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, and such clusters can consist of up to 32 physical nodes
31 (probably more, dependent on network latency).
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
39 Grouping nodes into a cluster has the following advantages:
41 * Centralized, web based management
43 * Multi-master clusters: each node can do all management task
45 * `pmxcfs`: database-driven file system for storing configuration files,
46 replicated in real-time on all nodes using `corosync`.
48 * Easy migration of virtual machines and containers between physical
53 * Cluster-wide services like firewall and HA
59 * All nodes must be able to connect to each other via UDP ports 5404 and 5405
62 * Date and time have to be synchronized.
64 * SSH tunnel on TCP port 22 between nodes is used.
66 * If you are interested in High Availability, you need to have at
67 least three nodes for reliable quorum. All nodes should have the
70 * We recommend a dedicated NIC for the cluster traffic, especially if
71 you use shared storage.
73 * Root password of a cluster node is required for adding nodes.
75 NOTE: It is not possible to mix {pve} 3.x and earlier with {pve} 4.X cluster
78 NOTE: While it's possible for {pve} 4.4 and {pve} 5.0 this is not supported as
79 production configuration and should only used temporarily during upgrading the
80 whole cluster from one to another major version.
82 NOTE: Running a cluster of {pve} 6.x with earlier versions is not possible. The
83 cluster protocol (corosync) between {pve} 6.x and earlier versions changed
84 fundamentally. The corosync 3 packages for {pve} 5.4 are only intended for the
85 upgrade procedure to {pve} 6.0.
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.
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 ->
99 While it's common to reference all nodenames and their IPs in `/etc/hosts` (or
100 make their names resolvable through other means), this is not necessary for a
101 cluster to work. It may be useful however, as you can then connect from one node
102 to the other with SSH via the easier to remember node name (see also
103 xref:pvecm_corosync_addresses[Link Address Types]). Note that we always
104 recommend to reference nodes by their IP addresses in the cluster configuration.
107 [[pvecm_create_cluster]]
111 Login via `ssh` to the first {pve} node. Use a unique name for your cluster.
112 This name cannot be changed later. The cluster name follows the same rules as
116 hp1# pvecm create CLUSTERNAME
119 NOTE: It is possible to create multiple clusters in the same physical or logical
120 network. Use unique cluster names if you do so. To avoid human confusion, it is
121 also recommended to choose different names even if clusters do not share the
124 To check the state of your cluster use:
131 [[pvecm_join_node_to_cluster]]
132 Adding Nodes to the Cluster
133 ---------------------------
135 Login via `ssh` to the node you want to add.
138 hp2# pvecm add IP-ADDRESS-CLUSTER
141 For `IP-ADDRESS-CLUSTER` use the IP or hostname of an existing cluster node.
142 An IP address is recommended (see xref:pvecm_corosync_addresses[Link Address Types]).
144 CAUTION: A new node cannot hold any VMs, because you would get
145 conflicts about identical VM IDs. Also, all existing configuration in
146 `/etc/pve` is overwritten when you join a new node to the cluster. To
147 workaround, use `vzdump` to backup and restore to a different VMID after
148 adding the node to the cluster.
150 To check the state of the cluster use:
156 .Cluster status after adding 4 nodes
161 Date: Mon Apr 20 12:30:13 2015
162 Quorum provider: corosync_votequorum
168 Votequorum information
169 ~~~~~~~~~~~~~~~~~~~~~~
176 Membership information
177 ~~~~~~~~~~~~~~~~~~~~~~
179 0x00000001 1 192.168.15.91
180 0x00000002 1 192.168.15.92 (local)
181 0x00000003 1 192.168.15.93
182 0x00000004 1 192.168.15.94
185 If you only want the list of all nodes use:
191 .List nodes in a cluster
195 Membership information
196 ~~~~~~~~~~~~~~~~~~~~~~
204 [[pvecm_adding_nodes_with_separated_cluster_network]]
205 Adding Nodes With Separated Cluster Network
206 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
208 When adding a node to a cluster with a separated cluster network you need to
209 use the 'link0' parameter to set the nodes address on that network:
213 pvecm add IP-ADDRESS-CLUSTER -link0 LOCAL-IP-ADDRESS-LINK0
216 If you want to use the built-in xref:pvecm_redundancy[redundancy] of the
217 kronosnet transport layer, also use the 'link1' parameter.
220 Remove a Cluster Node
221 ---------------------
223 CAUTION: Read carefully the procedure before proceeding, as it could
224 not be what you want or need.
226 Move all virtual machines from the node. Make sure you have no local
227 data or backups you want to keep, or save them accordingly.
228 In the following example we will remove the node hp4 from the cluster.
230 Log in to a *different* cluster node (not hp4), and issue a `pvecm nodes`
231 command to identify the node ID to remove:
236 Membership information
237 ~~~~~~~~~~~~~~~~~~~~~~
246 At this point you must power off hp4 and
247 make sure that it will not power on again (in the network) as it
250 IMPORTANT: As said above, it is critical to power off the node
251 *before* removal, and make sure that it will *never* power on again
252 (in the existing cluster network) as it is.
253 If you power on the node as it is, your cluster will be screwed up and
254 it could be difficult to restore a clean cluster state.
256 After powering off the node hp4, we can safely remove it from the cluster.
259 hp1# pvecm delnode hp4
262 If the operation succeeds no output is returned, just check the node
263 list again with `pvecm nodes` or `pvecm status`. You should see
271 Date: Mon Apr 20 12:44:28 2015
272 Quorum provider: corosync_votequorum
278 Votequorum information
279 ~~~~~~~~~~~~~~~~~~~~~~
286 Membership information
287 ~~~~~~~~~~~~~~~~~~~~~~
289 0x00000001 1 192.168.15.90 (local)
290 0x00000002 1 192.168.15.91
291 0x00000003 1 192.168.15.92
294 If, for whatever reason, you want this server to join the same cluster again,
297 * reinstall {pve} on it from scratch
299 * then join it, as explained in the previous section.
301 NOTE: After removal of the node, its SSH fingerprint will still reside in the
302 'known_hosts' of the other nodes. If you receive an SSH error after rejoining
303 a node with the same IP or hostname, run `pvecm updatecerts` once on the
304 re-added node to update its fingerprint cluster wide.
306 [[pvecm_separate_node_without_reinstall]]
307 Separate A Node Without Reinstalling
308 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
310 CAUTION: This is *not* the recommended method, proceed with caution. Use the
311 above mentioned method if you're unsure.
313 You can also separate a node from a cluster without reinstalling it from
314 scratch. But after removing the node from the cluster it will still have
315 access to the shared storages! This must be resolved before you start removing
316 the node from the cluster. A {pve} cluster cannot share the exact same
317 storage with another cluster, as storage locking doesn't work over cluster
318 boundary. Further, it may also lead to VMID conflicts.
320 Its suggested that you create a new storage where only the node which you want
321 to separate has access. This can be a new export on your NFS or a new Ceph
322 pool, to name a few examples. Its just important that the exact same storage
323 does not gets accessed by multiple clusters. After setting this storage up move
324 all data from the node and its VMs to it. Then you are ready to separate the
325 node from the cluster.
327 WARNING: Ensure all shared resources are cleanly separated! Otherwise you will
328 run into conflicts and problems.
330 First stop the corosync and the pve-cluster services on the node:
333 systemctl stop pve-cluster
334 systemctl stop corosync
337 Start the cluster filesystem again in local mode:
343 Delete the corosync configuration files:
346 rm /etc/pve/corosync.conf
350 You can now start the filesystem again as normal service:
354 systemctl start pve-cluster
357 The node is now separated from the cluster. You can deleted it from a remaining
358 node of the cluster with:
361 pvecm delnode oldnode
364 If the command failed, because the remaining node in the cluster lost quorum
365 when the now separate node exited, you may set the expected votes to 1 as a workaround:
371 And then repeat the 'pvecm delnode' command.
373 Now switch back to the separated node, here delete all remaining files left
374 from the old cluster. This ensures that the node can be added to another
375 cluster again without problems.
379 rm /var/lib/corosync/*
382 As the configuration files from the other nodes are still in the cluster
383 filesystem you may want to clean those up too. Remove simply the whole
384 directory recursive from '/etc/pve/nodes/NODENAME', but check three times that
385 you used the correct one before deleting it.
387 CAUTION: The nodes SSH keys are still in the 'authorized_key' file, this means
388 the nodes can still connect to each other with public key authentication. This
389 should be fixed by removing the respective keys from the
390 '/etc/pve/priv/authorized_keys' file.
396 {pve} use a quorum-based technique to provide a consistent state among
399 [quote, from Wikipedia, Quorum (distributed computing)]
401 A quorum is the minimum number of votes that a distributed transaction
402 has to obtain in order to be allowed to perform an operation in a
406 In case of network partitioning, state changes requires that a
407 majority of nodes are online. The cluster switches to read-only mode
410 NOTE: {pve} assigns a single vote to each node by default.
416 The cluster network is the core of a cluster. All messages sent over it have to
417 be delivered reliably to all nodes in their respective order. In {pve} this
418 part is done by corosync, an implementation of a high performance, low overhead
419 high availability development toolkit. It serves our decentralized
420 configuration file system (`pmxcfs`).
422 [[pvecm_cluster_network_requirements]]
425 This needs a reliable network with latencies under 2 milliseconds (LAN
426 performance) to work properly. The network should not be used heavily by other
427 members, ideally corosync runs on its own network. Do not use a shared network
428 for corosync and storage (except as a potential low-priority fallback in a
429 xref:pvecm_redundancy[redundant] configuration).
431 Before setting up a cluster, it is good practice to check if the network is fit
432 for that purpose. To make sure the nodes can connect to each other on the
433 cluster network, you can test the connectivity between them with the `ping`
436 If the {pve} firewall is enabled, ACCEPT rules for corosync will automatically
437 be generated - no manual action is required.
439 NOTE: Corosync used Multicast before version 3.0 (introduced in {pve} 6.0).
440 Modern versions rely on https://kronosnet.org/[Kronosnet] for cluster
441 communication, which, for now, only supports regular UDP unicast.
443 CAUTION: You can still enable Multicast or legacy unicast by setting your
444 transport to `udp` or `udpu` in your xref:pvecm_edit_corosync_conf[corosync.conf],
445 but keep in mind that this will disable all cryptography and redundancy support.
446 This is therefore not recommended.
448 Separate Cluster Network
449 ~~~~~~~~~~~~~~~~~~~~~~~~
451 When creating a cluster without any parameters the corosync cluster network is
452 generally shared with the Web UI and the VMs and their traffic. Depending on
453 your setup, even storage traffic may get sent over the same network. Its
454 recommended to change that, as corosync is a time critical real time
457 Setting Up A New Network
458 ^^^^^^^^^^^^^^^^^^^^^^^^
460 First you have to set up a new network interface. It should be on a physically
461 separate network. Ensure that your network fulfills the
462 xref:pvecm_cluster_network_requirements[cluster network requirements].
464 Separate On Cluster Creation
465 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
467 This is possible via the 'linkX' parameters of the 'pvecm create'
468 command used for creating a new cluster.
470 If you have set up an additional NIC with a static address on 10.10.10.1/25,
471 and want to send and receive all cluster communication over this interface,
476 pvecm create test --link0 10.10.10.1
479 To check if everything is working properly execute:
482 systemctl status corosync
485 Afterwards, proceed as described above to
486 xref:pvecm_adding_nodes_with_separated_cluster_network[add nodes with a separated cluster network].
488 [[pvecm_separate_cluster_net_after_creation]]
489 Separate After Cluster Creation
490 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
492 You can do this if you have already created a cluster and want to switch
493 its communication to another network, without rebuilding the whole cluster.
494 This change may lead to short durations of quorum loss in the cluster, as nodes
495 have to restart corosync and come up one after the other on the new network.
497 Check how to xref:pvecm_edit_corosync_conf[edit the corosync.conf file] first.
498 Then, open it and you should see a file similar to:
532 provider: corosync_votequorum
536 cluster_name: testcluster
548 NOTE: `ringX_addr` actually specifies a corosync *link address*, the name "ring"
549 is a remnant of older corosync versions that is kept for backwards
552 The first thing you want to do is add the 'name' properties in the node entries
553 if you do not see them already. Those *must* match the node name.
555 Then replace all addresses from the 'ring0_addr' properties of all nodes with
556 the new addresses. You may use plain IP addresses or hostnames here. If you use
557 hostnames ensure that they are resolvable from all nodes. (see also
558 xref:pvecm_corosync_addresses[Link Address Types])
560 In this example, we want to switch the cluster communication to the
561 10.10.10.1/25 network. So we replace all 'ring0_addr' respectively.
563 NOTE: The exact same procedure can be used to change other 'ringX_addr' values
564 as well, although we recommend to not change multiple addresses at once, to make
565 it easier to recover if something goes wrong.
567 After we increase the 'config_version' property, the new configuration file
582 ring0_addr: 10.10.10.2
589 ring0_addr: 10.10.10.3
596 ring0_addr: 10.10.10.1
602 provider: corosync_votequorum
606 cluster_name: testcluster
618 Then, after a final check if all changed information is correct, we save it and
619 once again follow the xref:pvecm_edit_corosync_conf[edit corosync.conf file]
620 section to bring it into effect.
622 The changes will be applied live, so restarting corosync is not strictly
623 necessary. If you changed other settings as well, or notice corosync
624 complaining, you can optionally trigger a restart.
626 On a single node execute:
630 systemctl restart corosync
633 Now check if everything is fine:
637 systemctl status corosync
640 If corosync runs again correct restart corosync also on all other nodes.
641 They will then join the cluster membership one by one on the new network.
643 [[pvecm_corosync_addresses]]
647 A corosync link address (for backwards compatibility denoted by 'ringX_addr' in
648 `corosync.conf`) can be specified in two ways:
650 * **IPv4/v6 addresses** will be used directly. They are recommended, since they
651 are static and usually not changed carelessly.
653 * **Hostnames** will be resolved using `getaddrinfo`, which means that per
654 default, IPv6 addresses will be used first, if available (see also
655 `man gai.conf`). Keep this in mind, especially when upgrading an existing
658 CAUTION: Hostnames should be used with care, since the address they
659 resolve to can be changed without touching corosync or the node it runs on -
660 which may lead to a situation where an address is changed without thinking
661 about implications for corosync.
663 A seperate, static hostname specifically for corosync is recommended, if
664 hostnames are preferred. Also, make sure that every node in the cluster can
665 resolve all hostnames correctly.
667 Since {pve} 5.1, while supported, hostnames will be resolved at the time of
668 entry. Only the resolved IP is then saved to the configuration.
670 Nodes that joined the cluster on earlier versions likely still use their
671 unresolved hostname in `corosync.conf`. It might be a good idea to replace
672 them with IPs or a seperate hostname, as mentioned above.
679 Corosync supports redundant networking via its integrated kronosnet layer by
680 default (it is not supported on the legacy udp/udpu transports). It can be
681 enabled by specifying more than one link address, either via the '--linkX'
682 parameters of `pvecm` (while creating a cluster or adding a new node) or by
683 specifying more than one 'ringX_addr' in `corosync.conf`.
685 NOTE: To provide useful failover, every link should be on its own
686 physical network connection.
688 Links are used according to a priority setting. You can configure this priority
689 by setting 'knet_link_priority' in the corresponding interface section in
690 `corosync.conf`, or, preferrably, using the 'priority' parameter when creating
691 your cluster with `pvecm`:
694 # pvecm create CLUSTERNAME --link0 10.10.10.1,priority=20 --link1 10.20.20.1,priority=15
697 This would cause 'link1' to be used first, since it has the lower priority.
699 If no priorities are configured manually (or two links have the same priority),
700 links will be used in order of their number, with the lower number having higher
703 Even if all links are working, only the one with the highest priority will see
704 corosync traffic. Link priorities cannot be mixed, i.e. links with different
705 priorities will not be able to communicate with each other.
707 Since lower priority links will not see traffic unless all higher priorities
708 have failed, it becomes a useful strategy to specify even networks used for
709 other tasks (VMs, storage, etc...) as low-priority links. If worst comes to
710 worst, a higher-latency or more congested connection might be better than no
713 Adding Redundant Links To An Existing Cluster
714 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
716 To add a new link to a running configuration, first check how to
717 xref:pvecm_edit_corosync_conf[edit the corosync.conf file].
719 Then, add a new 'ringX_addr' to every node in the `nodelist` section. Make
720 sure that your 'X' is the same for every node you add it to, and that it is
721 unique for each node.
723 Lastly, add a new 'interface', as shown below, to your `totem`
724 section, replacing 'X' with your link number chosen above.
726 Assuming you added a link with number 1, the new configuration file could look
741 ring0_addr: 10.10.10.2
742 ring1_addr: 10.20.20.2
749 ring0_addr: 10.10.10.3
750 ring1_addr: 10.20.20.3
757 ring0_addr: 10.10.10.1
758 ring1_addr: 10.20.20.1
764 provider: corosync_votequorum
768 cluster_name: testcluster
782 The new link will be enabled as soon as you follow the last steps to
783 xref:pvecm_edit_corosync_conf[edit the corosync.conf file]. A restart should not
784 be necessary. You can check that corosync loaded the new link using:
787 journalctl -b -u corosync
790 It might be a good idea to test the new link by temporarily disconnecting the
791 old link on one node and making sure that its status remains online while
798 If you see a healthy cluster state, it means that your new link is being used.
801 Corosync External Vote Support
802 ------------------------------
804 This section describes a way to deploy an external voter in a {pve} cluster.
805 When configured, the cluster can sustain more node failures without
806 violating safety properties of the cluster communication.
808 For this to work there are two services involved:
810 * a so called qdevice daemon which runs on each {pve} node
812 * an external vote daemon which runs on an independent server.
814 As a result you can achieve higher availability even in smaller setups (for
817 QDevice Technical Overview
818 ~~~~~~~~~~~~~~~~~~~~~~~~~~
820 The Corosync Quroum Device (QDevice) is a daemon which runs on each cluster
821 node. It provides a configured number of votes to the clusters quorum
822 subsystem based on an external running third-party arbitrator's decision.
823 Its primary use is to allow a cluster to sustain more node failures than
824 standard quorum rules allow. This can be done safely as the external device
825 can see all nodes and thus choose only one set of nodes to give its vote.
826 This will only be done if said set of nodes can have quorum (again) when
827 receiving the third-party vote.
829 Currently only 'QDevice Net' is supported as a third-party arbitrator. It is
830 a daemon which provides a vote to a cluster partition if it can reach the
831 partition members over the network. It will give only votes to one partition
832 of a cluster at any time.
833 It's designed to support multiple clusters and is almost configuration and
834 state free. New clusters are handled dynamically and no configuration file
835 is needed on the host running a QDevice.
837 The external host has the only requirement that it needs network access to the
838 cluster and a corosync-qnetd package available. We provide such a package
839 for Debian based hosts, other Linux distributions should also have a package
840 available through their respective package manager.
842 NOTE: In contrast to corosync itself, a QDevice connects to the cluster over
843 TCP/IP. The daemon may even run outside of the clusters LAN and can have longer
849 We support QDevices for clusters with an even number of nodes and recommend
850 it for 2 node clusters, if they should provide higher availability.
851 For clusters with an odd node count we discourage the use of QDevices
852 currently. The reason for this, is the difference of the votes the QDevice
853 provides for each cluster type. Even numbered clusters get single additional
854 vote, with this we can only increase availability, i.e. if the QDevice
855 itself fails we are in the same situation as with no QDevice at all.
857 Now, with an odd numbered cluster size the QDevice provides '(N-1)' votes --
858 where 'N' corresponds to the cluster node count. This difference makes
859 sense, if we had only one additional vote the cluster can get into a split
861 This algorithm would allow that all nodes but one (and naturally the
862 QDevice itself) could fail.
863 There are two drawbacks with this:
865 * If the QNet daemon itself fails, no other node may fail or the cluster
866 immediately loses quorum. For example, in a cluster with 15 nodes 7
867 could fail before the cluster becomes inquorate. But, if a QDevice is
868 configured here and said QDevice fails itself **no single node** of
869 the 15 may fail. The QDevice acts almost as a single point of failure in
872 * The fact that all but one node plus QDevice may fail sound promising at
873 first, but this may result in a mass recovery of HA services that would
874 overload the single node left. Also ceph server will stop to provide
875 services after only '((N-1)/2)' nodes are online.
877 If you understand the drawbacks and implications you can decide yourself if
878 you should use this technology in an odd numbered cluster setup.
883 We recommend to run any daemon which provides votes to corosync-qdevice as an
884 unprivileged user. {pve} and Debian provides a package which is already
886 The traffic between the daemon and the cluster must be encrypted to ensure a
887 safe and secure QDevice integration in {pve}.
889 First install the 'corosync-qnetd' package on your external server and
890 the 'corosync-qdevice' package on all cluster nodes.
892 After that, ensure that all your nodes on the cluster are online.
894 You can now easily set up your QDevice by running the following command on one
898 pve# pvecm qdevice setup <QDEVICE-IP>
901 The SSH key from the cluster will be automatically copied to the QDevice. You
902 might need to enter an SSH password during this step.
904 After you enter the password and all the steps are successfully completed, you
905 will see "Done". You can check the status now:
912 Votequorum information
913 ~~~~~~~~~~~~~~~~~~~~~
918 Flags: Quorate Qdevice
920 Membership information
921 ~~~~~~~~~~~~~~~~~~~~~~
922 Nodeid Votes Qdevice Name
923 0x00000001 1 A,V,NMW 192.168.22.180 (local)
924 0x00000002 1 A,V,NMW 192.168.22.181
929 which means the QDevice is set up.
931 Frequently Asked Questions
932 ~~~~~~~~~~~~~~~~~~~~~~~~~~
937 In case of a tie, where two same-sized cluster partitions cannot see each other
938 but the QDevice, the QDevice chooses randomly one of those partitions and
939 provides a vote to it.
941 Possible Negative Implications
942 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
944 For clusters with an even node count there are no negative implications when
945 setting up a QDevice. If it fails to work, you are as good as without QDevice at
948 Adding/Deleting Nodes After QDevice Setup
949 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
951 If you want to add a new node or remove an existing one from a cluster with a
952 QDevice setup, you need to remove the QDevice first. After that, you can add or
953 remove nodes normally. Once you have a cluster with an even node count again,
954 you can set up the QDevice again as described above.
959 If you used the official `pvecm` tool to add the QDevice, you can remove it
960 trivially by running:
963 pve# pvecm qdevice remove
968 //There is still stuff to add here
971 Corosync Configuration
972 ----------------------
974 The `/etc/pve/corosync.conf` file plays a central role in a {pve} cluster. It
975 controls the cluster membership and its network.
976 For further information about it, check the corosync.conf man page:
982 For node membership you should always use the `pvecm` tool provided by {pve}.
983 You may have to edit the configuration file manually for other changes.
984 Here are a few best practice tips for doing this.
986 [[pvecm_edit_corosync_conf]]
990 Editing the corosync.conf file is not always very straightforward. There are
991 two on each cluster node, one in `/etc/pve/corosync.conf` and the other in
992 `/etc/corosync/corosync.conf`. Editing the one in our cluster file system will
993 propagate the changes to the local one, but not vice versa.
995 The configuration will get updated automatically as soon as the file changes.
996 This means changes which can be integrated in a running corosync will take
997 effect immediately. So you should always make a copy and edit that instead, to
998 avoid triggering some unwanted changes by an in-between safe.
1002 cp /etc/pve/corosync.conf /etc/pve/corosync.conf.new
1005 Then open the config file with your favorite editor, `nano` and `vim.tiny` are
1006 preinstalled on any {pve} node for example.
1008 NOTE: Always increment the 'config_version' number on configuration changes,
1009 omitting this can lead to problems.
1011 After making the necessary changes create another copy of the current working
1012 configuration file. This serves as a backup if the new configuration fails to
1013 apply or makes problems in other ways.
1017 cp /etc/pve/corosync.conf /etc/pve/corosync.conf.bak
1020 Then move the new configuration file over the old one:
1023 mv /etc/pve/corosync.conf.new /etc/pve/corosync.conf
1026 You may check with the commands
1029 systemctl status corosync
1030 journalctl -b -u corosync
1033 If the change could be applied automatically. If not you may have to restart the
1034 corosync service via:
1037 systemctl restart corosync
1040 On errors check the troubleshooting section below.
1045 Issue: 'quorum.expected_votes must be configured'
1046 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1048 When corosync starts to fail and you get the following message in the system log:
1052 corosync[1647]: [QUORUM] Quorum provider: corosync_votequorum failed to initialize.
1053 corosync[1647]: [SERV ] Service engine 'corosync_quorum' failed to load for reason
1054 'configuration error: nodelist or quorum.expected_votes must be configured!'
1058 It means that the hostname you set for corosync 'ringX_addr' in the
1059 configuration could not be resolved.
1061 Write Configuration When Not Quorate
1062 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1064 If you need to change '/etc/pve/corosync.conf' on an node with no quorum, and you
1065 know what you do, use:
1071 This sets the expected vote count to 1 and makes the cluster quorate. You can
1072 now fix your configuration, or revert it back to the last working backup.
1074 This is not enough if corosync cannot start anymore. Here its best to edit the
1075 local copy of the corosync configuration in '/etc/corosync/corosync.conf' so
1076 that corosync can start again. Ensure that on all nodes this configuration has
1077 the same content to avoid split brains. If you are not sure what went wrong
1078 it's best to ask the Proxmox Community to help you.
1081 [[pvecm_corosync_conf_glossary]]
1082 Corosync Configuration Glossary
1083 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1086 This names the different link addresses for the kronosnet connections between
1093 It is obvious that a cluster is not quorate when all nodes are
1094 offline. This is a common case after a power failure.
1096 NOTE: It is always a good idea to use an uninterruptible power supply
1097 (``UPS'', also called ``battery backup'') to avoid this state, especially if
1100 On node startup, the `pve-guests` service is started and waits for
1101 quorum. Once quorate, it starts all guests which have the `onboot`
1104 When you turn on nodes, or when power comes back after power failure,
1105 it is likely that some nodes boots faster than others. Please keep in
1106 mind that guest startup is delayed until you reach quorum.
1112 Migrating virtual guests to other nodes is a useful feature in a
1113 cluster. There are settings to control the behavior of such
1114 migrations. This can be done via the configuration file
1115 `datacenter.cfg` or for a specific migration via API or command line
1118 It makes a difference if a Guest is online or offline, or if it has
1119 local resources (like a local disk).
1121 For Details about Virtual Machine Migration see the
1122 xref:qm_migration[QEMU/KVM Migration Chapter].
1124 For Details about Container Migration see the
1125 xref:pct_migration[Container Migration Chapter].
1130 The migration type defines if the migration data should be sent over an
1131 encrypted (`secure`) channel or an unencrypted (`insecure`) one.
1132 Setting the migration type to insecure means that the RAM content of a
1133 virtual guest gets also transferred unencrypted, which can lead to
1134 information disclosure of critical data from inside the guest (for
1135 example passwords or encryption keys).
1137 Therefore, we strongly recommend using the secure channel if you do
1138 not have full control over the network and can not guarantee that no
1139 one is eavesdropping to it.
1141 NOTE: Storage migration does not follow this setting. Currently, it
1142 always sends the storage content over a secure channel.
1144 Encryption requires a lot of computing power, so this setting is often
1145 changed to "unsafe" to achieve better performance. The impact on
1146 modern systems is lower because they implement AES encryption in
1147 hardware. The performance impact is particularly evident in fast
1148 networks where you can transfer 10 Gbps or more.
1153 By default, {pve} uses the network in which cluster communication
1154 takes place to send the migration traffic. This is not optimal because
1155 sensitive cluster traffic can be disrupted and this network may not
1156 have the best bandwidth available on the node.
1158 Setting the migration network parameter allows the use of a dedicated
1159 network for the entire migration traffic. In addition to the memory,
1160 this also affects the storage traffic for offline migrations.
1162 The migration network is set as a network in the CIDR notation. This
1163 has the advantage that you do not have to set individual IP addresses
1164 for each node. {pve} can determine the real address on the
1165 destination node from the network specified in the CIDR form. To
1166 enable this, the network must be specified so that each node has one,
1167 but only one IP in the respective network.
1172 We assume that we have a three-node setup with three separate
1173 networks. One for public communication with the Internet, one for
1174 cluster communication and a very fast one, which we want to use as a
1175 dedicated network for migration.
1177 A network configuration for such a setup might look as follows:
1180 iface eno1 inet manual
1184 iface vmbr0 inet static
1186 netmask 255.255.250.0
1194 iface eno2 inet static
1196 netmask 255.255.255.0
1200 iface eno3 inet static
1202 netmask 255.255.255.0
1205 Here, we will use the network 10.1.2.0/24 as a migration network. For
1206 a single migration, you can do this using the `migration_network`
1207 parameter of the command line tool:
1210 # qm migrate 106 tre --online --migration_network 10.1.2.0/24
1213 To configure this as the default network for all migrations in the
1214 cluster, set the `migration` property of the `/etc/pve/datacenter.cfg`
1218 # use dedicated migration network
1219 migration: secure,network=10.1.2.0/24
1222 NOTE: The migration type must always be set when the migration network
1223 gets set in `/etc/pve/datacenter.cfg`.
1227 include::pve-copyright.adoc[]