ---- | |
497 | ||
498 | TIP: If the pool already contains objects, all of these have to be moved | |
499 | accordingly. Depending on your setup this may introduce a big performance hit on | |
500 | your cluster. As an alternative, you can create a new pool and move disks | |
501 | separately. | |
502 | ||
503 | ||
21394e70 DM |
504 | Ceph Client |
505 | ----------- | |
506 | ||
1ff5e4e8 | 507 | [thumbnail="screenshot/gui-ceph-log.png"] |
8997dd6e | 508 | |
21394e70 DM |
509 | You can then configure {pve} to use such pools to store VM or |
510 | Container images. Simply use the GUI too add a new `RBD` storage (see | |
511 | section xref:ceph_rados_block_devices[Ceph RADOS Block Devices (RBD)]). | |
512 | ||
1d54c3b4 AA |
513 | You also need to copy the keyring to a predefined location for a external Ceph |
514 | cluster. If Ceph is installed on the Proxmox nodes itself, then this will be | |
515 | done automatically. | |
21394e70 DM |
516 | |
517 | NOTE: The file name needs to be `<storage_id> + `.keyring` - `<storage_id>` is | |
518 | the expression after 'rbd:' in `/etc/pve/storage.cfg` which is | |
519 | `my-ceph-storage` in the following example: | |
520 | ||
521 | [source,bash] | |
522 | ---- | |
523 | mkdir /etc/pve/priv/ceph | |
524 | cp /etc/ceph/ceph.client.admin.keyring /etc/pve/priv/ceph/my-ceph-storage.keyring | |
525 | ---- | |
0840a663 | 526 | |
58f95dd7 TL |
527 | [[pveceph_fs]] |
528 | CephFS | |
529 | ------ | |
530 | ||
531 | Ceph provides also a filesystem running on top of the same object storage as | |
532 | RADOS block devices do. A **M**eta**d**ata **S**erver (`MDS`) is used to map | |
533 | the RADOS backed objects to files and directories, allowing to provide a | |
534 | POSIX-compliant replicated filesystem. This allows one to have a clustered | |
535 | highly available shared filesystem in an easy way if ceph is already used. Its | |
536 | Metadata Servers guarantee that files get balanced out over the whole Ceph | |
537 | cluster, this way even high load will not overload a single host, which can be | |
d180eb39 | 538 | an issue with traditional shared filesystem approaches, like `NFS`, for |
58f95dd7 TL |
539 | example. |
540 | ||
2394c306 | 541 | {pve} supports both, using an existing xref:storage_cephfs[CephFS as storage] |
58f95dd7 TL |
542 | to save backups, ISO files or container templates and creating a |
543 | hyper-converged CephFS itself. | |
544 | ||
545 | ||
546 | [[pveceph_fs_mds]] | |
547 | Metadata Server (MDS) | |
548 | ~~~~~~~~~~~~~~~~~~~~~ | |
549 | ||
550 | CephFS needs at least one Metadata Server to be configured and running to be | |
551 | able to work. One can simply create one through the {pve} web GUI's `Node -> | |
552 | CephFS` panel or on the command line with: | |
553 | ||
554 | ---- | |
555 | pveceph mds create | |
556 | ---- | |
557 | ||
558 | Multiple metadata servers can be created in a cluster. But with the default | |
559 | settings only one can be active at any time. If an MDS, or its node, becomes | |
560 | unresponsive (or crashes), another `standby` MDS will get promoted to `active`. | |
561 | One can speed up the hand-over between the active and a standby MDS up by using | |
562 | the 'hotstandby' parameter option on create, or if you have already created it | |
563 | you may set/add: | |
564 | ||
565 | ---- | |
566 | mds standby replay = true | |
567 | ---- | |
568 | ||
569 | in the ceph.conf respective MDS section. With this enabled, this specific MDS | |
570 | will always poll the active one, so that it can take over faster as it is in a | |
3580eb13 | 571 | `warm` state. But naturally, the active polling will cause some additional |
58f95dd7 TL |
572 | performance impact on your system and active `MDS`. |
573 | ||
574 | Multiple Active MDS | |
575 | ^^^^^^^^^^^^^^^^^^^ | |
576 | ||
577 | Since Luminous (12.2.x) you can also have multiple active metadata servers | |
578 | running, but this is normally only useful for a high count on parallel clients, | |
579 | as else the `MDS` seldom is the bottleneck. If you want to set this up please | |
580 | refer to the ceph documentation. footnote:[Configuring multiple active MDS | |
127ca409 | 581 | daemons http://docs.ceph.com/docs/luminous/cephfs/multimds/] |
58f95dd7 TL |
582 | |
583 | [[pveceph_fs_create]] | |
584 | Create a CephFS | |
585 | ~~~~~~~~~~~~~~~ | |
586 | ||
587 | With {pve}'s CephFS integration into you can create a CephFS easily over the | |
588 | Web GUI, the CLI or an external API interface. Some prerequisites are required | |
589 | for this to work: | |
590 | ||
591 | .Prerequisites for a successful CephFS setup: | |
592 | - xref:pve_ceph_install[Install Ceph packages], if this was already done some | |
593 | time ago you might want to rerun it on an up to date system to ensure that | |
594 | also all CephFS related packages get installed. | |
595 | - xref:pve_ceph_monitors[Setup Monitors] | |
596 | - xref:pve_ceph_monitors[Setup your OSDs] | |
597 | - xref:pveceph_fs_mds[Setup at least one MDS] | |
598 | ||
599 | After this got all checked and done you can simply create a CephFS through | |
600 | either the Web GUI's `Node -> CephFS` panel or the command line tool `pveceph`, | |
601 | for example with: | |
602 | ||
603 | ---- | |
604 | pveceph fs create --pg_num 128 --add-storage | |
605 | ---- | |
606 | ||
607 | This creates a CephFS named `'cephfs'' using a pool for its data named | |
608 | `'cephfs_data'' with `128` placement groups and a pool for its metadata named | |
609 | `'cephfs_metadata'' with one quarter of the data pools placement groups (`32`). | |
610 | Check the xref:pve_ceph_pools[{pve} managed Ceph pool chapter] or visit the | |
611 | Ceph documentation for more information regarding a fitting placement group | |
612 | number (`pg_num`) for your setup footnote:[Ceph Placement Groups | |
127ca409 | 613 | http://docs.ceph.com/docs/luminous/rados/operations/placement-groups/]. |
58f95dd7 TL |
614 | Additionally, the `'--add-storage'' parameter will add the CephFS to the {pve} |
615 | storage configuration after it was created successfully. | |
616 | ||
617 | Destroy CephFS | |
618 | ~~~~~~~~~~~~~~ | |
619 | ||
fa9b4ee1 | 620 | WARNING: Destroying a CephFS will render all its data unusable, this cannot be |
58f95dd7 TL |
621 | undone! |
622 | ||
623 | If you really want to destroy an existing CephFS you first need to stop, or | |
624 | destroy, all metadata server (`M̀DS`). You can destroy them either over the Web | |
625 | GUI or the command line interface, with: | |
626 | ||
627 | ---- | |
628 | pveceph mds destroy NAME | |
629 | ---- | |
630 | on each {pve} node hosting a MDS daemon. | |
631 | ||
632 | Then, you can remove (destroy) CephFS by issuing a: | |
633 | ||
634 | ---- | |
de2f8225 | 635 | ceph fs rm NAME --yes-i-really-mean-it |
58f95dd7 TL |
636 | ---- |
637 | on a single node hosting Ceph. After this you may want to remove the created | |
638 | data and metadata pools, this can be done either over the Web GUI or the CLI | |
639 | with: | |
640 | ||
641 | ---- | |
642 | pveceph pool destroy NAME | |
643 | ---- | |
0840a663 | 644 | |
6ff32926 | 645 | |
10df14fb TL |
646 | Ceph monitoring and troubleshooting |
647 | ----------------------------------- | |
648 | A good start is to continuosly monitor the ceph health from the start of | |
649 | initial deployment. Either through the ceph tools itself, but also by accessing | |
650 | the status through the {pve} link:api-viewer/index.html[API]. | |
6ff32926 | 651 | |
10df14fb TL |
652 | The following ceph commands below can be used to see if the cluster is healthy |
653 | ('HEALTH_OK'), if there are warnings ('HEALTH_WARN'), or even errors | |
654 | ('HEALTH_ERR'). If the cluster is in an unhealthy state the status commands | |
655 | below will also give you an overview on the current events and actions take. | |
6ff32926 AA |
656 | |
657 | ---- | |
10df14fb TL |
658 | # single time output |
659 | pve# ceph -s | |
660 | # continuously output status changes (press CTRL+C to stop) | |
661 | pve# ceph -w | |
6ff32926 AA |
662 | ---- |
663 | ||
664 | To get a more detailed view, every ceph service has a log file under | |
665 | `/var/log/ceph/` and if there is not enough detail, the log level can be | |
666 | adjusted footnote:[Ceph log and debugging http://docs.ceph.com/docs/luminous/rados/troubleshooting/log-and-debug/]. | |
667 | ||
668 | You can find more information about troubleshooting | |
669 | footnote:[Ceph troubleshooting http://docs.ceph.com/docs/luminous/rados/troubleshooting/] | |
670 | a Ceph cluster on its website. | |
671 | ||
672 | ||
0840a663 DM |
673 | ifdef::manvolnum[] |
674 | include::pve-copyright.adoc[] | |
675 | endif::manvolnum[] |