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 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.
64 NOTE: Some switches do not support IP multicast by default and must be
65 manually enabled first.
67 * Date and time have to be synchronized.
69 * SSH tunnel on TCP port 22 between nodes is used.
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
75 * We recommend a dedicated NIC for the cluster traffic, especially if
76 you use shared storage.
78 * Root password of a cluster node is required for adding nodes.
80 NOTE: It is not possible to mix {pve} 3.x and earlier with {pve} 4.X cluster
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.
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 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.
104 [[pvecm_create_cluster]]
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
113 hp1# pvecm create CLUSTERNAME
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.
121 To check the state of your cluster use:
127 Multiple Clusters In Same Network
128 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
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
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.
142 [[pvecm_join_node_to_cluster]]
143 Adding Nodes to the Cluster
144 ---------------------------
146 Login via `ssh` to the node you want to add.
149 hp2# pvecm add IP-ADDRESS-CLUSTER
152 For `IP-ADDRESS-CLUSTER` use the IP or hostname of an existing cluster node.
153 An IP address is recommended (see <<corosync-addresses,Ring Address Types>>).
155 CAUTION: A new node cannot hold any VMs, because you would get
156 conflicts about identical VM IDs. Also, all existing configuration in
157 `/etc/pve` is overwritten when you join a new node to the cluster. To
158 workaround, use `vzdump` to backup and restore to a different VMID after
159 adding the node to the cluster.
161 To check the state of cluster:
167 .Cluster status after adding 4 nodes
172 Date: Mon Apr 20 12:30:13 2015
173 Quorum provider: corosync_votequorum
179 Votequorum information
180 ~~~~~~~~~~~~~~~~~~~~~~
187 Membership information
188 ~~~~~~~~~~~~~~~~~~~~~~
190 0x00000001 1 192.168.15.91
191 0x00000002 1 192.168.15.92 (local)
192 0x00000003 1 192.168.15.93
193 0x00000004 1 192.168.15.94
196 If you only want the list of all nodes use:
202 .List nodes in a cluster
206 Membership information
207 ~~~~~~~~~~~~~~~~~~~~~~
215 [[adding-nodes-with-separated-cluster-network]]
216 Adding Nodes With Separated Cluster Network
217 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
219 When adding a node to a cluster with a separated cluster network you need to
220 use the 'ringX_addr' parameters to set the nodes address on those networks:
224 pvecm add IP-ADDRESS-CLUSTER -ring0_addr IP-ADDRESS-RING0
227 If you want to use the Redundant Ring Protocol you will also want to pass the
228 'ring1_addr' parameter.
231 Remove a Cluster Node
232 ---------------------
234 CAUTION: Read carefully the procedure before proceeding, as it could
235 not be what you want or need.
237 Move all virtual machines from the node. Make sure you have no local
238 data or backups you want to keep, or save them accordingly.
239 In the following example we will remove the node hp4 from the cluster.
241 Log in to a *different* cluster node (not hp4), and issue a `pvecm nodes`
242 command to identify the node ID to remove:
247 Membership information
248 ~~~~~~~~~~~~~~~~~~~~~~
257 At this point you must power off hp4 and
258 make sure that it will not power on again (in the network) as it
261 IMPORTANT: As said above, it is critical to power off the node
262 *before* removal, and make sure that it will *never* power on again
263 (in the existing cluster network) as it is.
264 If you power on the node as it is, your cluster will be screwed up and
265 it could be difficult to restore a clean cluster state.
267 After powering off the node hp4, we can safely remove it from the cluster.
270 hp1# pvecm delnode hp4
273 If the operation succeeds no output is returned, just check the node
274 list again with `pvecm nodes` or `pvecm status`. You should see
282 Date: Mon Apr 20 12:44:28 2015
283 Quorum provider: corosync_votequorum
289 Votequorum information
290 ~~~~~~~~~~~~~~~~~~~~~~
297 Membership information
298 ~~~~~~~~~~~~~~~~~~~~~~
300 0x00000001 1 192.168.15.90 (local)
301 0x00000002 1 192.168.15.91
302 0x00000003 1 192.168.15.92
305 If, for whatever reason, you want that this server joins the same
306 cluster again, you have to
308 * reinstall {pve} on it from scratch
310 * then join it, as explained in the previous section.
312 NOTE: After removal of the node, its SSH fingerprint will still reside in the
313 'known_hosts' of the other nodes. If you receive an SSH error after rejoining
314 a node with the same IP or hostname, run `pvecm updatecerts` once to update
315 the fingerprint on all nodes.
317 [[pvecm_separate_node_without_reinstall]]
318 Separate A Node Without Reinstalling
319 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
321 CAUTION: This is *not* the recommended method, proceed with caution. Use the
322 above mentioned method if you're unsure.
324 You can also separate a node from a cluster without reinstalling it from
325 scratch. But after removing the node from the cluster it will still have
326 access to the shared storages! This must be resolved before you start removing
327 the node from the cluster. A {pve} cluster cannot share the exact same
328 storage with another cluster, as storage locking doesn't work over cluster
329 boundary. Further, it may also lead to VMID conflicts.
331 Its suggested that you create a new storage where only the node which you want
332 to separate has access. This can be an new export on your NFS or a new Ceph
333 pool, to name a few examples. Its just important that the exact same storage
334 does not gets accessed by multiple clusters. After setting this storage up move
335 all data from the node and its VMs to it. Then you are ready to separate the
336 node from the cluster.
338 WARNING: Ensure all shared resources are cleanly separated! You will run into
339 conflicts and problems else.
341 First stop the corosync and the pve-cluster services on the node:
344 systemctl stop pve-cluster
345 systemctl stop corosync
348 Start the cluster filesystem again in local mode:
354 Delete the corosync configuration files:
357 rm /etc/pve/corosync.conf
361 You can now start the filesystem again as normal service:
365 systemctl start pve-cluster
368 The node is now separated from the cluster. You can deleted it from a remaining
369 node of the cluster with:
372 pvecm delnode oldnode
375 If the command failed, because the remaining node in the cluster lost quorum
376 when the now separate node exited, you may set the expected votes to 1 as a workaround:
382 And then repeat the 'pvecm delnode' command.
384 Now switch back to the separated node, here delete all remaining files left
385 from the old cluster. This ensures that the node can be added to another
386 cluster again without problems.
390 rm /var/lib/corosync/*
393 As the configuration files from the other nodes are still in the cluster
394 filesystem you may want to clean those up too. Remove simply the whole
395 directory recursive from '/etc/pve/nodes/NODENAME', but check three times that
396 you used the correct one before deleting it.
398 CAUTION: The nodes SSH keys are still in the 'authorized_key' file, this means
399 the nodes can still connect to each other with public key authentication. This
400 should be fixed by removing the respective keys from the
401 '/etc/pve/priv/authorized_keys' file.
406 {pve} use a quorum-based technique to provide a consistent state among
409 [quote, from Wikipedia, Quorum (distributed computing)]
411 A quorum is the minimum number of votes that a distributed transaction
412 has to obtain in order to be allowed to perform an operation in a
416 In case of network partitioning, state changes requires that a
417 majority of nodes are online. The cluster switches to read-only mode
420 NOTE: {pve} assigns a single vote to each node by default.
425 The cluster network is the core of a cluster. All messages sent over it have to
426 be delivered reliable to all nodes in their respective order. In {pve} this
427 part is done by corosync, an implementation of a high performance low overhead
428 high availability development toolkit. It serves our decentralized
429 configuration file system (`pmxcfs`).
431 [[cluster-network-requirements]]
434 This needs a reliable network with latencies under 2 milliseconds (LAN
435 performance) to work properly. While corosync can also use unicast for
436 communication between nodes its **highly recommended** to have a multicast
437 capable network. The network should not be used heavily by other members,
438 ideally corosync runs on its own network.
439 *never* share it with network where storage communicates too.
441 Before setting up a cluster it is good practice to check if the network is fit
444 * Ensure that all nodes are in the same subnet. This must only be true for the
445 network interfaces used for cluster communication (corosync).
447 * Ensure all nodes can reach each other over those interfaces, using `ping` is
448 enough for a basic test.
450 * Ensure that multicast works in general and a high package rates. This can be
451 done with the `omping` tool. The final "%loss" number should be < 1%.
455 omping -c 10000 -i 0.001 -F -q NODE1-IP NODE2-IP ...
458 * Ensure that multicast communication works over an extended period of time.
459 This uncovers problems where IGMP snooping is activated on the network but
460 no multicast querier is active. This test has a duration of around 10
465 omping -c 600 -i 1 -q NODE1-IP NODE2-IP ...
468 Your network is not ready for clustering if any of these test fails. Recheck
469 your network configuration. Especially switches are notorious for having
470 multicast disabled by default or IGMP snooping enabled with no IGMP querier
473 In smaller cluster its also an option to use unicast if you really cannot get
476 Separate Cluster Network
477 ~~~~~~~~~~~~~~~~~~~~~~~~
479 When creating a cluster without any parameters the cluster network is generally
480 shared with the Web UI and the VMs and its traffic. Depending on your setup
481 even storage traffic may get sent over the same network. Its recommended to
482 change that, as corosync is a time critical real time application.
484 Setting Up A New Network
485 ^^^^^^^^^^^^^^^^^^^^^^^^
487 First you have to setup a new network interface. It should be on a physical
488 separate network. Ensure that your network fulfills the
489 <<cluster-network-requirements,cluster network requirements>>.
491 Separate On Cluster Creation
492 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
494 This is possible through the 'ring0_addr' and 'bindnet0_addr' parameter of
495 the 'pvecm create' command used for creating a new cluster.
497 If you have setup an additional NIC with a static address on 10.10.10.1/25
498 and want to send and receive all cluster communication over this interface
503 pvecm create test --ring0_addr 10.10.10.1 --bindnet0_addr 10.10.10.0
506 To check if everything is working properly execute:
509 systemctl status corosync
512 Afterwards, proceed as descripted in the section to
513 <<adding-nodes-with-separated-cluster-network,add nodes with a separated cluster network>>.
515 [[separate-cluster-net-after-creation]]
516 Separate After Cluster Creation
517 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
519 You can do this also if you have already created a cluster and want to switch
520 its communication to another network, without rebuilding the whole cluster.
521 This change may lead to short durations of quorum loss in the cluster, as nodes
522 have to restart corosync and come up one after the other on the new network.
524 Check how to <<edit-corosync-conf,edit the corosync.conf file>> first.
525 The open it and you should see a file similar to:
559 provider: corosync_votequorum
563 cluster_name: thomas-testcluster
569 bindnetaddr: 192.168.30.50
576 The first you want to do is add the 'name' properties in the node entries if
577 you do not see them already. Those *must* match the node name.
579 Then replace the address from the 'ring0_addr' properties with the new
580 addresses. You may use plain IP addresses or also hostnames here. If you use
581 hostnames ensure that they are resolvable from all nodes. (see also
582 <<corosync-addresses,Ring Address Types>>)
584 In my example I want to switch my cluster communication to the 10.10.10.1/25
585 network. So I replace all 'ring0_addr' respectively. I also set the bindnetaddr
586 in the totem section of the config to an address of the new network. It can be
587 any address from the subnet configured on the new network interface.
589 After you increased the 'config_version' property the new configuration file
605 ring0_addr: 10.10.10.2
612 ring0_addr: 10.10.10.3
619 ring0_addr: 10.10.10.1
625 provider: corosync_votequorum
629 cluster_name: thomas-testcluster
635 bindnetaddr: 10.10.10.1
642 Now after a final check whether all changed information is correct we save it
643 and see again the <<edit-corosync-conf,edit corosync.conf file>> section to
644 learn how to bring it in effect.
646 As our change cannot be enforced live from corosync we have to do an restart.
648 On a single node execute:
651 systemctl restart corosync
654 Now check if everything is fine:
658 systemctl status corosync
661 If corosync runs again correct restart corosync also on all other nodes.
662 They will then join the cluster membership one by one on the new network.
664 [[corosync-addresses]]
668 A corosync link or ring address can be specified in two ways:
670 * **IPv4/v6 addresses** will be used directly. They are recommended, since they
671 are static and usually not changed carelessly.
673 * **Hostnames** will be resolved using `getaddrinfo`, which means that per
674 default, IPv6 addresses will be used first, if available (see also
675 `man gai.conf`). Keep this in mind, especially when upgrading an existing
678 CAUTION: Hostnames should be used with care, since the address they
679 resolve to can be changed without touching corosync or the node it runs on -
680 which may lead to a situation where an address is changed without thinking
681 about implications for corosync.
683 A seperate, static hostname specifically for corosync is recommended, if
684 hostnames are preferred. Also, make sure that every node in the cluster can
685 resolve all hostnames correctly.
687 Since {pve} 5.1, while supported, hostnames will be resolved at the time of
688 entry. Only the resolved IP is then saved to the configuration.
690 Nodes that joined the cluster on earlier versions likely still use their
691 unresolved hostname in `corosync.conf`. It might be a good idea to replace
692 them with IPs or a seperate hostname, as mentioned above.
695 Redundant Ring Protocol
696 ~~~~~~~~~~~~~~~~~~~~~~~
697 To avoid a single point of failure you should implement counter measurements.
698 This can be on the hardware and operating system level through network bonding.
700 Corosync itself offers also a possibility to add redundancy through the so
701 called 'Redundant Ring Protocol'. This protocol allows running a second totem
702 ring on another network, this network should be physically separated from the
703 other rings network to actually increase availability.
705 RRP On Cluster Creation
706 ~~~~~~~~~~~~~~~~~~~~~~~
708 The 'pvecm create' command provides the additional parameters 'bindnetX_addr',
709 'ringX_addr' and 'rrp_mode', can be used for RRP configuration.
711 NOTE: See the <<corosync-conf-glossary,glossary>> if you do not know what each parameter means.
713 So if you have two networks, one on the 10.10.10.1/24 and the other on the
714 10.10.20.1/24 subnet you would execute:
718 pvecm create CLUSTERNAME -bindnet0_addr 10.10.10.1 -ring0_addr 10.10.10.1 \
719 -bindnet1_addr 10.10.20.1 -ring1_addr 10.10.20.1
722 RRP On Existing Clusters
723 ~~~~~~~~~~~~~~~~~~~~~~~~
725 You will take similar steps as described in
726 <<separate-cluster-net-after-creation,separating the cluster network>> to
727 enable RRP on an already running cluster. The single difference is, that you
728 will add `ring1` and use it instead of `ring0`.
730 First add a new `interface` subsection in the `totem` section, set its
731 `ringnumber` property to `1`. Set the interfaces `bindnetaddr` property to an
732 address of the subnet you have configured for your new ring.
733 Further set the `rrp_mode` to `passive`, this is the only stable mode.
735 Then add to each node entry in the `nodelist` section its new `ring1_addr`
736 property with the nodes additional ring address.
738 So if you have two networks, one on the 10.10.10.1/24 and the other on the
739 10.10.20.1/24 subnet, the final configuration file should look like:
750 bindnetaddr: 10.10.10.1
754 bindnetaddr: 10.10.20.1
764 ring0_addr: 10.10.10.1
765 ring1_addr: 10.10.20.1
772 ring0_addr: 10.10.10.2
773 ring1_addr: 10.10.20.2
776 [...] # other cluster nodes here
779 [...] # other remaining config sections here
783 Bring it in effect like described in the
784 <<edit-corosync-conf,edit the corosync.conf file>> section.
786 This is a change which cannot take live in effect and needs at least a restart
787 of corosync. Recommended is a restart of the whole cluster.
789 If you cannot reboot the whole cluster ensure no High Availability services are
790 configured and the stop the corosync service on all nodes. After corosync is
791 stopped on all nodes start it one after the other again.
793 Corosync External Vote Support
794 ------------------------------
796 This section describes a way to deploy an external voter in a {pve} cluster.
797 When configured, the cluster can sustain more node failures without
798 violating safety properties of the cluster communication.
800 For this to work there are two services involved:
802 * a so called qdevice daemon which runs on each {pve} node
804 * an external vote daemon which runs on an independent server.
806 As a result you can achieve higher availability even in smaller setups (for
809 QDevice Technical Overview
810 ~~~~~~~~~~~~~~~~~~~~~~~~~~
812 The Corosync Quroum Device (QDevice) is a daemon which runs on each cluster
813 node. It provides a configured number of votes to the clusters quorum
814 subsystem based on an external running third-party arbitrator's decision.
815 Its primary use is to allow a cluster to sustain more node failures than
816 standard quorum rules allow. This can be done safely as the external device
817 can see all nodes and thus choose only one set of nodes to give its vote.
818 This will only be done if said set of nodes can have quorum (again) when
819 receiving the third-party vote.
821 Currently only 'QDevice Net' is supported as a third-party arbitrator. It is
822 a daemon which provides a vote to a cluster partition if it can reach the
823 partition members over the network. It will give only votes to one partition
824 of a cluster at any time.
825 It's designed to support multiple clusters and is almost configuration and
826 state free. New clusters are handled dynamically and no configuration file
827 is needed on the host running a QDevice.
829 The external host has the only requirement that it needs network access to the
830 cluster and a corosync-qnetd package available. We provide such a package
831 for Debian based hosts, other Linux distributions should also have a package
832 available through their respective package manager.
834 NOTE: In contrast to corosync itself, a QDevice connects to the cluster over
835 TCP/IP and thus does not need a multicast capable network between itself and
836 the cluster. In fact the daemon may run outside of the LAN and can have
837 longer latencies than 2 ms.
843 We support QDevices for clusters with an even number of nodes and recommend
844 it for 2 node clusters, if they should provide higher availability.
845 For clusters with an odd node count we discourage the use of QDevices
846 currently. The reason for this, is the difference of the votes the QDevice
847 provides for each cluster type. Even numbered clusters get single additional
848 vote, with this we can only increase availability, i.e. if the QDevice
849 itself fails we are in the same situation as with no QDevice at all.
851 Now, with an odd numbered cluster size the QDevice provides '(N-1)' votes --
852 where 'N' corresponds to the cluster node count. This difference makes
853 sense, if we had only one additional vote the cluster can get into a split
855 This algorithm would allow that all nodes but one (and naturally the
856 QDevice itself) could fail.
857 There are two drawbacks with this:
859 * If the QNet daemon itself fails, no other node may fail or the cluster
860 immediately loses quorum. For example, in a cluster with 15 nodes 7
861 could fail before the cluster becomes inquorate. But, if a QDevice is
862 configured here and said QDevice fails itself **no single node** of
863 the 15 may fail. The QDevice acts almost as a single point of failure in
866 * The fact that all but one node plus QDevice may fail sound promising at
867 first, but this may result in a mass recovery of HA services that would
868 overload the single node left. Also ceph server will stop to provide
869 services after only '((N-1)/2)' nodes are online.
871 If you understand the drawbacks and implications you can decide yourself if
872 you should use this technology in an odd numbered cluster setup.
878 We recommend to run any daemon which provides votes to corosync-qdevice as an
879 unprivileged user. {pve} and Debian provides a package which is already
881 The traffic between the daemon and the cluster must be encrypted to ensure a
882 safe and secure QDevice integration in {pve}.
884 First install the 'corosync-qnetd' package on your external server and
885 the 'corosync-qdevice' package on all cluster nodes.
887 After that, ensure that all your nodes on the cluster are online.
889 You can now easily set up your QDevice by running the following command on one
893 pve# pvecm qdevice setup <QDEVICE-IP>
896 The SSH key from the cluster will be automatically copied to the QDevice. You
897 might need to enter an SSH password during this step.
899 After you enter the password and all the steps are successfully completed, you
900 will see "Done". You can check the status now:
907 Votequorum information
908 ~~~~~~~~~~~~~~~~~~~~~
913 Flags: Quorate Qdevice
915 Membership information
916 ~~~~~~~~~~~~~~~~~~~~~~
917 Nodeid Votes Qdevice Name
918 0x00000001 1 A,V,NMW 192.168.22.180 (local)
919 0x00000002 1 A,V,NMW 192.168.22.181
924 which means the QDevice is set up.
927 Frequently Asked Questions
928 ~~~~~~~~~~~~~~~~~~~~~~~~~~
933 In case of a tie, where two same-sized cluster partitions cannot see each other
934 but the QDevice, the QDevice chooses randomly one of those partitions and
935 provides a vote to it.
937 Possible Negative Implications
938 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
940 For clusters with an even node count there are no negative implications when
941 setting up a QDevice. If it fails to work, you are as good as without QDevice at
944 Adding/Deleting Nodes After QDevice Setup
945 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
947 If you want to add a new node or remove an existing one from a cluster with a
948 QDevice setup, you need to remove the QDevice first. After that, you can add or
949 remove nodes normally. Once you have a cluster with an even node count again,
950 you can set up the QDevice again as described above.
955 If you used the official `pvecm` tool to add the QDevice, you can remove it
956 trivially by running:
959 pve# pvecm qdevice remove
964 //There ist still stuff to add here
967 Corosync Configuration
968 ----------------------
970 The `/etc/pve/corosync.conf` file plays a central role in {pve} cluster. It
971 controls the cluster member ship and its network.
972 For reading more about it check the corosync.conf man page:
978 For node membership you should always use the `pvecm` tool provided by {pve}.
979 You may have to edit the configuration file manually for other changes.
980 Here are a few best practice tips for doing this.
982 [[edit-corosync-conf]]
986 Editing the corosync.conf file can be not always straight forward. There are
987 two on each cluster, one in `/etc/pve/corosync.conf` and the other in
988 `/etc/corosync/corosync.conf`. Editing the one in our cluster file system will
989 propagate the changes to the local one, but not vice versa.
991 The configuration will get updated automatically as soon as the file changes.
992 This means changes which can be integrated in a running corosync will take
993 instantly effect. So you should always make a copy and edit that instead, to
994 avoid triggering some unwanted changes by an in between safe.
998 cp /etc/pve/corosync.conf /etc/pve/corosync.conf.new
1001 Then open the Config file with your favorite editor, `nano` and `vim.tiny` are
1002 preinstalled on {pve} for example.
1004 NOTE: Always increment the 'config_version' number on configuration changes,
1005 omitting this can lead to problems.
1007 After making the necessary changes create another copy of the current working
1008 configuration file. This serves as a backup if the new configuration fails to
1009 apply or makes problems in other ways.
1013 cp /etc/pve/corosync.conf /etc/pve/corosync.conf.bak
1016 Then move the new configuration file over the old one:
1019 mv /etc/pve/corosync.conf.new /etc/pve/corosync.conf
1022 You may check with the commands
1025 systemctl status corosync
1026 journalctl -b -u corosync
1029 If the change could applied automatically. If not you may have to restart the
1030 corosync service via:
1033 systemctl restart corosync
1036 On errors check the troubleshooting section below.
1041 Issue: 'quorum.expected_votes must be configured'
1042 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1044 When corosync starts to fail and you get the following message in the system log:
1048 corosync[1647]: [QUORUM] Quorum provider: corosync_votequorum failed to initialize.
1049 corosync[1647]: [SERV ] Service engine 'corosync_quorum' failed to load for reason
1050 'configuration error: nodelist or quorum.expected_votes must be configured!'
1054 It means that the hostname you set for corosync 'ringX_addr' in the
1055 configuration could not be resolved.
1058 Write Configuration When Not Quorate
1059 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1061 If you need to change '/etc/pve/corosync.conf' on an node with no quorum, and you
1062 know what you do, use:
1068 This sets the expected vote count to 1 and makes the cluster quorate. You can
1069 now fix your configuration, or revert it back to the last working backup.
1071 This is not enough if corosync cannot start anymore. Here its best to edit the
1072 local copy of the corosync configuration in '/etc/corosync/corosync.conf' so
1073 that corosync can start again. Ensure that on all nodes this configuration has
1074 the same content to avoid split brains. If you are not sure what went wrong
1075 it's best to ask the Proxmox Community to help you.
1078 [[corosync-conf-glossary]]
1079 Corosync Configuration Glossary
1080 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1083 This names the different ring addresses for the corosync totem rings used for
1084 the cluster communication.
1087 Defines to which interface the ring should bind to. It may be any address of
1088 the subnet configured on the interface we want to use. In general its the
1089 recommended to just use an address a node uses on this interface.
1092 Specifies the mode of the redundant ring protocol and may be passive, active or
1093 none. Note that use of active is highly experimental and not official
1094 supported. Passive is the preferred mode, it may double the cluster
1095 communication throughput and increases availability.
1101 It is obvious that a cluster is not quorate when all nodes are
1102 offline. This is a common case after a power failure.
1104 NOTE: It is always a good idea to use an uninterruptible power supply
1105 (``UPS'', also called ``battery backup'') to avoid this state, especially if
1108 On node startup, the `pve-guests` service is started and waits for
1109 quorum. Once quorate, it starts all guests which have the `onboot`
1112 When you turn on nodes, or when power comes back after power failure,
1113 it is likely that some nodes boots faster than others. Please keep in
1114 mind that guest startup is delayed until you reach quorum.
1120 Migrating virtual guests to other nodes is a useful feature in a
1121 cluster. There are settings to control the behavior of such
1122 migrations. This can be done via the configuration file
1123 `datacenter.cfg` or for a specific migration via API or command line
1126 It makes a difference if a Guest is online or offline, or if it has
1127 local resources (like a local disk).
1129 For Details about Virtual Machine Migration see the
1130 xref:qm_migration[QEMU/KVM Migration Chapter]
1132 For Details about Container Migration see the
1133 xref:pct_migration[Container Migration Chapter]
1138 The migration type defines if the migration data should be sent over an
1139 encrypted (`secure`) channel or an unencrypted (`insecure`) one.
1140 Setting the migration type to insecure means that the RAM content of a
1141 virtual guest gets also transferred unencrypted, which can lead to
1142 information disclosure of critical data from inside the guest (for
1143 example passwords or encryption keys).
1145 Therefore, we strongly recommend using the secure channel if you do
1146 not have full control over the network and can not guarantee that no
1147 one is eavesdropping to it.
1149 NOTE: Storage migration does not follow this setting. Currently, it
1150 always sends the storage content over a secure channel.
1152 Encryption requires a lot of computing power, so this setting is often
1153 changed to "unsafe" to achieve better performance. The impact on
1154 modern systems is lower because they implement AES encryption in
1155 hardware. The performance impact is particularly evident in fast
1156 networks where you can transfer 10 Gbps or more.
1162 By default, {pve} uses the network in which cluster communication
1163 takes place to send the migration traffic. This is not optimal because
1164 sensitive cluster traffic can be disrupted and this network may not
1165 have the best bandwidth available on the node.
1167 Setting the migration network parameter allows the use of a dedicated
1168 network for the entire migration traffic. In addition to the memory,
1169 this also affects the storage traffic for offline migrations.
1171 The migration network is set as a network in the CIDR notation. This
1172 has the advantage that you do not have to set individual IP addresses
1173 for each node. {pve} can determine the real address on the
1174 destination node from the network specified in the CIDR form. To
1175 enable this, the network must be specified so that each node has one,
1176 but only one IP in the respective network.
1182 We assume that we have a three-node setup with three separate
1183 networks. One for public communication with the Internet, one for
1184 cluster communication and a very fast one, which we want to use as a
1185 dedicated network for migration.
1187 A network configuration for such a setup might look as follows:
1190 iface eno1 inet manual
1194 iface vmbr0 inet static
1196 netmask 255.255.250.0
1204 iface eno2 inet static
1206 netmask 255.255.255.0
1210 iface eno3 inet static
1212 netmask 255.255.255.0
1215 Here, we will use the network 10.1.2.0/24 as a migration network. For
1216 a single migration, you can do this using the `migration_network`
1217 parameter of the command line tool:
1220 # qm migrate 106 tre --online --migration_network 10.1.2.0/24
1223 To configure this as the default network for all migrations in the
1224 cluster, set the `migration` property of the `/etc/pve/datacenter.cfg`
1228 # use dedicated migration network
1229 migration: secure,network=10.1.2.0/24
1232 NOTE: The migration type must always be set when the migration network
1233 gets set in `/etc/pve/datacenter.cfg`.
1237 include::pve-copyright.adoc[]