]>
Commit | Line | Data |
---|---|---|
1 | ifdef::manvolnum[] | |
2 | PVE({manvolnum}) | |
3 | ================ | |
4 | include::attributes.txt[] | |
5 | ||
6 | NAME | |
7 | ---- | |
8 | ||
9 | pvecm - Proxmox VE Cluster Manager | |
10 | ||
11 | SYNOPSIS | |
12 | -------- | |
13 | ||
14 | include::pvecm.1-synopsis.adoc[] | |
15 | ||
16 | DESCRIPTION | |
17 | ----------- | |
18 | endif::manvolnum[] | |
19 | ||
20 | ifndef::manvolnum[] | |
21 | Cluster Manager | |
22 | =============== | |
23 | include::attributes.txt[] | |
24 | endif::manvolnum[] | |
25 | ||
26 | The {PVE} cluster manager `pvecm` is a tool to create a group of | |
27 | physical servers. Such a group is called a *cluster*. We use the | |
28 | http://www.corosync.org[Corosync Cluster Engine] for reliable group | |
29 | communication, and such clusters can consist of up to 32 physical nodes | |
30 | (probably more, dependent on network latency). | |
31 | ||
32 | `pvecm` can be used to create a new cluster, join nodes to a cluster, | |
33 | leave the cluster, get status information and do various other cluster | |
34 | related tasks. The **P**rox**m**o**x** **C**luster **F**ile **S**ystem (``pmxcfs'') | |
35 | is used to transparently distribute the cluster configuration to all cluster | |
36 | nodes. | |
37 | ||
38 | Grouping nodes into a cluster has the following advantages: | |
39 | ||
40 | * Centralized, web based management | |
41 | ||
42 | * Multi-master clusters: each node can do all management task | |
43 | ||
44 | * `pmxcfs`: database-driven file system for storing configuration files, | |
45 | replicated in real-time on all nodes using `corosync`. | |
46 | ||
47 | * Easy migration of virtual machines and containers between physical | |
48 | hosts | |
49 | ||
50 | * Fast deployment | |
51 | ||
52 | * Cluster-wide services like firewall and HA | |
53 | ||
54 | ||
55 | Requirements | |
56 | ------------ | |
57 | ||
58 | * All nodes must be in the same network as `corosync` uses IP Multicast | |
59 | to communicate between nodes (also see | |
60 | http://www.corosync.org[Corosync Cluster Engine]). Corosync uses UDP | |
61 | ports 5404 and 5405 for cluster communication. | |
62 | + | |
63 | NOTE: Some switches do not support IP multicast by default and must be | |
64 | manually enabled first. | |
65 | ||
66 | * Date and time have to be synchronized. | |
67 | ||
68 | * SSH tunnel on TCP port 22 between nodes is used. | |
69 | ||
70 | * If you are interested in High Availability, you need to have at | |
71 | least three nodes for reliable quorum. All nodes should have the | |
72 | same version. | |
73 | ||
74 | * We recommend a dedicated NIC for the cluster traffic, especially if | |
75 | you use shared storage. | |
76 | ||
77 | NOTE: It is not possible to mix Proxmox VE 3.x and earlier with | |
78 | Proxmox VE 4.0 cluster nodes. | |
79 | ||
80 | ||
81 | Preparing Nodes | |
82 | --------------- | |
83 | ||
84 | First, install {PVE} on all nodes. Make sure that each node is | |
85 | installed with the final hostname and IP configuration. Changing the | |
86 | hostname and IP is not possible after cluster creation. | |
87 | ||
88 | Currently the cluster creation has to be done on the console, so you | |
89 | need to login via `ssh`. | |
90 | ||
91 | Create the Cluster | |
92 | ------------------ | |
93 | ||
94 | Login via `ssh` to the first {pve} node. Use a unique name for your cluster. | |
95 | This name cannot be changed later. | |
96 | ||
97 | hp1# pvecm create YOUR-CLUSTER-NAME | |
98 | ||
99 | CAUTION: The cluster name is used to compute the default multicast | |
100 | address. Please use unique cluster names if you run more than one | |
101 | cluster inside your network. | |
102 | ||
103 | To check the state of your cluster use: | |
104 | ||
105 | hp1# pvecm status | |
106 | ||
107 | ||
108 | Adding Nodes to the Cluster | |
109 | --------------------------- | |
110 | ||
111 | Login via `ssh` to the node you want to add. | |
112 | ||
113 | hp2# pvecm add IP-ADDRESS-CLUSTER | |
114 | ||
115 | For `IP-ADDRESS-CLUSTER` use the IP from an existing cluster node. | |
116 | ||
117 | CAUTION: A new node cannot hold any VMs, because you would get | |
118 | conflicts about identical VM IDs. Also, all existing configuration in | |
119 | `/etc/pve` is overwritten when you join a new node to the cluster. To | |
120 | workaround, use `vzdump` to backup and restore to a different VMID after | |
121 | adding the node to the cluster. | |
122 | ||
123 | To check the state of cluster: | |
124 | ||
125 | # pvecm status | |
126 | ||
127 | .Cluster status after adding 4 nodes | |
128 | ---- | |
129 | hp2# pvecm status | |
130 | Quorum information | |
131 | ~~~~~~~~~~~~~~~~~~ | |
132 | Date: Mon Apr 20 12:30:13 2015 | |
133 | Quorum provider: corosync_votequorum | |
134 | Nodes: 4 | |
135 | Node ID: 0x00000001 | |
136 | Ring ID: 1928 | |
137 | Quorate: Yes | |
138 | ||
139 | Votequorum information | |
140 | ~~~~~~~~~~~~~~~~~~~~~~ | |
141 | Expected votes: 4 | |
142 | Highest expected: 4 | |
143 | Total votes: 4 | |
144 | Quorum: 2 | |
145 | Flags: Quorate | |
146 | ||
147 | Membership information | |
148 | ~~~~~~~~~~~~~~~~~~~~~~ | |
149 | Nodeid Votes Name | |
150 | 0x00000001 1 192.168.15.91 | |
151 | 0x00000002 1 192.168.15.92 (local) | |
152 | 0x00000003 1 192.168.15.93 | |
153 | 0x00000004 1 192.168.15.94 | |
154 | ---- | |
155 | ||
156 | If you only want the list of all nodes use: | |
157 | ||
158 | # pvecm nodes | |
159 | ||
160 | .List nodes in a cluster | |
161 | ---- | |
162 | hp2# pvecm nodes | |
163 | ||
164 | Membership information | |
165 | ~~~~~~~~~~~~~~~~~~~~~~ | |
166 | Nodeid Votes Name | |
167 | 1 1 hp1 | |
168 | 2 1 hp2 (local) | |
169 | 3 1 hp3 | |
170 | 4 1 hp4 | |
171 | ---- | |
172 | ||
173 | Adding Nodes With Separated Cluster Network | |
174 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ | |
175 | ||
176 | When adding a node to a cluster with a separated cluster network you need to | |
177 | use the 'ringX_addr' parameters to set the nodes address on those networks: | |
178 | ||
179 | [source,bash] | |
180 | ---- | |
181 | pvecm add IP-ADDRESS-CLUSTER -ring0_addr IP-ADDRESS-RING0 | |
182 | ---- | |
183 | ||
184 | If you want to use the Redundant Ring Protocol you will also want to pass the | |
185 | 'ring1_addr' parameter. | |
186 | ||
187 | ||
188 | Remove a Cluster Node | |
189 | --------------------- | |
190 | ||
191 | CAUTION: Read carefully the procedure before proceeding, as it could | |
192 | not be what you want or need. | |
193 | ||
194 | Move all virtual machines from the node. Make sure you have no local | |
195 | data or backups you want to keep, or save them accordingly. | |
196 | ||
197 | Log in to one remaining node via ssh. Issue a `pvecm nodes` command to | |
198 | identify the node ID: | |
199 | ||
200 | ---- | |
201 | hp1# pvecm status | |
202 | ||
203 | Quorum information | |
204 | ~~~~~~~~~~~~~~~~~~ | |
205 | Date: Mon Apr 20 12:30:13 2015 | |
206 | Quorum provider: corosync_votequorum | |
207 | Nodes: 4 | |
208 | Node ID: 0x00000001 | |
209 | Ring ID: 1928 | |
210 | Quorate: Yes | |
211 | ||
212 | Votequorum information | |
213 | ~~~~~~~~~~~~~~~~~~~~~~ | |
214 | Expected votes: 4 | |
215 | Highest expected: 4 | |
216 | Total votes: 4 | |
217 | Quorum: 2 | |
218 | Flags: Quorate | |
219 | ||
220 | Membership information | |
221 | ~~~~~~~~~~~~~~~~~~~~~~ | |
222 | Nodeid Votes Name | |
223 | 0x00000001 1 192.168.15.91 (local) | |
224 | 0x00000002 1 192.168.15.92 | |
225 | 0x00000003 1 192.168.15.93 | |
226 | 0x00000004 1 192.168.15.94 | |
227 | ---- | |
228 | ||
229 | IMPORTANT: at this point you must power off the node to be removed and | |
230 | make sure that it will not power on again (in the network) as it | |
231 | is. | |
232 | ||
233 | ---- | |
234 | hp1# pvecm nodes | |
235 | ||
236 | Membership information | |
237 | ~~~~~~~~~~~~~~~~~~~~~~ | |
238 | Nodeid Votes Name | |
239 | 1 1 hp1 (local) | |
240 | 2 1 hp2 | |
241 | 3 1 hp3 | |
242 | 4 1 hp4 | |
243 | ---- | |
244 | ||
245 | Log in to one remaining node via ssh. Issue the delete command (here | |
246 | deleting node `hp4`): | |
247 | ||
248 | hp1# pvecm delnode hp4 | |
249 | ||
250 | If the operation succeeds no output is returned, just check the node | |
251 | list again with `pvecm nodes` or `pvecm status`. You should see | |
252 | something like: | |
253 | ||
254 | ---- | |
255 | hp1# pvecm status | |
256 | ||
257 | Quorum information | |
258 | ~~~~~~~~~~~~~~~~~~ | |
259 | Date: Mon Apr 20 12:44:28 2015 | |
260 | Quorum provider: corosync_votequorum | |
261 | Nodes: 3 | |
262 | Node ID: 0x00000001 | |
263 | Ring ID: 1992 | |
264 | Quorate: Yes | |
265 | ||
266 | Votequorum information | |
267 | ~~~~~~~~~~~~~~~~~~~~~~ | |
268 | Expected votes: 3 | |
269 | Highest expected: 3 | |
270 | Total votes: 3 | |
271 | Quorum: 3 | |
272 | Flags: Quorate | |
273 | ||
274 | Membership information | |
275 | ~~~~~~~~~~~~~~~~~~~~~~ | |
276 | Nodeid Votes Name | |
277 | 0x00000001 1 192.168.15.90 (local) | |
278 | 0x00000002 1 192.168.15.91 | |
279 | 0x00000003 1 192.168.15.92 | |
280 | ---- | |
281 | ||
282 | IMPORTANT: as said above, it is very important to power off the node | |
283 | *before* removal, and make sure that it will *never* power on again | |
284 | (in the existing cluster network) as it is. | |
285 | ||
286 | If you power on the node as it is, your cluster will be screwed up and | |
287 | it could be difficult to restore a clean cluster state. | |
288 | ||
289 | If, for whatever reason, you want that this server joins the same | |
290 | cluster again, you have to | |
291 | ||
292 | * reinstall {pve} on it from scratch | |
293 | ||
294 | * then join it, as explained in the previous section. | |
295 | ||
296 | Separate A Node Without Reinstalling | |
297 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ | |
298 | ||
299 | CAUTION: This is *not* the recommended method, proceed with caution. Use the | |
300 | above mentioned method if you're unsure. | |
301 | ||
302 | You can also separate a node from a cluster without reinstalling it from | |
303 | scratch. But after removing the node from the cluster it will still have | |
304 | access to the shared storages! This must be resolved before you start removing | |
305 | the node from the cluster. A {pve} cluster cannot share the exact same | |
306 | storage with another cluster, as it leads to VMID conflicts. | |
307 | ||
308 | Its suggested that you create a new storage where only the node which you want | |
309 | to separate has access. This can be an new export on your NFS or a new Ceph | |
310 | pool, to name a few examples. Its just important that the exact same storage | |
311 | does not gets accessed by multiple clusters. After setting this storage up move | |
312 | all data from the node and its VMs to it. Then you are ready to separate the | |
313 | node from the cluster. | |
314 | ||
315 | WARNING: Ensure all shared resources are cleanly separated! You will run into | |
316 | conflicts and problems else. | |
317 | ||
318 | First stop the corosync and the pve-cluster services on the node: | |
319 | [source,bash] | |
320 | ---- | |
321 | systemctl stop pve-cluster | |
322 | systemctl stop corosync | |
323 | ---- | |
324 | ||
325 | Start the cluster filesystem again in local mode: | |
326 | [source,bash] | |
327 | ---- | |
328 | pmxcfs -l | |
329 | ---- | |
330 | ||
331 | Delete the corosync configuration files: | |
332 | [source,bash] | |
333 | ---- | |
334 | rm /etc/pve/corosync.conf | |
335 | rm /etc/corosync/* | |
336 | ---- | |
337 | ||
338 | You can now start the filesystem again as normal service: | |
339 | [source,bash] | |
340 | ---- | |
341 | killall pmxcfs | |
342 | systemctl start pve-cluster | |
343 | ---- | |
344 | ||
345 | The node is now separated from the cluster. You can deleted it from a remaining | |
346 | node of the cluster with: | |
347 | [source,bash] | |
348 | ---- | |
349 | pvecm delnode oldnode | |
350 | ---- | |
351 | ||
352 | If the command failed, because the remaining node in the cluster lost quorum | |
353 | when the now separate node exited, you may set the expected votes to 1 as a workaround: | |
354 | [source,bash] | |
355 | ---- | |
356 | pvecm expected 1 | |
357 | ---- | |
358 | ||
359 | And the repeat the 'pvecm delnode' command. | |
360 | ||
361 | Now switch back to the separated node, here delete all remaining files left | |
362 | from the old cluster. This ensures that the node can be added to another | |
363 | cluster again without problems. | |
364 | ||
365 | [source,bash] | |
366 | ---- | |
367 | rm /var/lib/corosync/* | |
368 | ---- | |
369 | ||
370 | As the configuration files from the other nodes are still in the cluster | |
371 | filesystem you may want to clean those up too. Remove simply the whole | |
372 | directory recursive from '/etc/pve/nodes/NODENAME', but check three times that | |
373 | you used the correct one before deleting it. | |
374 | ||
375 | CAUTION: The nodes SSH keys are still in the 'authorized_key' file, this means | |
376 | the nodes can still connect to each other with public key authentication. This | |
377 | should be fixed by removing the respective keys from the | |
378 | '/etc/pve/priv/authorized_keys' file. | |
379 | ||
380 | Quorum | |
381 | ------ | |
382 | ||
383 | {pve} use a quorum-based technique to provide a consistent state among | |
384 | all cluster nodes. | |
385 | ||
386 | [quote, from Wikipedia, Quorum (distributed computing)] | |
387 | ____ | |
388 | A quorum is the minimum number of votes that a distributed transaction | |
389 | has to obtain in order to be allowed to perform an operation in a | |
390 | distributed system. | |
391 | ____ | |
392 | ||
393 | In case of network partitioning, state changes requires that a | |
394 | majority of nodes are online. The cluster switches to read-only mode | |
395 | if it loses quorum. | |
396 | ||
397 | NOTE: {pve} assigns a single vote to each node by default. | |
398 | ||
399 | Cluster Network | |
400 | --------------- | |
401 | ||
402 | The cluster network is the core of a cluster. All messages sent over it have to | |
403 | be delivered reliable to all nodes in their respective order. In {pve} this | |
404 | part is done by corosync, an implementation of a high performance low overhead | |
405 | high availability development toolkit. It serves our decentralized | |
406 | configuration file system (`pmxcfs`). | |
407 | ||
408 | [[cluster-network-requirements]] | |
409 | Network Requirements | |
410 | ~~~~~~~~~~~~~~~~~~~~ | |
411 | This needs a reliable network with latencies under 2 milliseconds (LAN | |
412 | performance) to work properly. While corosync can also use unicast for | |
413 | communication between nodes its **highly recommended** to have a multicast | |
414 | capable network. The network should not be used heavily by other members, | |
415 | ideally corosync runs on its own network. | |
416 | *never* share it with network where storage communicates too. | |
417 | ||
418 | Before setting up a cluster it is good practice to check if the network is fit | |
419 | for that purpose. | |
420 | ||
421 | * Ensure that all nodes are in the same subnet. This must only be true for the | |
422 | network interfaces used for cluster communication (corosync). | |
423 | ||
424 | * Ensure all nodes can reach each other over those interfaces, using `ping` is | |
425 | enough for a basic test. | |
426 | ||
427 | * Ensure that multicast works in general and a high package rates. This can be | |
428 | done with the `omping` tool. The final "%loss" number should be < 1%. | |
429 | [source,bash] | |
430 | ---- | |
431 | omping -c 10000 -i 0.001 -F -q NODE1-IP NODE2-IP ... | |
432 | ---- | |
433 | ||
434 | * Ensure that multicast communication works over an extended period of time. | |
435 | This covers up problems where IGMP snooping is activated on the network but | |
436 | no multicast querier is active. This test has a duration of around 10 | |
437 | minutes. | |
438 | [source,bash] | |
439 | ---- | |
440 | omping -c 600 -i 1 -q NODE1-IP NODE2-IP ... | |
441 | ---- | |
442 | ||
443 | Your network is not ready for clustering if any of these test fails. Recheck | |
444 | your network configuration. Especially switches are notorious for having | |
445 | multicast disabled by default or IGMP snooping enabled with no IGMP querier | |
446 | active. | |
447 | ||
448 | In smaller cluster its also an option to use unicast if you really cannot get | |
449 | multicast to work. | |
450 | ||
451 | Separate Cluster Network | |
452 | ~~~~~~~~~~~~~~~~~~~~~~~~ | |
453 | ||
454 | When creating a cluster without any parameters the cluster network is generally | |
455 | shared with the Web UI and the VMs and its traffic. Depending on your setup | |
456 | even storage traffic may get sent over the same network. Its recommended to | |
457 | change that, as corosync is a time critical real time application. | |
458 | ||
459 | Setting Up A New Network | |
460 | ^^^^^^^^^^^^^^^^^^^^^^^^ | |
461 | ||
462 | First you have to setup a new network interface. It should be on a physical | |
463 | separate network. Ensure that your network fulfills the | |
464 | <<cluster-network-requirements,cluster network requirements>>. | |
465 | ||
466 | Separate On Cluster Creation | |
467 | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^ | |
468 | ||
469 | This is possible through the 'ring0_addr' and 'bindnet0_addr' parameter of | |
470 | the 'pvecm create' command used for creating a new cluster. | |
471 | ||
472 | If you have setup a additional NIC with a static address on 10.10.10.1/25 | |
473 | and want to send and receive all cluster communication over this interface | |
474 | you would execute: | |
475 | ||
476 | [source,bash] | |
477 | ---- | |
478 | pvecm create test --ring0_addr 10.10.10.1 --bindnet0_addr 10.10.10.0 | |
479 | ---- | |
480 | ||
481 | To check if everything is working properly execute: | |
482 | [source,bash] | |
483 | ---- | |
484 | systemctl status corosync | |
485 | ---- | |
486 | ||
487 | [[separate-cluster-net-after-creation]] | |
488 | Separate After Cluster Creation | |
489 | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ | |
490 | ||
491 | You can do this also if you have already created a cluster and want to switch | |
492 | its communication to another network, without rebuilding the whole cluster. | |
493 | This change may lead to short durations of quorum loss in the cluster, as nodes | |
494 | have to restart corosync and come up one after the other on the new network. | |
495 | ||
496 | Check how to <<edit-corosync-conf,edit the corosync.conf file>> first. | |
497 | The open it and you should see a file similar to: | |
498 | ||
499 | ---- | |
500 | logging { | |
501 | debug: off | |
502 | to_syslog: yes | |
503 | } | |
504 | ||
505 | nodelist { | |
506 | ||
507 | node { | |
508 | name: due | |
509 | nodeid: 2 | |
510 | quorum_votes: 1 | |
511 | ring0_addr: due | |
512 | } | |
513 | ||
514 | node { | |
515 | name: tre | |
516 | nodeid: 3 | |
517 | quorum_votes: 1 | |
518 | ring0_addr: tre | |
519 | } | |
520 | ||
521 | node { | |
522 | name: uno | |
523 | nodeid: 1 | |
524 | quorum_votes: 1 | |
525 | ring0_addr: uno | |
526 | } | |
527 | ||
528 | } | |
529 | ||
530 | quorum { | |
531 | provider: corosync_votequorum | |
532 | } | |
533 | ||
534 | totem { | |
535 | cluster_name: thomas-testcluster | |
536 | config_version: 3 | |
537 | ip_version: ipv4 | |
538 | secauth: on | |
539 | version: 2 | |
540 | interface { | |
541 | bindnetaddr: 192.168.30.50 | |
542 | ringnumber: 0 | |
543 | } | |
544 | ||
545 | } | |
546 | ---- | |
547 | ||
548 | The first you want to do is add the 'name' properties in the node entries if | |
549 | you do not see them already. Those *must* match the node name. | |
550 | ||
551 | Then replace the address from the 'ring0_addr' properties with the new | |
552 | addresses. You may use plain IP addresses or also hostnames here. If you use | |
553 | hostnames ensure that they are resolvable from all nodes. | |
554 | ||
555 | In my example I want to switch my cluster communication to the 10.10.10.1/25 | |
556 | network. So I replace all 'ring0_addr' respectively. I also set the bindetaddr | |
557 | in the totem section of the config to an address of the new network. It can be | |
558 | any address from the subnet configured on the new network interface. | |
559 | ||
560 | After you increased the 'config_version' property the new configuration file | |
561 | should look like: | |
562 | ||
563 | ---- | |
564 | ||
565 | logging { | |
566 | debug: off | |
567 | to_syslog: yes | |
568 | } | |
569 | ||
570 | nodelist { | |
571 | ||
572 | node { | |
573 | name: due | |
574 | nodeid: 2 | |
575 | quorum_votes: 1 | |
576 | ring0_addr: 10.10.10.2 | |
577 | } | |
578 | ||
579 | node { | |
580 | name: tre | |
581 | nodeid: 3 | |
582 | quorum_votes: 1 | |
583 | ring0_addr: 10.10.10.3 | |
584 | } | |
585 | ||
586 | node { | |
587 | name: uno | |
588 | nodeid: 1 | |
589 | quorum_votes: 1 | |
590 | ring0_addr: 10.10.10.1 | |
591 | } | |
592 | ||
593 | } | |
594 | ||
595 | quorum { | |
596 | provider: corosync_votequorum | |
597 | } | |
598 | ||
599 | totem { | |
600 | cluster_name: thomas-testcluster | |
601 | config_version: 4 | |
602 | ip_version: ipv4 | |
603 | secauth: on | |
604 | version: 2 | |
605 | interface { | |
606 | bindnetaddr: 10.10.10.1 | |
607 | ringnumber: 0 | |
608 | } | |
609 | ||
610 | } | |
611 | ---- | |
612 | ||
613 | Now after a final check whether all changed information is correct we save it | |
614 | and see again the <<edit-corosync-conf,edit corosync.conf file>> section to | |
615 | learn how to bring it in effect. | |
616 | ||
617 | As our change cannot be enforced live from corosync we have to do an restart. | |
618 | ||
619 | On a single node execute: | |
620 | [source,bash] | |
621 | ---- | |
622 | systemctl restart corosync | |
623 | ---- | |
624 | ||
625 | Now check if everything is fine: | |
626 | ||
627 | [source,bash] | |
628 | ---- | |
629 | systemctl status corosync | |
630 | ---- | |
631 | ||
632 | If corosync runs again correct restart corosync also on all other nodes. | |
633 | They will then join the cluster membership one by one on the new network. | |
634 | ||
635 | Redundant Ring Protocol | |
636 | ~~~~~~~~~~~~~~~~~~~~~~~ | |
637 | To avoid a single point of failure you should implement counter measurements. | |
638 | This can be on the hardware and operating system level through network bonding. | |
639 | ||
640 | Corosync itself offers also a possibility to add redundancy through the so | |
641 | called 'Redundant Ring Protocol'. This protocol allows running a second totem | |
642 | ring on another network, this network should be physically separated from the | |
643 | other rings network to actually increase availability. | |
644 | ||
645 | RRP On Cluster Creation | |
646 | ~~~~~~~~~~~~~~~~~~~~~~~ | |
647 | ||
648 | The 'pvecm create' command provides the additional parameters 'bindnetX_addr', | |
649 | 'ringX_addr' and 'rrp_mode', can be used for RRP configuration. | |
650 | ||
651 | NOTE: See the <<corosync-conf-glossary,glossary>> if you do not know what each parameter means. | |
652 | ||
653 | So if you have two networks, one on the 10.10.10.1/24 and the other on the | |
654 | 10.10.20.1/24 subnet you would execute: | |
655 | ||
656 | [source,bash] | |
657 | ---- | |
658 | pvecm create CLUSTERNAME -bindnet0_addr 10.10.10.1 -ring0_addr 10.10.10.1 \ | |
659 | -bindnet1_addr 10.10.20.1 -ring1_addr 10.10.20.1 | |
660 | ---- | |
661 | ||
662 | RRP On A Created Cluster | |
663 | ~~~~~~~~~~~~~~~~~~~~~~~~ | |
664 | ||
665 | When enabling an already running cluster to use RRP you will take similar steps | |
666 | as describe in <<separate-cluster-net-after-creation,separating the cluster | |
667 | network>>. You just do it on another ring. | |
668 | ||
669 | First add a new `interface` subsection in the `totem` section, set its | |
670 | `ringnumber` property to `1`. Set the interfaces `bindnetaddr` property to an | |
671 | address of the subnet you have configured for your new ring. | |
672 | Further set the `rrp_mode` to `passive`, this is the only stable mode. | |
673 | ||
674 | Then add to each node entry in the `nodelist` section its new `ring1_addr` | |
675 | property with the nodes additional ring address. | |
676 | ||
677 | So if you have two networks, one on the 10.10.10.1/24 and the other on the | |
678 | 10.10.20.1/24 subnet, the final configuration file should look like: | |
679 | ||
680 | ---- | |
681 | totem { | |
682 | cluster_name: tweak | |
683 | config_version: 9 | |
684 | ip_version: ipv4 | |
685 | rrp_mode: passive | |
686 | secauth: on | |
687 | version: 2 | |
688 | interface { | |
689 | bindnetaddr: 10.10.10.1 | |
690 | ringnumber: 0 | |
691 | } | |
692 | interface { | |
693 | bindnetaddr: 10.10.20.1 | |
694 | ringnumber: 1 | |
695 | } | |
696 | } | |
697 | ||
698 | nodelist { | |
699 | node { | |
700 | name: pvecm1 | |
701 | nodeid: 1 | |
702 | quorum_votes: 1 | |
703 | ring0_addr: 10.10.10.1 | |
704 | ring1_addr: 10.10.20.1 | |
705 | } | |
706 | ||
707 | node { | |
708 | name: pvecm2 | |
709 | nodeid: 2 | |
710 | quorum_votes: 1 | |
711 | ring0_addr: 10.10.10.2 | |
712 | ring1_addr: 10.10.20.2 | |
713 | } | |
714 | ||
715 | [...] # other cluster nodes here | |
716 | } | |
717 | ||
718 | [...] # other remaining config sections here | |
719 | ||
720 | ---- | |
721 | ||
722 | Bring it in effect like described in the <<edit-corosync-conf,edit the | |
723 | corosync.conf file>> section. | |
724 | ||
725 | This is a change which cannot take live in effect and needs at least a restart | |
726 | of corosync. Recommended is a restart of the whole cluster. | |
727 | ||
728 | If you cannot reboot the whole cluster ensure no High Availability services are | |
729 | configured and the stop the corosync service on all nodes. After corosync is | |
730 | stopped on all nodes start it one after the other again. | |
731 | ||
732 | Corosync Configuration | |
733 | ---------------------- | |
734 | ||
735 | The `/ect/pve/corosync.conf` file plays a central role in {pve} cluster. It | |
736 | controls the cluster member ship and its network. | |
737 | For reading more about it check the corosync.conf man page: | |
738 | [source,bash] | |
739 | ---- | |
740 | man corosync.conf | |
741 | ---- | |
742 | ||
743 | For node membership you should always use the `pvecm` tool provided by {pve}. | |
744 | You may have to edit the configuration file manually for other changes. | |
745 | Here are a few best practice tips for doing this. | |
746 | ||
747 | [[edit-corosync-conf]] | |
748 | Edit corosync.conf | |
749 | ~~~~~~~~~~~~~~~~~~ | |
750 | ||
751 | Editing the corosync.conf file can be not always straight forward. There are | |
752 | two on each cluster, one in `/etc/pve/corosync.conf` and the other in | |
753 | `/etc/corosync/corosync.conf`. Editing the one in our cluster file system will | |
754 | propagate the changes to the local one, but not vice versa. | |
755 | ||
756 | The configuration will get updated automatically as soon as the file changes. | |
757 | This means changes which can be integrated in a running corosync will take | |
758 | instantly effect. So you should always make a copy and edit that instead, to | |
759 | avoid triggering some unwanted changes by an in between safe. | |
760 | ||
761 | [source,bash] | |
762 | ---- | |
763 | cp /etc/pve/corosync.conf /etc/pve/corosync.conf.new | |
764 | ---- | |
765 | ||
766 | Then open the Config file with your favorite editor, `nano` and `vim.tiny` are | |
767 | preinstalled on {pve} for example. | |
768 | ||
769 | NOTE: Always increment the 'config_version' number on configuration changes, | |
770 | omitting this can lead to problems. | |
771 | ||
772 | After making the necessary changes create another copy of the current working | |
773 | configuration file. This serves as a backup if the new configuration fails to | |
774 | apply or makes problems in other ways. | |
775 | ||
776 | [source,bash] | |
777 | ---- | |
778 | cp /etc/pve/corosync.conf /etc/pve/corosync.conf.bak | |
779 | ---- | |
780 | ||
781 | Then move the new configuration file over the old one: | |
782 | [source,bash] | |
783 | ---- | |
784 | mv /etc/pve/corosync.conf.new /etc/pve/corosync.conf | |
785 | ---- | |
786 | ||
787 | You may check with the commands | |
788 | [source,bash] | |
789 | ---- | |
790 | systemctl status corosync | |
791 | journalctl -b -u corosync | |
792 | ---- | |
793 | ||
794 | If the change could applied automatically. If not you may have to restart the | |
795 | corosync service via: | |
796 | [source,bash] | |
797 | ---- | |
798 | systemctl restart corosync | |
799 | ---- | |
800 | ||
801 | On errors check the troubleshooting section below. | |
802 | ||
803 | Troubleshooting | |
804 | ~~~~~~~~~~~~~~~ | |
805 | ||
806 | Issue: 'quorum.expected_votes must be configured' | |
807 | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ | |
808 | ||
809 | When corosync starts to fail and you get the following message in the system log: | |
810 | ||
811 | ---- | |
812 | [...] | |
813 | corosync[1647]: [QUORUM] Quorum provider: corosync_votequorum failed to initialize. | |
814 | corosync[1647]: [SERV ] Service engine 'corosync_quorum' failed to load for reason | |
815 | 'configuration error: nodelist or quorum.expected_votes must be configured!' | |
816 | [...] | |
817 | ---- | |
818 | ||
819 | It means that the hostname you set for corosync 'ringX_addr' in the | |
820 | configuration could not be resolved. | |
821 | ||
822 | ||
823 | Write Configuration When Not Quorate | |
824 | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ | |
825 | ||
826 | If you need to change '/etc/pve/corosync.conf' on an node with no quorum, and you | |
827 | know what you do, use: | |
828 | [source,bash] | |
829 | ---- | |
830 | pvecm expected 1 | |
831 | ---- | |
832 | ||
833 | This sets the expected vote count to 1 and makes the cluster quorate. You can | |
834 | now fix your configuration, or revert it back to the last working backup. | |
835 | ||
836 | This is not enough if corosync cannot start anymore. Here its best to edit the | |
837 | local copy of the corosync configuration in '/etc/corosync/corosync.conf' so | |
838 | that corosync can start again. Ensure that on all nodes this configuration has | |
839 | the same content to avoid split brains. If you are not sure what went wrong | |
840 | it's best to ask the Proxmox Community to help you. | |
841 | ||
842 | ||
843 | [[corosync-conf-glossary]] | |
844 | Corosync Configuration Glossary | |
845 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ | |
846 | ||
847 | ringX_addr:: | |
848 | This names the different ring addresses for the corosync totem rings used for | |
849 | the cluster communication. | |
850 | ||
851 | bindnetaddr:: | |
852 | Defines to which interface the ring should bind to. It may be any address of | |
853 | the subnet configured on the interface we want to use. In general its the | |
854 | recommended to just use an address a node uses on this interface. | |
855 | ||
856 | rrp_mode:: | |
857 | Specifies the mode of the redundant ring protocol and may be passive, active or | |
858 | none. Note that use of active is highly experimental and not official | |
859 | supported. Passive is the preferred mode, it may double the cluster | |
860 | communication throughput and increases availability. | |
861 | ||
862 | ||
863 | Cluster Cold Start | |
864 | ------------------ | |
865 | ||
866 | It is obvious that a cluster is not quorate when all nodes are | |
867 | offline. This is a common case after a power failure. | |
868 | ||
869 | NOTE: It is always a good idea to use an uninterruptible power supply | |
870 | (``UPS'', also called ``battery backup'') to avoid this state, especially if | |
871 | you want HA. | |
872 | ||
873 | On node startup, service `pve-manager` is started and waits for | |
874 | quorum. Once quorate, it starts all guests which have the `onboot` | |
875 | flag set. | |
876 | ||
877 | When you turn on nodes, or when power comes back after power failure, | |
878 | it is likely that some nodes boots faster than others. Please keep in | |
879 | mind that guest startup is delayed until you reach quorum. | |
880 | ||
881 | ||
882 | ifdef::manvolnum[] | |
883 | include::pve-copyright.adoc[] | |
884 | endif::manvolnum[] |