]> git.proxmox.com Git - pve-docs.git/blob - pvecm.adoc
btrfs: Add link to upstream docs regarding its status
[pve-docs.git] / pvecm.adoc
1 [[chapter_pvecm]]
2 ifdef::manvolnum[]
3 pvecm(1)
4 ========
5 :pve-toplevel:
6
7 NAME
8 ----
9
10 pvecm - Proxmox VE Cluster Manager
11
12 SYNOPSIS
13 --------
14
15 include::pvecm.1-synopsis.adoc[]
16
17 DESCRIPTION
18 -----------
19 endif::manvolnum[]
20
21 ifndef::manvolnum[]
22 Cluster Manager
23 ===============
24 :pve-toplevel:
25 endif::manvolnum[]
26
27 The {PVE} cluster manager `pvecm` is a tool to create a group of
28 physical servers. Such a group is called a *cluster*. We use the
29 http://www.corosync.org[Corosync Cluster Engine] for reliable group
30 communication, and such clusters can consist of up to 32 physical nodes
31 (probably more, dependent on network latency).
32
33 `pvecm` can be used to create a new cluster, join nodes to a cluster,
34 leave the cluster, get status information and do various other cluster
35 related tasks. The **P**rox**m**o**x** **C**luster **F**ile **S**ystem (``pmxcfs'')
36 is used to transparently distribute the cluster configuration to all cluster
37 nodes.
38
39 Grouping nodes into a cluster has the following advantages:
40
41 * Centralized, web based management
42
43 * Multi-master clusters: each node can do all management task
44
45 * `pmxcfs`: database-driven file system for storing configuration files,
46 replicated in real-time on all nodes using `corosync`.
47
48 * Easy migration of virtual machines and containers between physical
49 hosts
50
51 * Fast deployment
52
53 * Cluster-wide services like firewall and HA
54
55
56 Requirements
57 ------------
58
59 * All nodes must be in the same network as `corosync` uses IP Multicast
60 to communicate between nodes (also see
61 http://www.corosync.org[Corosync Cluster Engine]). Corosync uses UDP
62 ports 5404 and 5405 for cluster communication.
63 +
64 NOTE: Some switches do not support IP multicast by default and must be
65 manually enabled first.
66
67 * Date and time have to be synchronized.
68
69 * SSH tunnel on TCP port 22 between nodes is used.
70
71 * If you are interested in High Availability, you need to have at
72 least three nodes for reliable quorum. All nodes should have the
73 same version.
74
75 * We recommend a dedicated NIC for the cluster traffic, especially if
76 you use shared storage.
77
78 * Root password of a cluster node is required for adding nodes.
79
80 NOTE: It is not possible to mix {pve} 3.x and earlier with {pve} 4.X cluster
81 nodes.
82
83 NOTE: While it's possible for {pve} 4.4 and {pve} 5.0 this is not supported as
84 production configuration and should only used temporarily during upgrading the
85 whole cluster from one to another major version.
86
87
88 Preparing Nodes
89 ---------------
90
91 First, install {PVE} on all nodes. Make sure that each node is
92 installed with the final hostname and IP configuration. Changing the
93 hostname and IP is not possible after cluster creation.
94
95 Currently the cluster creation can either be done on the console (login via
96 `ssh`) or the API, which we have a GUI implementation for (__Datacenter ->
97 Cluster__).
98
99 While it's often common use to reference all other nodenames in `/etc/hosts`
100 with their IP this is not strictly necessary for a cluster, which normally uses
101 multicast, to work. It maybe useful as you then can connect from one node to
102 the other with SSH through the easier to remember node name.
103
104 [[pvecm_create_cluster]]
105 Create the Cluster
106 ------------------
107
108 Login via `ssh` to the first {pve} node. Use a unique name for your cluster.
109 This name cannot be changed later. The cluster name follows the same rules as
110 node names.
111
112 ----
113 hp1# pvecm create CLUSTERNAME
114 ----
115
116 CAUTION: The cluster name is used to compute the default multicast address.
117 Please use unique cluster names if you run more than one cluster inside your
118 network. To avoid human confusion, it is also recommended to choose different
119 names even if clusters do not share the cluster network.
120
121 To check the state of your cluster use:
122
123 ----
124 hp1# pvecm status
125 ----
126
127 Multiple Clusters In Same Network
128 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
129
130 It is possible to create multiple clusters in the same physical or logical
131 network. Each cluster must have a unique name, which is used to generate the
132 cluster's multicast group address. As long as no duplicate cluster names are
133 configured in one network segment, the different clusters won't interfere with
134 each other.
135
136 If multiple clusters operate in a single network it may be beneficial to setup
137 an IGMP querier and enable IGMP Snooping in said network. This may reduce the
138 load of the network significantly because multicast packets are only delivered
139 to endpoints of the respective member nodes.
140
141
142 [[pvecm_join_node_to_cluster]]
143 Adding Nodes to the Cluster
144 ---------------------------
145
146 Login via `ssh` to the node you want to add.
147
148 ----
149 hp2# pvecm add IP-ADDRESS-CLUSTER
150 ----
151
152 For `IP-ADDRESS-CLUSTER` use the IP from an existing cluster node.
153
154 CAUTION: A new node cannot hold any VMs, because you would get
155 conflicts about identical VM IDs. Also, all existing configuration in
156 `/etc/pve` is overwritten when you join a new node to the cluster. To
157 workaround, use `vzdump` to backup and restore to a different VMID after
158 adding the node to the cluster.
159
160 To check the state of cluster:
161
162 ----
163 # pvecm status
164 ----
165
166 .Cluster status after adding 4 nodes
167 ----
168 hp2# pvecm status
169 Quorum information
170 ~~~~~~~~~~~~~~~~~~
171 Date: Mon Apr 20 12:30:13 2015
172 Quorum provider: corosync_votequorum
173 Nodes: 4
174 Node ID: 0x00000001
175 Ring ID: 1928
176 Quorate: Yes
177
178 Votequorum information
179 ~~~~~~~~~~~~~~~~~~~~~~
180 Expected votes: 4
181 Highest expected: 4
182 Total votes: 4
183 Quorum: 3
184 Flags: Quorate
185
186 Membership information
187 ~~~~~~~~~~~~~~~~~~~~~~
188 Nodeid Votes Name
189 0x00000001 1 192.168.15.91
190 0x00000002 1 192.168.15.92 (local)
191 0x00000003 1 192.168.15.93
192 0x00000004 1 192.168.15.94
193 ----
194
195 If you only want the list of all nodes use:
196
197 ----
198 # pvecm nodes
199 ----
200
201 .List nodes in a cluster
202 ----
203 hp2# pvecm nodes
204
205 Membership information
206 ~~~~~~~~~~~~~~~~~~~~~~
207 Nodeid Votes Name
208 1 1 hp1
209 2 1 hp2 (local)
210 3 1 hp3
211 4 1 hp4
212 ----
213
214 [[adding-nodes-with-separated-cluster-network]]
215 Adding Nodes With Separated Cluster Network
216 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
217
218 When adding a node to a cluster with a separated cluster network you need to
219 use the 'ringX_addr' parameters to set the nodes address on those networks:
220
221 [source,bash]
222 ----
223 pvecm add IP-ADDRESS-CLUSTER -ring0_addr IP-ADDRESS-RING0
224 ----
225
226 If you want to use the Redundant Ring Protocol you will also want to pass the
227 'ring1_addr' parameter.
228
229
230 Remove a Cluster Node
231 ---------------------
232
233 CAUTION: Read carefully the procedure before proceeding, as it could
234 not be what you want or need.
235
236 Move all virtual machines from the node. Make sure you have no local
237 data or backups you want to keep, or save them accordingly.
238 In the following example we will remove the node hp4 from the cluster.
239
240 Log in to a *different* cluster node (not hp4), and issue a `pvecm nodes`
241 command to identify the node ID to remove:
242
243 ----
244 hp1# pvecm nodes
245
246 Membership information
247 ~~~~~~~~~~~~~~~~~~~~~~
248 Nodeid Votes Name
249 1 1 hp1 (local)
250 2 1 hp2
251 3 1 hp3
252 4 1 hp4
253 ----
254
255
256 At this point you must power off hp4 and
257 make sure that it will not power on again (in the network) as it
258 is.
259
260 IMPORTANT: As said above, it is critical to power off the node
261 *before* removal, and make sure that it will *never* power on again
262 (in the existing cluster network) as it is.
263 If you power on the node as it is, your cluster will be screwed up and
264 it could be difficult to restore a clean cluster state.
265
266 After powering off the node hp4, we can safely remove it from the cluster.
267
268 ----
269 hp1# pvecm delnode hp4
270 ----
271
272 If the operation succeeds no output is returned, just check the node
273 list again with `pvecm nodes` or `pvecm status`. You should see
274 something like:
275
276 ----
277 hp1# pvecm status
278
279 Quorum information
280 ~~~~~~~~~~~~~~~~~~
281 Date: Mon Apr 20 12:44:28 2015
282 Quorum provider: corosync_votequorum
283 Nodes: 3
284 Node ID: 0x00000001
285 Ring ID: 1992
286 Quorate: Yes
287
288 Votequorum information
289 ~~~~~~~~~~~~~~~~~~~~~~
290 Expected votes: 3
291 Highest expected: 3
292 Total votes: 3
293 Quorum: 2
294 Flags: Quorate
295
296 Membership information
297 ~~~~~~~~~~~~~~~~~~~~~~
298 Nodeid Votes Name
299 0x00000001 1 192.168.15.90 (local)
300 0x00000002 1 192.168.15.91
301 0x00000003 1 192.168.15.92
302 ----
303
304 If, for whatever reason, you want that this server joins the same
305 cluster again, you have to
306
307 * reinstall {pve} on it from scratch
308
309 * then join it, as explained in the previous section.
310
311 [[pvecm_separate_node_without_reinstall]]
312 Separate A Node Without Reinstalling
313 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
314
315 CAUTION: This is *not* the recommended method, proceed with caution. Use the
316 above mentioned method if you're unsure.
317
318 You can also separate a node from a cluster without reinstalling it from
319 scratch. But after removing the node from the cluster it will still have
320 access to the shared storages! This must be resolved before you start removing
321 the node from the cluster. A {pve} cluster cannot share the exact same
322 storage with another cluster, as storage locking doesn't work over cluster
323 boundary. Further, it may also lead to VMID conflicts.
324
325 Its suggested that you create a new storage where only the node which you want
326 to separate has access. This can be an new export on your NFS or a new Ceph
327 pool, to name a few examples. Its just important that the exact same storage
328 does not gets accessed by multiple clusters. After setting this storage up move
329 all data from the node and its VMs to it. Then you are ready to separate the
330 node from the cluster.
331
332 WARNING: Ensure all shared resources are cleanly separated! You will run into
333 conflicts and problems else.
334
335 First stop the corosync and the pve-cluster services on the node:
336 [source,bash]
337 ----
338 systemctl stop pve-cluster
339 systemctl stop corosync
340 ----
341
342 Start the cluster filesystem again in local mode:
343 [source,bash]
344 ----
345 pmxcfs -l
346 ----
347
348 Delete the corosync configuration files:
349 [source,bash]
350 ----
351 rm /etc/pve/corosync.conf
352 rm /etc/corosync/*
353 ----
354
355 You can now start the filesystem again as normal service:
356 [source,bash]
357 ----
358 killall pmxcfs
359 systemctl start pve-cluster
360 ----
361
362 The node is now separated from the cluster. You can deleted it from a remaining
363 node of the cluster with:
364 [source,bash]
365 ----
366 pvecm delnode oldnode
367 ----
368
369 If the command failed, because the remaining node in the cluster lost quorum
370 when the now separate node exited, you may set the expected votes to 1 as a workaround:
371 [source,bash]
372 ----
373 pvecm expected 1
374 ----
375
376 And then repeat the 'pvecm delnode' command.
377
378 Now switch back to the separated node, here delete all remaining files left
379 from the old cluster. This ensures that the node can be added to another
380 cluster again without problems.
381
382 [source,bash]
383 ----
384 rm /var/lib/corosync/*
385 ----
386
387 As the configuration files from the other nodes are still in the cluster
388 filesystem you may want to clean those up too. Remove simply the whole
389 directory recursive from '/etc/pve/nodes/NODENAME', but check three times that
390 you used the correct one before deleting it.
391
392 CAUTION: The nodes SSH keys are still in the 'authorized_key' file, this means
393 the nodes can still connect to each other with public key authentication. This
394 should be fixed by removing the respective keys from the
395 '/etc/pve/priv/authorized_keys' file.
396
397 Quorum
398 ------
399
400 {pve} use a quorum-based technique to provide a consistent state among
401 all cluster nodes.
402
403 [quote, from Wikipedia, Quorum (distributed computing)]
404 ____
405 A quorum is the minimum number of votes that a distributed transaction
406 has to obtain in order to be allowed to perform an operation in a
407 distributed system.
408 ____
409
410 In case of network partitioning, state changes requires that a
411 majority of nodes are online. The cluster switches to read-only mode
412 if it loses quorum.
413
414 NOTE: {pve} assigns a single vote to each node by default.
415
416 Cluster Network
417 ---------------
418
419 The cluster network is the core of a cluster. All messages sent over it have to
420 be delivered reliable to all nodes in their respective order. In {pve} this
421 part is done by corosync, an implementation of a high performance low overhead
422 high availability development toolkit. It serves our decentralized
423 configuration file system (`pmxcfs`).
424
425 [[cluster-network-requirements]]
426 Network Requirements
427 ~~~~~~~~~~~~~~~~~~~~
428 This needs a reliable network with latencies under 2 milliseconds (LAN
429 performance) to work properly. While corosync can also use unicast for
430 communication between nodes its **highly recommended** to have a multicast
431 capable network. The network should not be used heavily by other members,
432 ideally corosync runs on its own network.
433 *never* share it with network where storage communicates too.
434
435 Before setting up a cluster it is good practice to check if the network is fit
436 for that purpose.
437
438 * Ensure that all nodes are in the same subnet. This must only be true for the
439 network interfaces used for cluster communication (corosync).
440
441 * Ensure all nodes can reach each other over those interfaces, using `ping` is
442 enough for a basic test.
443
444 * Ensure that multicast works in general and a high package rates. This can be
445 done with the `omping` tool. The final "%loss" number should be < 1%.
446 +
447 [source,bash]
448 ----
449 omping -c 10000 -i 0.001 -F -q NODE1-IP NODE2-IP ...
450 ----
451
452 * Ensure that multicast communication works over an extended period of time.
453 This uncovers problems where IGMP snooping is activated on the network but
454 no multicast querier is active. This test has a duration of around 10
455 minutes.
456 +
457 [source,bash]
458 ----
459 omping -c 600 -i 1 -q NODE1-IP NODE2-IP ...
460 ----
461
462 Your network is not ready for clustering if any of these test fails. Recheck
463 your network configuration. Especially switches are notorious for having
464 multicast disabled by default or IGMP snooping enabled with no IGMP querier
465 active.
466
467 In smaller cluster its also an option to use unicast if you really cannot get
468 multicast to work.
469
470 Separate Cluster Network
471 ~~~~~~~~~~~~~~~~~~~~~~~~
472
473 When creating a cluster without any parameters the cluster network is generally
474 shared with the Web UI and the VMs and its traffic. Depending on your setup
475 even storage traffic may get sent over the same network. Its recommended to
476 change that, as corosync is a time critical real time application.
477
478 Setting Up A New Network
479 ^^^^^^^^^^^^^^^^^^^^^^^^
480
481 First you have to setup a new network interface. It should be on a physical
482 separate network. Ensure that your network fulfills the
483 <<cluster-network-requirements,cluster network requirements>>.
484
485 Separate On Cluster Creation
486 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
487
488 This is possible through the 'ring0_addr' and 'bindnet0_addr' parameter of
489 the 'pvecm create' command used for creating a new cluster.
490
491 If you have setup an additional NIC with a static address on 10.10.10.1/25
492 and want to send and receive all cluster communication over this interface
493 you would execute:
494
495 [source,bash]
496 ----
497 pvecm create test --ring0_addr 10.10.10.1 --bindnet0_addr 10.10.10.0
498 ----
499
500 To check if everything is working properly execute:
501 [source,bash]
502 ----
503 systemctl status corosync
504 ----
505
506 Afterwards, proceed as descripted in the section to
507 <<adding-nodes-with-separated-cluster-network,add nodes with a separated cluster network>>.
508
509 [[separate-cluster-net-after-creation]]
510 Separate After Cluster Creation
511 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
512
513 You can do this also if you have already created a cluster and want to switch
514 its communication to another network, without rebuilding the whole cluster.
515 This change may lead to short durations of quorum loss in the cluster, as nodes
516 have to restart corosync and come up one after the other on the new network.
517
518 Check how to <<edit-corosync-conf,edit the corosync.conf file>> first.
519 The open it and you should see a file similar to:
520
521 ----
522 logging {
523 debug: off
524 to_syslog: yes
525 }
526
527 nodelist {
528
529 node {
530 name: due
531 nodeid: 2
532 quorum_votes: 1
533 ring0_addr: due
534 }
535
536 node {
537 name: tre
538 nodeid: 3
539 quorum_votes: 1
540 ring0_addr: tre
541 }
542
543 node {
544 name: uno
545 nodeid: 1
546 quorum_votes: 1
547 ring0_addr: uno
548 }
549
550 }
551
552 quorum {
553 provider: corosync_votequorum
554 }
555
556 totem {
557 cluster_name: thomas-testcluster
558 config_version: 3
559 ip_version: ipv4
560 secauth: on
561 version: 2
562 interface {
563 bindnetaddr: 192.168.30.50
564 ringnumber: 0
565 }
566
567 }
568 ----
569
570 The first you want to do is add the 'name' properties in the node entries if
571 you do not see them already. Those *must* match the node name.
572
573 Then replace the address from the 'ring0_addr' properties with the new
574 addresses. You may use plain IP addresses or also hostnames here. If you use
575 hostnames ensure that they are resolvable from all nodes.
576
577 In my example I want to switch my cluster communication to the 10.10.10.1/25
578 network. So I replace all 'ring0_addr' respectively. I also set the bindnetaddr
579 in the totem section of the config to an address of the new network. It can be
580 any address from the subnet configured on the new network interface.
581
582 After you increased the 'config_version' property the new configuration file
583 should look like:
584
585 ----
586
587 logging {
588 debug: off
589 to_syslog: yes
590 }
591
592 nodelist {
593
594 node {
595 name: due
596 nodeid: 2
597 quorum_votes: 1
598 ring0_addr: 10.10.10.2
599 }
600
601 node {
602 name: tre
603 nodeid: 3
604 quorum_votes: 1
605 ring0_addr: 10.10.10.3
606 }
607
608 node {
609 name: uno
610 nodeid: 1
611 quorum_votes: 1
612 ring0_addr: 10.10.10.1
613 }
614
615 }
616
617 quorum {
618 provider: corosync_votequorum
619 }
620
621 totem {
622 cluster_name: thomas-testcluster
623 config_version: 4
624 ip_version: ipv4
625 secauth: on
626 version: 2
627 interface {
628 bindnetaddr: 10.10.10.1
629 ringnumber: 0
630 }
631
632 }
633 ----
634
635 Now after a final check whether all changed information is correct we save it
636 and see again the <<edit-corosync-conf,edit corosync.conf file>> section to
637 learn how to bring it in effect.
638
639 As our change cannot be enforced live from corosync we have to do an restart.
640
641 On a single node execute:
642 [source,bash]
643 ----
644 systemctl restart corosync
645 ----
646
647 Now check if everything is fine:
648
649 [source,bash]
650 ----
651 systemctl status corosync
652 ----
653
654 If corosync runs again correct restart corosync also on all other nodes.
655 They will then join the cluster membership one by one on the new network.
656
657 [[pvecm_rrp]]
658 Redundant Ring Protocol
659 ~~~~~~~~~~~~~~~~~~~~~~~
660 To avoid a single point of failure you should implement counter measurements.
661 This can be on the hardware and operating system level through network bonding.
662
663 Corosync itself offers also a possibility to add redundancy through the so
664 called 'Redundant Ring Protocol'. This protocol allows running a second totem
665 ring on another network, this network should be physically separated from the
666 other rings network to actually increase availability.
667
668 RRP On Cluster Creation
669 ~~~~~~~~~~~~~~~~~~~~~~~
670
671 The 'pvecm create' command provides the additional parameters 'bindnetX_addr',
672 'ringX_addr' and 'rrp_mode', can be used for RRP configuration.
673
674 NOTE: See the <<corosync-conf-glossary,glossary>> if you do not know what each parameter means.
675
676 So if you have two networks, one on the 10.10.10.1/24 and the other on the
677 10.10.20.1/24 subnet you would execute:
678
679 [source,bash]
680 ----
681 pvecm create CLUSTERNAME -bindnet0_addr 10.10.10.1 -ring0_addr 10.10.10.1 \
682 -bindnet1_addr 10.10.20.1 -ring1_addr 10.10.20.1
683 ----
684
685 RRP On Existing Clusters
686 ~~~~~~~~~~~~~~~~~~~~~~~~
687
688 You will take similar steps as described in
689 <<separate-cluster-net-after-creation,separating the cluster network>> to
690 enable RRP on an already running cluster. The single difference is, that you
691 will add `ring1` and use it instead of `ring0`.
692
693 First add a new `interface` subsection in the `totem` section, set its
694 `ringnumber` property to `1`. Set the interfaces `bindnetaddr` property to an
695 address of the subnet you have configured for your new ring.
696 Further set the `rrp_mode` to `passive`, this is the only stable mode.
697
698 Then add to each node entry in the `nodelist` section its new `ring1_addr`
699 property with the nodes additional ring address.
700
701 So if you have two networks, one on the 10.10.10.1/24 and the other on the
702 10.10.20.1/24 subnet, the final configuration file should look like:
703
704 ----
705 totem {
706 cluster_name: tweak
707 config_version: 9
708 ip_version: ipv4
709 rrp_mode: passive
710 secauth: on
711 version: 2
712 interface {
713 bindnetaddr: 10.10.10.1
714 ringnumber: 0
715 }
716 interface {
717 bindnetaddr: 10.10.20.1
718 ringnumber: 1
719 }
720 }
721
722 nodelist {
723 node {
724 name: pvecm1
725 nodeid: 1
726 quorum_votes: 1
727 ring0_addr: 10.10.10.1
728 ring1_addr: 10.10.20.1
729 }
730
731 node {
732 name: pvecm2
733 nodeid: 2
734 quorum_votes: 1
735 ring0_addr: 10.10.10.2
736 ring1_addr: 10.10.20.2
737 }
738
739 [...] # other cluster nodes here
740 }
741
742 [...] # other remaining config sections here
743
744 ----
745
746 Bring it in effect like described in the
747 <<edit-corosync-conf,edit the corosync.conf file>> section.
748
749 This is a change which cannot take live in effect and needs at least a restart
750 of corosync. Recommended is a restart of the whole cluster.
751
752 If you cannot reboot the whole cluster ensure no High Availability services are
753 configured and the stop the corosync service on all nodes. After corosync is
754 stopped on all nodes start it one after the other again.
755
756 Corosync External Vote Support
757 ------------------------------
758
759 This section describes a way to deploy an external voter in a {pve} cluster.
760 When configured, the cluster can sustain more node failures without
761 violating safety properties of the cluster communication.
762
763 For this to work there are two services involved:
764
765 * a so called qdevice daemon which runs on each {pve} node
766
767 * an external vote daemon which runs on an independent server.
768
769 As a result you can achieve higher availability even in smaller setups (for
770 example 2+1 nodes).
771
772 QDevice Technical Overview
773 ~~~~~~~~~~~~~~~~~~~~~~~~~~
774
775 The Corosync Quroum Device (QDevice) is a daemon which runs on each cluster
776 node. It provides a configured number of votes to the clusters quorum
777 subsystem based on an external running third-party arbitrator's decision.
778 Its primary use is to allow a cluster to sustain more node failures than
779 standard quorum rules allow. This can be done safely as the external device
780 can see all nodes and thus choose only one set of nodes to give its vote.
781 This will only be done if said set of nodes can have quorum (again) when
782 receiving the third-party vote.
783
784 Currently only 'QDevice Net' is supported as a third-party arbitrator. It is
785 a daemon which provides a vote to a cluster partition if it can reach the
786 partition members over the network. It will give only votes to one partition
787 of a cluster at any time.
788 It's designed to support multiple clusters and is almost configuration and
789 state free. New clusters are handled dynamically and no configuration file
790 is needed on the host running a QDevice.
791
792 The external host has the only requirement that it needs network access to the
793 cluster and a corosync-qnetd package available. We provide such a package
794 for Debian based hosts, other Linux distributions should also have a package
795 available through their respective package manager.
796
797 NOTE: In contrast to corosync itself, a QDevice connects to the cluster over
798 TCP/IP and thus does not need a multicast capable network between itself and
799 the cluster. In fact the daemon may run outside of the LAN and can have
800 longer latencies than 2 ms.
801
802
803 Supported Setups
804 ~~~~~~~~~~~~~~~~
805
806 We support QDevices for clusters with an even number of nodes and recommend
807 it for 2 node clusters, if they should provide higher availability.
808 For clusters with an odd node count we discourage the use of QDevices
809 currently. The reason for this, is the difference of the votes the QDevice
810 provides for each cluster type. Even numbered clusters get single additional
811 vote, with this we can only increase availability, i.e. if the QDevice
812 itself fails we are in the same situation as with no QDevice at all.
813
814 Now, with an odd numbered cluster size the QDevice provides '(N-1)' votes --
815 where 'N' corresponds to the cluster node count. This difference makes
816 sense, if we had only one additional vote the cluster can get into a split
817 brain situation.
818 This algorithm would allow that all nodes but one (and naturally the
819 QDevice itself) could fail.
820 There are two drawbacks with this:
821
822 * If the QNet daemon itself fails, no other node may fail or the cluster
823 immediately loses quorum. For example, in a cluster with 15 nodes 7
824 could fail before the cluster becomes inquorate. But, if a QDevice is
825 configured here and said QDevice fails itself **no single node** of
826 the 15 may fail. The QDevice acts almost as a single point of failure in
827 this case.
828
829 * The fact that all but one node plus QDevice may fail sound promising at
830 first, but this may result in a mass recovery of HA services that would
831 overload the single node left. Also ceph server will stop to provide
832 services after only '((N-1)/2)' nodes are online.
833
834 If you understand the drawbacks and implications you can decide yourself if
835 you should use this technology in an odd numbered cluster setup.
836
837
838 QDevice-Net Setup
839 ~~~~~~~~~~~~~~~~~
840
841 We recommend to run any daemon which provides votes to corosync-qdevice as an
842 unprivileged user. {pve} and Debian Stretch provide a package which is
843 already configured to do so.
844 The traffic between the daemon and the cluster must be encrypted to ensure a
845 safe and secure QDevice integration in {pve}.
846
847 First install the 'corosync-qnetd' package on your external server and
848 the 'corosync-qdevice' package on all cluster nodes.
849
850 After that, ensure that all your nodes on the cluster are online.
851
852 You can now easily set up your QDevice by running the following command on one
853 of the {pve} nodes:
854
855 ----
856 pve# pvecm qdevice setup <QDEVICE-IP>
857 ----
858
859 The SSH key from the cluster will be automatically copied to the QDevice. You
860 might need to enter an SSH password during this step.
861
862 After you enter the password and all the steps are successfully completed, you
863 will see "Done". You can check the status now:
864
865 ----
866 pve# pvecm status
867
868 ...
869
870 Votequorum information
871 ~~~~~~~~~~~~~~~~~~~~~
872 Expected votes: 3
873 Highest expected: 3
874 Total votes: 3
875 Quorum: 2
876 Flags: Quorate Qdevice
877
878 Membership information
879 ~~~~~~~~~~~~~~~~~~~~~~
880 Nodeid Votes Qdevice Name
881 0x00000001 1 A,V,NMW 192.168.22.180 (local)
882 0x00000002 1 A,V,NMW 192.168.22.181
883 0x00000000 1 Qdevice
884
885 ----
886
887 which means the QDevice is set up.
888
889
890 Frequently Asked Questions
891 ~~~~~~~~~~~~~~~~~~~~~~~~~~
892
893 Tie Breaking
894 ^^^^^^^^^^^^
895
896 In case of a tie, where two same-sized cluster partitions cannot see each other
897 but the QDevice, the QDevice chooses randomly one of those partitions and
898 provides a vote to it.
899
900 Possible Negative Implications
901 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
902
903 For clusters with an even node count there are no negative implications when
904 setting up a QDevice. If it fails to work, you are as good as without QDevice at
905 all.
906
907 Adding/Deleting Nodes After QDevice Setup
908 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
909
910 If you want to add a new node or remove an existing one from a cluster with a
911 QDevice setup, you need to remove the QDevice first. After that, you can add or
912 remove nodes normally. Once you have a cluster with an even node count again,
913 you can set up the QDevice again as described above.
914
915 Removing the QDevice
916 ^^^^^^^^^^^^^^^^^^^^
917
918 If you used the official `pvecm` tool to add the QDevice, you can remove it
919 trivially by running:
920
921 ----
922 pve# pvecm qdevice remove
923 ----
924
925 //Still TODO
926 //^^^^^^^^^^
927 //There ist still stuff to add here
928
929
930 Corosync Configuration
931 ----------------------
932
933 The `/etc/pve/corosync.conf` file plays a central role in {pve} cluster. It
934 controls the cluster member ship and its network.
935 For reading more about it check the corosync.conf man page:
936 [source,bash]
937 ----
938 man corosync.conf
939 ----
940
941 For node membership you should always use the `pvecm` tool provided by {pve}.
942 You may have to edit the configuration file manually for other changes.
943 Here are a few best practice tips for doing this.
944
945 [[edit-corosync-conf]]
946 Edit corosync.conf
947 ~~~~~~~~~~~~~~~~~~
948
949 Editing the corosync.conf file can be not always straight forward. There are
950 two on each cluster, one in `/etc/pve/corosync.conf` and the other in
951 `/etc/corosync/corosync.conf`. Editing the one in our cluster file system will
952 propagate the changes to the local one, but not vice versa.
953
954 The configuration will get updated automatically as soon as the file changes.
955 This means changes which can be integrated in a running corosync will take
956 instantly effect. So you should always make a copy and edit that instead, to
957 avoid triggering some unwanted changes by an in between safe.
958
959 [source,bash]
960 ----
961 cp /etc/pve/corosync.conf /etc/pve/corosync.conf.new
962 ----
963
964 Then open the Config file with your favorite editor, `nano` and `vim.tiny` are
965 preinstalled on {pve} for example.
966
967 NOTE: Always increment the 'config_version' number on configuration changes,
968 omitting this can lead to problems.
969
970 After making the necessary changes create another copy of the current working
971 configuration file. This serves as a backup if the new configuration fails to
972 apply or makes problems in other ways.
973
974 [source,bash]
975 ----
976 cp /etc/pve/corosync.conf /etc/pve/corosync.conf.bak
977 ----
978
979 Then move the new configuration file over the old one:
980 [source,bash]
981 ----
982 mv /etc/pve/corosync.conf.new /etc/pve/corosync.conf
983 ----
984
985 You may check with the commands
986 [source,bash]
987 ----
988 systemctl status corosync
989 journalctl -b -u corosync
990 ----
991
992 If the change could applied automatically. If not you may have to restart the
993 corosync service via:
994 [source,bash]
995 ----
996 systemctl restart corosync
997 ----
998
999 On errors check the troubleshooting section below.
1000
1001 Troubleshooting
1002 ~~~~~~~~~~~~~~~
1003
1004 Issue: 'quorum.expected_votes must be configured'
1005 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1006
1007 When corosync starts to fail and you get the following message in the system log:
1008
1009 ----
1010 [...]
1011 corosync[1647]: [QUORUM] Quorum provider: corosync_votequorum failed to initialize.
1012 corosync[1647]: [SERV ] Service engine 'corosync_quorum' failed to load for reason
1013 'configuration error: nodelist or quorum.expected_votes must be configured!'
1014 [...]
1015 ----
1016
1017 It means that the hostname you set for corosync 'ringX_addr' in the
1018 configuration could not be resolved.
1019
1020
1021 Write Configuration When Not Quorate
1022 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1023
1024 If you need to change '/etc/pve/corosync.conf' on an node with no quorum, and you
1025 know what you do, use:
1026 [source,bash]
1027 ----
1028 pvecm expected 1
1029 ----
1030
1031 This sets the expected vote count to 1 and makes the cluster quorate. You can
1032 now fix your configuration, or revert it back to the last working backup.
1033
1034 This is not enough if corosync cannot start anymore. Here its best to edit the
1035 local copy of the corosync configuration in '/etc/corosync/corosync.conf' so
1036 that corosync can start again. Ensure that on all nodes this configuration has
1037 the same content to avoid split brains. If you are not sure what went wrong
1038 it's best to ask the Proxmox Community to help you.
1039
1040
1041 [[corosync-conf-glossary]]
1042 Corosync Configuration Glossary
1043 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1044
1045 ringX_addr::
1046 This names the different ring addresses for the corosync totem rings used for
1047 the cluster communication.
1048
1049 bindnetaddr::
1050 Defines to which interface the ring should bind to. It may be any address of
1051 the subnet configured on the interface we want to use. In general its the
1052 recommended to just use an address a node uses on this interface.
1053
1054 rrp_mode::
1055 Specifies the mode of the redundant ring protocol and may be passive, active or
1056 none. Note that use of active is highly experimental and not official
1057 supported. Passive is the preferred mode, it may double the cluster
1058 communication throughput and increases availability.
1059
1060
1061 Cluster Cold Start
1062 ------------------
1063
1064 It is obvious that a cluster is not quorate when all nodes are
1065 offline. This is a common case after a power failure.
1066
1067 NOTE: It is always a good idea to use an uninterruptible power supply
1068 (``UPS'', also called ``battery backup'') to avoid this state, especially if
1069 you want HA.
1070
1071 On node startup, the `pve-guests` service is started and waits for
1072 quorum. Once quorate, it starts all guests which have the `onboot`
1073 flag set.
1074
1075 When you turn on nodes, or when power comes back after power failure,
1076 it is likely that some nodes boots faster than others. Please keep in
1077 mind that guest startup is delayed until you reach quorum.
1078
1079
1080 Guest Migration
1081 ---------------
1082
1083 Migrating virtual guests to other nodes is a useful feature in a
1084 cluster. There are settings to control the behavior of such
1085 migrations. This can be done via the configuration file
1086 `datacenter.cfg` or for a specific migration via API or command line
1087 parameters.
1088
1089 It makes a difference if a Guest is online or offline, or if it has
1090 local resources (like a local disk).
1091
1092 For Details about Virtual Machine Migration see the
1093 xref:qm_migration[QEMU/KVM Migration Chapter]
1094
1095 For Details about Container Migration see the
1096 xref:pct_migration[Container Migration Chapter]
1097
1098 Migration Type
1099 ~~~~~~~~~~~~~~
1100
1101 The migration type defines if the migration data should be sent over an
1102 encrypted (`secure`) channel or an unencrypted (`insecure`) one.
1103 Setting the migration type to insecure means that the RAM content of a
1104 virtual guest gets also transferred unencrypted, which can lead to
1105 information disclosure of critical data from inside the guest (for
1106 example passwords or encryption keys).
1107
1108 Therefore, we strongly recommend using the secure channel if you do
1109 not have full control over the network and can not guarantee that no
1110 one is eavesdropping to it.
1111
1112 NOTE: Storage migration does not follow this setting. Currently, it
1113 always sends the storage content over a secure channel.
1114
1115 Encryption requires a lot of computing power, so this setting is often
1116 changed to "unsafe" to achieve better performance. The impact on
1117 modern systems is lower because they implement AES encryption in
1118 hardware. The performance impact is particularly evident in fast
1119 networks where you can transfer 10 Gbps or more.
1120
1121
1122 Migration Network
1123 ~~~~~~~~~~~~~~~~~
1124
1125 By default, {pve} uses the network in which cluster communication
1126 takes place to send the migration traffic. This is not optimal because
1127 sensitive cluster traffic can be disrupted and this network may not
1128 have the best bandwidth available on the node.
1129
1130 Setting the migration network parameter allows the use of a dedicated
1131 network for the entire migration traffic. In addition to the memory,
1132 this also affects the storage traffic for offline migrations.
1133
1134 The migration network is set as a network in the CIDR notation. This
1135 has the advantage that you do not have to set individual IP addresses
1136 for each node. {pve} can determine the real address on the
1137 destination node from the network specified in the CIDR form. To
1138 enable this, the network must be specified so that each node has one,
1139 but only one IP in the respective network.
1140
1141
1142 Example
1143 ^^^^^^^
1144
1145 We assume that we have a three-node setup with three separate
1146 networks. One for public communication with the Internet, one for
1147 cluster communication and a very fast one, which we want to use as a
1148 dedicated network for migration.
1149
1150 A network configuration for such a setup might look as follows:
1151
1152 ----
1153 iface eno1 inet manual
1154
1155 # public network
1156 auto vmbr0
1157 iface vmbr0 inet static
1158 address 192.X.Y.57
1159 netmask 255.255.250.0
1160 gateway 192.X.Y.1
1161 bridge_ports eno1
1162 bridge_stp off
1163 bridge_fd 0
1164
1165 # cluster network
1166 auto eno2
1167 iface eno2 inet static
1168 address 10.1.1.1
1169 netmask 255.255.255.0
1170
1171 # fast network
1172 auto eno3
1173 iface eno3 inet static
1174 address 10.1.2.1
1175 netmask 255.255.255.0
1176 ----
1177
1178 Here, we will use the network 10.1.2.0/24 as a migration network. For
1179 a single migration, you can do this using the `migration_network`
1180 parameter of the command line tool:
1181
1182 ----
1183 # qm migrate 106 tre --online --migration_network 10.1.2.0/24
1184 ----
1185
1186 To configure this as the default network for all migrations in the
1187 cluster, set the `migration` property of the `/etc/pve/datacenter.cfg`
1188 file:
1189
1190 ----
1191 # use dedicated migration network
1192 migration: secure,network=10.1.2.0/24
1193 ----
1194
1195 NOTE: The migration type must always be set when the migration network
1196 gets set in `/etc/pve/datacenter.cfg`.
1197
1198
1199 ifdef::manvolnum[]
1200 include::pve-copyright.adoc[]
1201 endif::manvolnum[]