10 pvecm - Proxmox VE Cluster Manager
15 include::pvecm.1-synopsis.adoc[]
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. There's no explicit limit for the number of nodes in a cluster.
31 In practice, the actual possible node count may be limited by the host and
32 network performance. Currently (2021), there are reports of clusters (using
33 high-end enterprise hardware) with over 50 nodes in production.
35 `pvecm` can be used to create a new cluster, join nodes to a cluster,
36 leave the cluster, get status information, and do various other cluster-related
37 tasks. The **P**rox**m**o**x** **C**luster **F**ile **S**ystem (``pmxcfs'')
38 is used to transparently distribute the cluster configuration to all cluster
41 Grouping nodes into a cluster has the following advantages:
43 * Centralized, web-based management
45 * Multi-master clusters: each node can do all management tasks
47 * Use of `pmxcfs`, a database-driven file system, for storing configuration
48 files, replicated in real-time on all nodes using `corosync`
50 * Easy migration of virtual machines and containers between physical
55 * Cluster-wide services like firewall and HA
61 * All nodes must be able to connect to each other via UDP ports 5404 and 5405
64 * Date and time must be synchronized.
66 * An SSH tunnel on TCP port 22 between nodes is required.
68 * If you are interested in High Availability, you need to have at
69 least three nodes for reliable quorum. All nodes should have the
72 * We recommend a dedicated NIC for the cluster traffic, especially if
73 you use shared storage.
75 * The root password of a cluster node is required for adding nodes.
77 * Online migration of virtual machines is only supported when nodes have CPUs
78 from the same vendor. It might work otherwise, but this is never guaranteed.
80 NOTE: It is not possible to mix {pve} 3.x and earlier with {pve} 4.X cluster
83 NOTE: While it's possible to mix {pve} 4.4 and {pve} 5.0 nodes, doing so is
84 not supported as a production configuration and should only be done temporarily,
85 during an upgrade of the whole cluster from one major version to another.
87 NOTE: Running a cluster of {pve} 6.x with earlier versions is not possible. The
88 cluster protocol (corosync) between {pve} 6.x and earlier versions changed
89 fundamentally. The corosync 3 packages for {pve} 5.4 are only intended for the
90 upgrade procedure to {pve} 6.0.
96 First, install {pve} on all nodes. Make sure that each node is
97 installed with the final hostname and IP configuration. Changing the
98 hostname and IP is not possible after cluster creation.
100 While it's common to reference all node names and their IPs in `/etc/hosts` (or
101 make their names resolvable through other means), this is not necessary for a
102 cluster to work. It may be useful however, as you can then connect from one node
103 to another via SSH, using the easier to remember node name (see also
104 xref:pvecm_corosync_addresses[Link Address Types]). Note that we always
105 recommend referencing nodes by their IP addresses in the cluster configuration.
108 [[pvecm_create_cluster]]
112 You can either create a cluster on the console (login via `ssh`), or through
113 the API using the {pve} web interface (__Datacenter -> Cluster__).
115 NOTE: Use a unique name for your cluster. This name cannot be changed later.
116 The cluster name follows the same rules as node names.
118 [[pvecm_cluster_create_via_gui]]
122 [thumbnail="screenshot/gui-cluster-create.png"]
124 Under __Datacenter -> Cluster__, click on *Create Cluster*. Enter the cluster
125 name and select a network connection from the drop-down list to serve as the
126 main cluster network (Link 0). It defaults to the IP resolved via the node's
129 As of {pve} 6.2, up to 8 fallback links can be added to a cluster. To add a
130 redundant link, click the 'Add' button and select a link number and IP address
131 from the respective fields. Prior to {pve} 6.2, to add a second link as
132 fallback, you can select the 'Advanced' checkbox and choose an additional
133 network interface (Link 1, see also xref:pvecm_redundancy[Corosync Redundancy]).
135 NOTE: Ensure that the network selected for cluster communication is not used for
136 any high traffic purposes, like network storage or live-migration.
137 While the cluster network itself produces small amounts of data, it is very
138 sensitive to latency. Check out full
139 xref:pvecm_cluster_network_requirements[cluster network requirements].
141 [[pvecm_cluster_create_via_cli]]
142 Create via the Command Line
143 ~~~~~~~~~~~~~~~~~~~~~~~~~~~
145 Login via `ssh` to the first {pve} node and run the following command:
148 hp1# pvecm create CLUSTERNAME
151 To check the state of the new cluster use:
157 Multiple Clusters in the Same Network
158 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
160 It is possible to create multiple clusters in the same physical or logical
161 network. In this case, each cluster must have a unique name to avoid possible
162 clashes in the cluster communication stack. Furthermore, this helps avoid human
163 confusion by making clusters clearly distinguishable.
165 While the bandwidth requirement of a corosync cluster is relatively low, the
166 latency of packages and the package per second (PPS) rate is the limiting
167 factor. Different clusters in the same network can compete with each other for
168 these resources, so it may still make sense to use separate physical network
169 infrastructure for bigger clusters.
171 [[pvecm_join_node_to_cluster]]
172 Adding Nodes to the Cluster
173 ---------------------------
175 CAUTION: A node that is about to be added to the cluster cannot hold any guests.
176 All existing configuration in `/etc/pve` is overwritten when joining a cluster,
177 since guest IDs could otherwise conflict. As a workaround, you can create a
178 backup of the guest (`vzdump`) and restore it under a different ID, after the
179 node has been added to the cluster.
181 Join Node to Cluster via GUI
182 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
184 [thumbnail="screenshot/gui-cluster-join-information.png"]
186 Log in to the web interface on an existing cluster node. Under __Datacenter ->
187 Cluster__, click the *Join Information* button at the top. Then, click on the
188 button *Copy Information*. Alternatively, copy the string from the 'Information'
191 [thumbnail="screenshot/gui-cluster-join.png"]
193 Next, log in to the web interface on the node you want to add.
194 Under __Datacenter -> Cluster__, click on *Join Cluster*. Fill in the
195 'Information' field with the 'Join Information' text you copied earlier.
196 Most settings required for joining the cluster will be filled out
197 automatically. For security reasons, the cluster password has to be entered
200 NOTE: To enter all required data manually, you can disable the 'Assisted Join'
203 After clicking the *Join* button, the cluster join process will start
204 immediately. After the node has joined the cluster, its current node certificate
205 will be replaced by one signed from the cluster certificate authority (CA).
206 This means that the current session will stop working after a few seconds. You
207 then might need to force-reload the web interface and log in again with the
210 Now your node should be visible under __Datacenter -> Cluster__.
212 Join Node to Cluster via Command Line
213 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
215 Log in to the node you want to join into an existing cluster via `ssh`.
218 # pvecm add IP-ADDRESS-CLUSTER
221 For `IP-ADDRESS-CLUSTER`, use the IP or hostname of an existing cluster node.
222 An IP address is recommended (see xref:pvecm_corosync_addresses[Link Address Types]).
225 To check the state of the cluster use:
231 .Cluster status after adding 4 nodes
243 Date: Tue Sep 14 11:06:47 2021
244 Quorum provider: corosync_votequorum
250 Votequorum information
251 ~~~~~~~~~~~~~~~~~~~~~~
258 Membership information
259 ~~~~~~~~~~~~~~~~~~~~~~
261 0x00000001 1 192.168.15.91
262 0x00000002 1 192.168.15.92 (local)
263 0x00000003 1 192.168.15.93
264 0x00000004 1 192.168.15.94
267 If you only want a list of all nodes, use:
273 .List nodes in a cluster
277 Membership information
278 ~~~~~~~~~~~~~~~~~~~~~~
286 [[pvecm_adding_nodes_with_separated_cluster_network]]
287 Adding Nodes with Separated Cluster Network
288 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
290 When adding a node to a cluster with a separated cluster network, you need to
291 use the 'link0' parameter to set the nodes address on that network:
295 pvecm add IP-ADDRESS-CLUSTER -link0 LOCAL-IP-ADDRESS-LINK0
298 If you want to use the built-in xref:pvecm_redundancy[redundancy] of the
299 Kronosnet transport layer, also use the 'link1' parameter.
301 Using the GUI, you can select the correct interface from the corresponding
302 'Link X' fields in the *Cluster Join* dialog.
304 Remove a Cluster Node
305 ---------------------
307 CAUTION: Read the procedure carefully before proceeding, as it may
308 not be what you want or need.
310 Move all virtual machines from the node. Ensure that you have made copies of any
311 local data or backups that you want to keep. In addition, make sure to remove
312 any scheduled replication jobs to the node to be removed.
314 CAUTION: Failure to remove replication jobs to a node before removing said node
315 will result in the replication job becoming irremovable. Especially note that
316 replication automatically switches direction if a replicated VM is migrated, so
317 by migrating a replicated VM from a node to be deleted, replication jobs will be
318 set up to that node automatically.
320 In the following example, we will remove the node hp4 from the cluster.
322 Log in to a *different* cluster node (not hp4), and issue a `pvecm nodes`
323 command to identify the node ID to remove:
328 Membership information
329 ~~~~~~~~~~~~~~~~~~~~~~
338 At this point, you must power off hp4 and ensure that it will not power on
339 again (in the network) with its current configuration.
341 IMPORTANT: As mentioned above, it is critical to power off the node
342 *before* removal, and make sure that it will *not* power on again
343 (in the existing cluster network) with its current configuration.
344 If you power on the node as it is, the cluster could end up broken,
345 and it could be difficult to restore it to a functioning state.
347 After powering off the node hp4, we can safely remove it from the cluster.
350 hp1# pvecm delnode hp4
354 NOTE: At this point, it is possible that you will receive an error message
355 stating `Could not kill node (error = CS_ERR_NOT_EXIST)`. This does not
356 signify an actual failure in the deletion of the node, but rather a failure in
357 corosync trying to kill an offline node. Thus, it can be safely ignored.
359 Use `pvecm nodes` or `pvecm status` to check the node list again. It should
367 Votequorum information
368 ~~~~~~~~~~~~~~~~~~~~~~
375 Membership information
376 ~~~~~~~~~~~~~~~~~~~~~~
378 0x00000001 1 192.168.15.90 (local)
379 0x00000002 1 192.168.15.91
380 0x00000003 1 192.168.15.92
383 If, for whatever reason, you want this server to join the same cluster again,
386 * do a fresh install of {pve} on it,
388 * then join it, as explained in the previous section.
390 NOTE: After removal of the node, its SSH fingerprint will still reside in the
391 'known_hosts' of the other nodes. If you receive an SSH error after rejoining
392 a node with the same IP or hostname, run `pvecm updatecerts` once on the
393 re-added node to update its fingerprint cluster wide.
395 [[pvecm_separate_node_without_reinstall]]
396 Separate a Node Without Reinstalling
397 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
399 CAUTION: This is *not* the recommended method, proceed with caution. Use the
400 previous method if you're unsure.
402 You can also separate a node from a cluster without reinstalling it from
403 scratch. But after removing the node from the cluster, it will still have
404 access to any shared storage. This must be resolved before you start removing
405 the node from the cluster. A {pve} cluster cannot share the exact same
406 storage with another cluster, as storage locking doesn't work over the cluster
407 boundary. Furthermore, it may also lead to VMID conflicts.
409 It's suggested that you create a new storage, where only the node which you want
410 to separate has access. This can be a new export on your NFS or a new Ceph
411 pool, to name a few examples. It's just important that the exact same storage
412 does not get accessed by multiple clusters. After setting up this storage, move
413 all data and VMs from the node to it. Then you are ready to separate the
414 node from the cluster.
416 WARNING: Ensure that all shared resources are cleanly separated! Otherwise you
417 will run into conflicts and problems.
419 First, stop the corosync and pve-cluster services on the node:
422 systemctl stop pve-cluster
423 systemctl stop corosync
426 Start the cluster file system again in local mode:
432 Delete the corosync configuration files:
435 rm /etc/pve/corosync.conf
436 rm -r /etc/corosync/*
439 You can now start the file system again as a normal service:
443 systemctl start pve-cluster
446 The node is now separated from the cluster. You can deleted it from any
447 remaining node of the cluster with:
450 pvecm delnode oldnode
453 If the command fails due to a loss of quorum in the remaining node, you can set
454 the expected votes to 1 as a workaround:
460 And then repeat the 'pvecm delnode' command.
462 Now switch back to the separated node and delete all the remaining cluster
463 files on it. This ensures that the node can be added to another cluster again
468 rm /var/lib/corosync/*
471 As the configuration files from the other nodes are still in the cluster
472 file system, you may want to clean those up too. After making absolutely sure
473 that you have the correct node name, you can simply remove the entire
474 directory recursively from '/etc/pve/nodes/NODENAME'.
476 CAUTION: The node's SSH keys will remain in the 'authorized_key' file. This
477 means that the nodes can still connect to each other with public key
478 authentication. You should fix this by removing the respective keys from the
479 '/etc/pve/priv/authorized_keys' file.
485 {pve} use a quorum-based technique to provide a consistent state among
488 [quote, from Wikipedia, Quorum (distributed computing)]
490 A quorum is the minimum number of votes that a distributed transaction
491 has to obtain in order to be allowed to perform an operation in a
495 In case of network partitioning, state changes requires that a
496 majority of nodes are online. The cluster switches to read-only mode
499 NOTE: {pve} assigns a single vote to each node by default.
505 The cluster network is the core of a cluster. All messages sent over it have to
506 be delivered reliably to all nodes in their respective order. In {pve} this
507 part is done by corosync, an implementation of a high performance, low overhead,
508 high availability development toolkit. It serves our decentralized configuration
509 file system (`pmxcfs`).
511 [[pvecm_cluster_network_requirements]]
514 This needs a reliable network with latencies under 2 milliseconds (LAN
515 performance) to work properly. The network should not be used heavily by other
516 members; ideally corosync runs on its own network. Do not use a shared network
517 for corosync and storage (except as a potential low-priority fallback in a
518 xref:pvecm_redundancy[redundant] configuration).
520 Before setting up a cluster, it is good practice to check if the network is fit
521 for that purpose. To ensure that the nodes can connect to each other on the
522 cluster network, you can test the connectivity between them with the `ping`
525 If the {pve} firewall is enabled, ACCEPT rules for corosync will automatically
526 be generated - no manual action is required.
528 NOTE: Corosync used Multicast before version 3.0 (introduced in {pve} 6.0).
529 Modern versions rely on https://kronosnet.org/[Kronosnet] for cluster
530 communication, which, for now, only supports regular UDP unicast.
532 CAUTION: You can still enable Multicast or legacy unicast by setting your
533 transport to `udp` or `udpu` in your xref:pvecm_edit_corosync_conf[corosync.conf],
534 but keep in mind that this will disable all cryptography and redundancy support.
535 This is therefore not recommended.
537 Separate Cluster Network
538 ~~~~~~~~~~~~~~~~~~~~~~~~
540 When creating a cluster without any parameters, the corosync cluster network is
541 generally shared with the web interface and the VMs' network. Depending on
542 your setup, even storage traffic may get sent over the same network. It's
543 recommended to change that, as corosync is a time-critical, real-time
546 Setting Up a New Network
547 ^^^^^^^^^^^^^^^^^^^^^^^^
549 First, you have to set up a new network interface. It should be on a physically
550 separate network. Ensure that your network fulfills the
551 xref:pvecm_cluster_network_requirements[cluster network requirements].
553 Separate On Cluster Creation
554 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
556 This is possible via the 'linkX' parameters of the 'pvecm create'
557 command, used for creating a new cluster.
559 If you have set up an additional NIC with a static address on 10.10.10.1/25,
560 and want to send and receive all cluster communication over this interface,
565 pvecm create test --link0 10.10.10.1
568 To check if everything is working properly, execute:
571 systemctl status corosync
574 Afterwards, proceed as described above to
575 xref:pvecm_adding_nodes_with_separated_cluster_network[add nodes with a separated cluster network].
577 [[pvecm_separate_cluster_net_after_creation]]
578 Separate After Cluster Creation
579 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
581 You can do this if you have already created a cluster and want to switch
582 its communication to another network, without rebuilding the whole cluster.
583 This change may lead to short periods of quorum loss in the cluster, as nodes
584 have to restart corosync and come up one after the other on the new network.
586 Check how to xref:pvecm_edit_corosync_conf[edit the corosync.conf file] first.
587 Then, open it and you should see a file similar to:
621 provider: corosync_votequorum
625 cluster_name: testcluster
637 NOTE: `ringX_addr` actually specifies a corosync *link address*. The name "ring"
638 is a remnant of older corosync versions that is kept for backwards
641 The first thing you want to do is add the 'name' properties in the node entries,
642 if you do not see them already. Those *must* match the node name.
644 Then replace all addresses from the 'ring0_addr' properties of all nodes with
645 the new addresses. You may use plain IP addresses or hostnames here. If you use
646 hostnames, ensure that they are resolvable from all nodes (see also
647 xref:pvecm_corosync_addresses[Link Address Types]).
649 In this example, we want to switch cluster communication to the
650 10.10.10.1/25 network, so we change the 'ring0_addr' of each node respectively.
652 NOTE: The exact same procedure can be used to change other 'ringX_addr' values
653 as well. However, we recommend only changing one link address at a time, so
654 that it's easier to recover if something goes wrong.
656 After we increase the 'config_version' property, the new configuration file
671 ring0_addr: 10.10.10.2
678 ring0_addr: 10.10.10.3
685 ring0_addr: 10.10.10.1
691 provider: corosync_votequorum
695 cluster_name: testcluster
707 Then, after a final check to see that all changed information is correct, we
708 save it and once again follow the
709 xref:pvecm_edit_corosync_conf[edit corosync.conf file] section to bring it into
712 The changes will be applied live, so restarting corosync is not strictly
713 necessary. If you changed other settings as well, or notice corosync
714 complaining, you can optionally trigger a restart.
716 On a single node execute:
720 systemctl restart corosync
723 Now check if everything is okay:
727 systemctl status corosync
730 If corosync begins to work again, restart it on all other nodes too.
731 They will then join the cluster membership one by one on the new network.
733 [[pvecm_corosync_addresses]]
737 A corosync link address (for backwards compatibility denoted by 'ringX_addr' in
738 `corosync.conf`) can be specified in two ways:
740 * **IPv4/v6 addresses** can be used directly. They are recommended, since they
741 are static and usually not changed carelessly.
743 * **Hostnames** will be resolved using `getaddrinfo`, which means that by
744 default, IPv6 addresses will be used first, if available (see also
745 `man gai.conf`). Keep this in mind, especially when upgrading an existing
748 CAUTION: Hostnames should be used with care, since the addresses they
749 resolve to can be changed without touching corosync or the node it runs on -
750 which may lead to a situation where an address is changed without thinking
751 about implications for corosync.
753 A separate, static hostname specifically for corosync is recommended, if
754 hostnames are preferred. Also, make sure that every node in the cluster can
755 resolve all hostnames correctly.
757 Since {pve} 5.1, while supported, hostnames will be resolved at the time of
758 entry. Only the resolved IP is saved to the configuration.
760 Nodes that joined the cluster on earlier versions likely still use their
761 unresolved hostname in `corosync.conf`. It might be a good idea to replace
762 them with IPs or a separate hostname, as mentioned above.
769 Corosync supports redundant networking via its integrated Kronosnet layer by
770 default (it is not supported on the legacy udp/udpu transports). It can be
771 enabled by specifying more than one link address, either via the '--linkX'
772 parameters of `pvecm`, in the GUI as **Link 1** (while creating a cluster or
773 adding a new node) or by specifying more than one 'ringX_addr' in
776 NOTE: To provide useful failover, every link should be on its own
777 physical network connection.
779 Links are used according to a priority setting. You can configure this priority
780 by setting 'knet_link_priority' in the corresponding interface section in
781 `corosync.conf`, or, preferably, using the 'priority' parameter when creating
782 your cluster with `pvecm`:
785 # pvecm create CLUSTERNAME --link0 10.10.10.1,priority=15 --link1 10.20.20.1,priority=20
788 This would cause 'link1' to be used first, since it has the higher priority.
790 If no priorities are configured manually (or two links have the same priority),
791 links will be used in order of their number, with the lower number having higher
794 Even if all links are working, only the one with the highest priority will see
795 corosync traffic. Link priorities cannot be mixed, meaning that links with
796 different priorities will not be able to communicate with each other.
798 Since lower priority links will not see traffic unless all higher priorities
799 have failed, it becomes a useful strategy to specify networks used for
800 other tasks (VMs, storage, etc.) as low-priority links. If worst comes to
801 worst, a higher latency or more congested connection might be better than no
804 Adding Redundant Links To An Existing Cluster
805 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
807 To add a new link to a running configuration, first check how to
808 xref:pvecm_edit_corosync_conf[edit the corosync.conf file].
810 Then, add a new 'ringX_addr' to every node in the `nodelist` section. Make
811 sure that your 'X' is the same for every node you add it to, and that it is
812 unique for each node.
814 Lastly, add a new 'interface', as shown below, to your `totem`
815 section, replacing 'X' with the link number chosen above.
817 Assuming you added a link with number 1, the new configuration file could look
832 ring0_addr: 10.10.10.2
833 ring1_addr: 10.20.20.2
840 ring0_addr: 10.10.10.3
841 ring1_addr: 10.20.20.3
848 ring0_addr: 10.10.10.1
849 ring1_addr: 10.20.20.1
855 provider: corosync_votequorum
859 cluster_name: testcluster
873 The new link will be enabled as soon as you follow the last steps to
874 xref:pvecm_edit_corosync_conf[edit the corosync.conf file]. A restart should not
875 be necessary. You can check that corosync loaded the new link using:
878 journalctl -b -u corosync
881 It might be a good idea to test the new link by temporarily disconnecting the
882 old link on one node and making sure that its status remains online while
889 If you see a healthy cluster state, it means that your new link is being used.
892 Role of SSH in {pve} Clusters
893 -----------------------------
895 {pve} utilizes SSH tunnels for various features.
897 * Proxying console/shell sessions (node and guests)
899 When using the shell for node B while being connected to node A, connects to a
900 terminal proxy on node A, which is in turn connected to the login shell on node
901 B via a non-interactive SSH tunnel.
903 * VM and CT memory and local-storage migration in 'secure' mode.
905 During the migration, one or more SSH tunnel(s) are established between the
906 source and target nodes, in order to exchange migration information and
907 transfer memory and disk contents.
909 * Storage replication
911 .Pitfalls due to automatic execution of `.bashrc` and siblings
914 In case you have a custom `.bashrc`, or similar files that get executed on
915 login by the configured shell, `ssh` will automatically run it once the session
916 is established successfully. This can cause some unexpected behavior, as those
917 commands may be executed with root permissions on any of the operations
918 described above. This can cause possible problematic side-effects!
920 In order to avoid such complications, it's recommended to add a check in
921 `/root/.bashrc` to make sure the session is interactive, and only then run
924 You can add this snippet at the beginning of your `.bashrc` file:
927 # Early exit if not running interactively to avoid side-effects!
936 Corosync External Vote Support
937 ------------------------------
939 This section describes a way to deploy an external voter in a {pve} cluster.
940 When configured, the cluster can sustain more node failures without
941 violating safety properties of the cluster communication.
943 For this to work, there are two services involved:
945 * A QDevice daemon which runs on each {pve} node
947 * An external vote daemon which runs on an independent server
949 As a result, you can achieve higher availability, even in smaller setups (for
952 QDevice Technical Overview
953 ~~~~~~~~~~~~~~~~~~~~~~~~~~
955 The Corosync Quorum Device (QDevice) is a daemon which runs on each cluster
956 node. It provides a configured number of votes to the cluster's quorum
957 subsystem, based on an externally running third-party arbitrator's decision.
958 Its primary use is to allow a cluster to sustain more node failures than
959 standard quorum rules allow. This can be done safely as the external device
960 can see all nodes and thus choose only one set of nodes to give its vote.
961 This will only be done if said set of nodes can have quorum (again) after
962 receiving the third-party vote.
964 Currently, only 'QDevice Net' is supported as a third-party arbitrator. This is
965 a daemon which provides a vote to a cluster partition, if it can reach the
966 partition members over the network. It will only give votes to one partition
967 of a cluster at any time.
968 It's designed to support multiple clusters and is almost configuration and
969 state free. New clusters are handled dynamically and no configuration file
970 is needed on the host running a QDevice.
972 The only requirements for the external host are that it needs network access to
973 the cluster and to have a corosync-qnetd package available. We provide a package
974 for Debian based hosts, and other Linux distributions should also have a package
975 available through their respective package manager.
977 NOTE: In contrast to corosync itself, a QDevice connects to the cluster over
978 TCP/IP. The daemon may even run outside of the cluster's LAN and can have longer
984 We support QDevices for clusters with an even number of nodes and recommend
985 it for 2 node clusters, if they should provide higher availability.
986 For clusters with an odd node count, we currently discourage the use of
987 QDevices. The reason for this is the difference in the votes which the QDevice
988 provides for each cluster type. Even numbered clusters get a single additional
989 vote, which only increases availability, because if the QDevice
990 itself fails, you are in the same position as with no QDevice at all.
992 On the other hand, with an odd numbered cluster size, the QDevice provides
993 '(N-1)' votes -- where 'N' corresponds to the cluster node count. This
994 alternative behavior makes sense; if it had only one additional vote, the
995 cluster could get into a split-brain situation. This algorithm allows for all
996 nodes but one (and naturally the QDevice itself) to fail. However, there are two
999 * If the QNet daemon itself fails, no other node may fail or the cluster
1000 immediately loses quorum. For example, in a cluster with 15 nodes, 7
1001 could fail before the cluster becomes inquorate. But, if a QDevice is
1002 configured here and it itself fails, **no single node** of the 15 may fail.
1003 The QDevice acts almost as a single point of failure in this case.
1005 * The fact that all but one node plus QDevice may fail sounds promising at
1006 first, but this may result in a mass recovery of HA services, which could
1007 overload the single remaining node. Furthermore, a Ceph server will stop
1008 providing services if only '((N-1)/2)' nodes or less remain online.
1010 If you understand the drawbacks and implications, you can decide yourself if
1011 you want to use this technology in an odd numbered cluster setup.
1016 We recommend running any daemon which provides votes to corosync-qdevice as an
1017 unprivileged user. {pve} and Debian provide a package which is already
1018 configured to do so.
1019 The traffic between the daemon and the cluster must be encrypted to ensure a
1020 safe and secure integration of the QDevice in {pve}.
1022 First, install the 'corosync-qnetd' package on your external server
1025 external# apt install corosync-qnetd
1028 and the 'corosync-qdevice' package on all cluster nodes
1031 pve# apt install corosync-qdevice
1034 After doing this, ensure that all the nodes in the cluster are online.
1036 You can now set up your QDevice by running the following command on one
1040 pve# pvecm qdevice setup <QDEVICE-IP>
1043 The SSH key from the cluster will be automatically copied to the QDevice.
1045 NOTE: Make sure that the SSH configuration on your external server allows root
1046 login via password, if you are asked for a password during this step.
1048 After you enter the password and all the steps have successfully completed, you
1049 will see "Done". You can verify that the QDevice has been set up with:
1056 Votequorum information
1057 ~~~~~~~~~~~~~~~~~~~~~
1062 Flags: Quorate Qdevice
1064 Membership information
1065 ~~~~~~~~~~~~~~~~~~~~~~
1066 Nodeid Votes Qdevice Name
1067 0x00000001 1 A,V,NMW 192.168.22.180 (local)
1068 0x00000002 1 A,V,NMW 192.168.22.181
1069 0x00000000 1 Qdevice
1074 Frequently Asked Questions
1075 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1080 In case of a tie, where two same-sized cluster partitions cannot see each other
1081 but can see the QDevice, the QDevice chooses one of those partitions randomly
1082 and provides a vote to it.
1084 Possible Negative Implications
1085 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1087 For clusters with an even node count, there are no negative implications when
1088 using a QDevice. If it fails to work, it is the same as not having a QDevice
1091 Adding/Deleting Nodes After QDevice Setup
1092 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1094 If you want to add a new node or remove an existing one from a cluster with a
1095 QDevice setup, you need to remove the QDevice first. After that, you can add or
1096 remove nodes normally. Once you have a cluster with an even node count again,
1097 you can set up the QDevice again as described previously.
1099 Removing the QDevice
1100 ^^^^^^^^^^^^^^^^^^^^
1102 If you used the official `pvecm` tool to add the QDevice, you can remove it
1106 pve# pvecm qdevice remove
1111 //There is still stuff to add here
1114 Corosync Configuration
1115 ----------------------
1117 The `/etc/pve/corosync.conf` file plays a central role in a {pve} cluster. It
1118 controls the cluster membership and its network.
1119 For further information about it, check the corosync.conf man page:
1125 For node membership, you should always use the `pvecm` tool provided by {pve}.
1126 You may have to edit the configuration file manually for other changes.
1127 Here are a few best practice tips for doing this.
1129 [[pvecm_edit_corosync_conf]]
1133 Editing the corosync.conf file is not always very straightforward. There are
1134 two on each cluster node, one in `/etc/pve/corosync.conf` and the other in
1135 `/etc/corosync/corosync.conf`. Editing the one in our cluster file system will
1136 propagate the changes to the local one, but not vice versa.
1138 The configuration will get updated automatically, as soon as the file changes.
1139 This means that changes which can be integrated in a running corosync will take
1140 effect immediately. Thus, you should always make a copy and edit that instead,
1141 to avoid triggering unintended changes when saving the file while editing.
1145 cp /etc/pve/corosync.conf /etc/pve/corosync.conf.new
1148 Then, open the config file with your favorite editor, such as `nano` or
1149 `vim.tiny`, which come pre-installed on every {pve} node.
1151 NOTE: Always increment the 'config_version' number after configuration changes;
1152 omitting this can lead to problems.
1154 After making the necessary changes, create another copy of the current working
1155 configuration file. This serves as a backup if the new configuration fails to
1156 apply or causes other issues.
1160 cp /etc/pve/corosync.conf /etc/pve/corosync.conf.bak
1163 Then replace the old configuration file with the new one:
1166 mv /etc/pve/corosync.conf.new /etc/pve/corosync.conf
1169 You can check if the changes could be applied automatically, using the following
1173 systemctl status corosync
1174 journalctl -b -u corosync
1177 If the changes could not be applied automatically, you may have to restart the
1178 corosync service via:
1181 systemctl restart corosync
1184 On errors, check the troubleshooting section below.
1189 Issue: 'quorum.expected_votes must be configured'
1190 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1192 When corosync starts to fail and you get the following message in the system log:
1196 corosync[1647]: [QUORUM] Quorum provider: corosync_votequorum failed to initialize.
1197 corosync[1647]: [SERV ] Service engine 'corosync_quorum' failed to load for reason
1198 'configuration error: nodelist or quorum.expected_votes must be configured!'
1202 It means that the hostname you set for a corosync 'ringX_addr' in the
1203 configuration could not be resolved.
1205 Write Configuration When Not Quorate
1206 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1208 If you need to change '/etc/pve/corosync.conf' on a node with no quorum, and you
1209 understand what you are doing, use:
1215 This sets the expected vote count to 1 and makes the cluster quorate. You can
1216 then fix your configuration, or revert it back to the last working backup.
1218 This is not enough if corosync cannot start anymore. In that case, it is best to
1219 edit the local copy of the corosync configuration in
1220 '/etc/corosync/corosync.conf', so that corosync can start again. Ensure that on
1221 all nodes, this configuration has the same content to avoid split-brain
1225 [[pvecm_corosync_conf_glossary]]
1226 Corosync Configuration Glossary
1227 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1230 This names the different link addresses for the Kronosnet connections between
1237 It is obvious that a cluster is not quorate when all nodes are
1238 offline. This is a common case after a power failure.
1240 NOTE: It is always a good idea to use an uninterruptible power supply
1241 (``UPS'', also called ``battery backup'') to avoid this state, especially if
1244 On node startup, the `pve-guests` service is started and waits for
1245 quorum. Once quorate, it starts all guests which have the `onboot`
1248 When you turn on nodes, or when power comes back after power failure,
1249 it is likely that some nodes will boot faster than others. Please keep in
1250 mind that guest startup is delayed until you reach quorum.
1256 Migrating virtual guests to other nodes is a useful feature in a
1257 cluster. There are settings to control the behavior of such
1258 migrations. This can be done via the configuration file
1259 `datacenter.cfg` or for a specific migration via API or command line
1262 It makes a difference if a guest is online or offline, or if it has
1263 local resources (like a local disk).
1265 For details about virtual machine migration, see the
1266 xref:qm_migration[QEMU/KVM Migration Chapter].
1268 For details about container migration, see the
1269 xref:pct_migration[Container Migration Chapter].
1274 The migration type defines if the migration data should be sent over an
1275 encrypted (`secure`) channel or an unencrypted (`insecure`) one.
1276 Setting the migration type to insecure means that the RAM content of a
1277 virtual guest is also transferred unencrypted, which can lead to
1278 information disclosure of critical data from inside the guest (for
1279 example, passwords or encryption keys).
1281 Therefore, we strongly recommend using the secure channel if you do
1282 not have full control over the network and can not guarantee that no
1283 one is eavesdropping on it.
1285 NOTE: Storage migration does not follow this setting. Currently, it
1286 always sends the storage content over a secure channel.
1288 Encryption requires a lot of computing power, so this setting is often
1289 changed to "unsafe" to achieve better performance. The impact on
1290 modern systems is lower because they implement AES encryption in
1291 hardware. The performance impact is particularly evident in fast
1292 networks, where you can transfer 10 Gbps or more.
1297 By default, {pve} uses the network in which cluster communication
1298 takes place to send the migration traffic. This is not optimal both because
1299 sensitive cluster traffic can be disrupted and this network may not
1300 have the best bandwidth available on the node.
1302 Setting the migration network parameter allows the use of a dedicated
1303 network for all migration traffic. In addition to the memory,
1304 this also affects the storage traffic for offline migrations.
1306 The migration network is set as a network using CIDR notation. This
1307 has the advantage that you don't have to set individual IP addresses
1308 for each node. {pve} can determine the real address on the
1309 destination node from the network specified in the CIDR form. To
1310 enable this, the network must be specified so that each node has exactly one
1311 IP in the respective network.
1316 We assume that we have a three-node setup, with three separate
1317 networks. One for public communication with the Internet, one for
1318 cluster communication, and a very fast one, which we want to use as a
1319 dedicated network for migration.
1321 A network configuration for such a setup might look as follows:
1324 iface eno1 inet manual
1328 iface vmbr0 inet static
1329 address 192.X.Y.57/24
1337 iface eno2 inet static
1342 iface eno3 inet static
1346 Here, we will use the network 10.1.2.0/24 as a migration network. For
1347 a single migration, you can do this using the `migration_network`
1348 parameter of the command line tool:
1351 # qm migrate 106 tre --online --migration_network 10.1.2.0/24
1354 To configure this as the default network for all migrations in the
1355 cluster, set the `migration` property of the `/etc/pve/datacenter.cfg`
1359 # use dedicated migration network
1360 migration: secure,network=10.1.2.0/24
1363 NOTE: The migration type must always be set when the migration network
1364 is set in `/etc/pve/datacenter.cfg`.
1368 include::pve-copyright.adoc[]