9 pvecm - Proxmox VE Cluster Manager
14 include::pvecm.1-synopsis.adoc[]
26 The {PVE} cluster manager `pvecm` is a tool to create a group of
27 physical servers. Such a group is called a *cluster*. We use the
28 http://www.corosync.org[Corosync Cluster Engine] for reliable group
29 communication, and such clusters can consist of up to 32 physical nodes
30 (probably more, dependent on network latency).
32 `pvecm` can be used to create a new cluster, join nodes to a cluster,
33 leave the cluster, get status information and do various other cluster
34 related tasks. The **P**rox**m**o**x** **C**luster **F**ile **S**ystem (``pmxcfs'')
35 is used to transparently distribute the cluster configuration to all cluster
38 Grouping nodes into a cluster has the following advantages:
40 * Centralized, web based management
42 * Multi-master clusters: each node can do all management task
44 * `pmxcfs`: database-driven file system for storing configuration files,
45 replicated in real-time on all nodes using `corosync`.
47 * Easy migration of virtual machines and containers between physical
52 * Cluster-wide services like firewall and HA
58 * All nodes must be in the same network as `corosync` uses IP Multicast
59 to communicate between nodes (also see
60 http://www.corosync.org[Corosync Cluster Engine]). Corosync uses UDP
61 ports 5404 and 5405 for cluster communication.
63 NOTE: Some switches do not support IP multicast by default and must be
64 manually enabled first.
66 * Date and time have to be synchronized.
68 * SSH tunnel on TCP port 22 between nodes is used.
70 * If you are interested in High Availability, you need to have at
71 least three nodes for reliable quorum. All nodes should have the
74 * We recommend a dedicated NIC for the cluster traffic, especially if
75 you use shared storage.
77 NOTE: It is not possible to mix Proxmox VE 3.x and earlier with
78 Proxmox VE 4.0 cluster nodes.
84 First, install {PVE} on all nodes. Make sure that each node is
85 installed with the final hostname and IP configuration. Changing the
86 hostname and IP is not possible after cluster creation.
88 Currently the cluster creation has to be done on the console, so you
89 need to login via `ssh`.
94 Login via `ssh` to the first {pve} node. Use a unique name for your cluster.
95 This name cannot be changed later.
97 hp1# pvecm create YOUR-CLUSTER-NAME
99 CAUTION: The cluster name is used to compute the default multicast
100 address. Please use unique cluster names if you run more than one
101 cluster inside your network.
103 To check the state of your cluster use:
107 Multiple Clusters In Same Network
108 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
110 It is possible to create multiple clusters in the same physical or logical
111 network. Each cluster must have a unique name, which is used to generate the
112 cluster's multicast group address. As long as no duplicate cluster names are
113 configured in one network segment, the different clusters won't interfere with
116 If multiple clusters operate in a single network it may be beneficial to setup
117 an IGMP querier and enable IGMP Snooping in said network. This may reduce the
118 load of the network significantly because multicast packets are only delivered
119 to endpoints of the respective member nodes.
122 Adding Nodes to the Cluster
123 ---------------------------
125 Login via `ssh` to the node you want to add.
127 hp2# pvecm add IP-ADDRESS-CLUSTER
129 For `IP-ADDRESS-CLUSTER` use the IP from an existing cluster node.
131 CAUTION: A new node cannot hold any VMs, because you would get
132 conflicts about identical VM IDs. Also, all existing configuration in
133 `/etc/pve` is overwritten when you join a new node to the cluster. To
134 workaround, use `vzdump` to backup and restore to a different VMID after
135 adding the node to the cluster.
137 To check the state of cluster:
141 .Cluster status after adding 4 nodes
146 Date: Mon Apr 20 12:30:13 2015
147 Quorum provider: corosync_votequorum
153 Votequorum information
154 ~~~~~~~~~~~~~~~~~~~~~~
161 Membership information
162 ~~~~~~~~~~~~~~~~~~~~~~
164 0x00000001 1 192.168.15.91
165 0x00000002 1 192.168.15.92 (local)
166 0x00000003 1 192.168.15.93
167 0x00000004 1 192.168.15.94
170 If you only want the list of all nodes use:
174 .List nodes in a cluster
178 Membership information
179 ~~~~~~~~~~~~~~~~~~~~~~
187 Adding Nodes With Separated Cluster Network
188 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
190 When adding a node to a cluster with a separated cluster network you need to
191 use the 'ringX_addr' parameters to set the nodes address on those networks:
195 pvecm add IP-ADDRESS-CLUSTER -ring0_addr IP-ADDRESS-RING0
198 If you want to use the Redundant Ring Protocol you will also want to pass the
199 'ring1_addr' parameter.
202 Remove a Cluster Node
203 ---------------------
205 CAUTION: Read carefully the procedure before proceeding, as it could
206 not be what you want or need.
208 Move all virtual machines from the node. Make sure you have no local
209 data or backups you want to keep, or save them accordingly.
210 In the following example we will remove the node hp4 from the cluster.
212 Log in to a *different* cluster node (not hp4), and issue a `pvecm nodes`
213 command to identify the node ID to remove:
218 Membership information
219 ~~~~~~~~~~~~~~~~~~~~~~
228 At this point you must power off hp4 and
229 make sure that it will not power on again (in the network) as it
232 IMPORTANT: As said above, it is critical to power off the node
233 *before* removal, and make sure that it will *never* power on again
234 (in the existing cluster network) as it is.
235 If you power on the node as it is, your cluster will be screwed up and
236 it could be difficult to restore a clean cluster state.
238 After powering off the node hp4, we can safely remove it from the cluster.
240 hp1# pvecm delnode hp4
242 If the operation succeeds no output is returned, just check the node
243 list again with `pvecm nodes` or `pvecm status`. You should see
251 Date: Mon Apr 20 12:44:28 2015
252 Quorum provider: corosync_votequorum
258 Votequorum information
259 ~~~~~~~~~~~~~~~~~~~~~~
266 Membership information
267 ~~~~~~~~~~~~~~~~~~~~~~
269 0x00000001 1 192.168.15.90 (local)
270 0x00000002 1 192.168.15.91
271 0x00000003 1 192.168.15.92
274 If, for whatever reason, you want that this server joins the same
275 cluster again, you have to
277 * reinstall {pve} on it from scratch
279 * then join it, as explained in the previous section.
281 [[pvecm_separate_node_without_reinstall]]
282 Separate A Node Without Reinstalling
283 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
285 CAUTION: This is *not* the recommended method, proceed with caution. Use the
286 above mentioned method if you're unsure.
288 You can also separate a node from a cluster without reinstalling it from
289 scratch. But after removing the node from the cluster it will still have
290 access to the shared storages! This must be resolved before you start removing
291 the node from the cluster. A {pve} cluster cannot share the exact same
292 storage with another cluster, as storage locking doesn't work over cluster
293 boundary. Further, it may also lead to VMID conflicts.
295 Its suggested that you create a new storage where only the node which you want
296 to separate has access. This can be an new export on your NFS or a new Ceph
297 pool, to name a few examples. Its just important that the exact same storage
298 does not gets accessed by multiple clusters. After setting this storage up move
299 all data from the node and its VMs to it. Then you are ready to separate the
300 node from the cluster.
302 WARNING: Ensure all shared resources are cleanly separated! You will run into
303 conflicts and problems else.
305 First stop the corosync and the pve-cluster services on the node:
308 systemctl stop pve-cluster
309 systemctl stop corosync
312 Start the cluster filesystem again in local mode:
318 Delete the corosync configuration files:
321 rm /etc/pve/corosync.conf
325 You can now start the filesystem again as normal service:
329 systemctl start pve-cluster
332 The node is now separated from the cluster. You can deleted it from a remaining
333 node of the cluster with:
336 pvecm delnode oldnode
339 If the command failed, because the remaining node in the cluster lost quorum
340 when the now separate node exited, you may set the expected votes to 1 as a workaround:
346 And the repeat the 'pvecm delnode' command.
348 Now switch back to the separated node, here delete all remaining files left
349 from the old cluster. This ensures that the node can be added to another
350 cluster again without problems.
354 rm /var/lib/corosync/*
357 As the configuration files from the other nodes are still in the cluster
358 filesystem you may want to clean those up too. Remove simply the whole
359 directory recursive from '/etc/pve/nodes/NODENAME', but check three times that
360 you used the correct one before deleting it.
362 CAUTION: The nodes SSH keys are still in the 'authorized_key' file, this means
363 the nodes can still connect to each other with public key authentication. This
364 should be fixed by removing the respective keys from the
365 '/etc/pve/priv/authorized_keys' file.
370 {pve} use a quorum-based technique to provide a consistent state among
373 [quote, from Wikipedia, Quorum (distributed computing)]
375 A quorum is the minimum number of votes that a distributed transaction
376 has to obtain in order to be allowed to perform an operation in a
380 In case of network partitioning, state changes requires that a
381 majority of nodes are online. The cluster switches to read-only mode
384 NOTE: {pve} assigns a single vote to each node by default.
389 The cluster network is the core of a cluster. All messages sent over it have to
390 be delivered reliable to all nodes in their respective order. In {pve} this
391 part is done by corosync, an implementation of a high performance low overhead
392 high availability development toolkit. It serves our decentralized
393 configuration file system (`pmxcfs`).
395 [[cluster-network-requirements]]
398 This needs a reliable network with latencies under 2 milliseconds (LAN
399 performance) to work properly. While corosync can also use unicast for
400 communication between nodes its **highly recommended** to have a multicast
401 capable network. The network should not be used heavily by other members,
402 ideally corosync runs on its own network.
403 *never* share it with network where storage communicates too.
405 Before setting up a cluster it is good practice to check if the network is fit
408 * Ensure that all nodes are in the same subnet. This must only be true for the
409 network interfaces used for cluster communication (corosync).
411 * Ensure all nodes can reach each other over those interfaces, using `ping` is
412 enough for a basic test.
414 * Ensure that multicast works in general and a high package rates. This can be
415 done with the `omping` tool. The final "%loss" number should be < 1%.
419 omping -c 10000 -i 0.001 -F -q NODE1-IP NODE2-IP ...
422 * Ensure that multicast communication works over an extended period of time.
423 This uncovers problems where IGMP snooping is activated on the network but
424 no multicast querier is active. This test has a duration of around 10
429 omping -c 600 -i 1 -q NODE1-IP NODE2-IP ...
432 Your network is not ready for clustering if any of these test fails. Recheck
433 your network configuration. Especially switches are notorious for having
434 multicast disabled by default or IGMP snooping enabled with no IGMP querier
437 In smaller cluster its also an option to use unicast if you really cannot get
440 Separate Cluster Network
441 ~~~~~~~~~~~~~~~~~~~~~~~~
443 When creating a cluster without any parameters the cluster network is generally
444 shared with the Web UI and the VMs and its traffic. Depending on your setup
445 even storage traffic may get sent over the same network. Its recommended to
446 change that, as corosync is a time critical real time application.
448 Setting Up A New Network
449 ^^^^^^^^^^^^^^^^^^^^^^^^
451 First you have to setup a new network interface. It should be on a physical
452 separate network. Ensure that your network fulfills the
453 <<cluster-network-requirements,cluster network requirements>>.
455 Separate On Cluster Creation
456 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
458 This is possible through the 'ring0_addr' and 'bindnet0_addr' parameter of
459 the 'pvecm create' command used for creating a new cluster.
461 If you have setup an additional NIC with a static address on 10.10.10.1/25
462 and want to send and receive all cluster communication over this interface
467 pvecm create test --ring0_addr 10.10.10.1 --bindnet0_addr 10.10.10.0
470 To check if everything is working properly execute:
473 systemctl status corosync
476 [[separate-cluster-net-after-creation]]
477 Separate After Cluster Creation
478 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
480 You can do this also if you have already created a cluster and want to switch
481 its communication to another network, without rebuilding the whole cluster.
482 This change may lead to short durations of quorum loss in the cluster, as nodes
483 have to restart corosync and come up one after the other on the new network.
485 Check how to <<edit-corosync-conf,edit the corosync.conf file>> first.
486 The open it and you should see a file similar to:
520 provider: corosync_votequorum
524 cluster_name: thomas-testcluster
530 bindnetaddr: 192.168.30.50
537 The first you want to do is add the 'name' properties in the node entries if
538 you do not see them already. Those *must* match the node name.
540 Then replace the address from the 'ring0_addr' properties with the new
541 addresses. You may use plain IP addresses or also hostnames here. If you use
542 hostnames ensure that they are resolvable from all nodes.
544 In my example I want to switch my cluster communication to the 10.10.10.1/25
545 network. So I replace all 'ring0_addr' respectively. I also set the bindnetaddr
546 in the totem section of the config to an address of the new network. It can be
547 any address from the subnet configured on the new network interface.
549 After you increased the 'config_version' property the new configuration file
565 ring0_addr: 10.10.10.2
572 ring0_addr: 10.10.10.3
579 ring0_addr: 10.10.10.1
585 provider: corosync_votequorum
589 cluster_name: thomas-testcluster
595 bindnetaddr: 10.10.10.1
602 Now after a final check whether all changed information is correct we save it
603 and see again the <<edit-corosync-conf,edit corosync.conf file>> section to
604 learn how to bring it in effect.
606 As our change cannot be enforced live from corosync we have to do an restart.
608 On a single node execute:
611 systemctl restart corosync
614 Now check if everything is fine:
618 systemctl status corosync
621 If corosync runs again correct restart corosync also on all other nodes.
622 They will then join the cluster membership one by one on the new network.
624 Redundant Ring Protocol
625 ~~~~~~~~~~~~~~~~~~~~~~~
626 To avoid a single point of failure you should implement counter measurements.
627 This can be on the hardware and operating system level through network bonding.
629 Corosync itself offers also a possibility to add redundancy through the so
630 called 'Redundant Ring Protocol'. This protocol allows running a second totem
631 ring on another network, this network should be physically separated from the
632 other rings network to actually increase availability.
634 RRP On Cluster Creation
635 ~~~~~~~~~~~~~~~~~~~~~~~
637 The 'pvecm create' command provides the additional parameters 'bindnetX_addr',
638 'ringX_addr' and 'rrp_mode', can be used for RRP configuration.
640 NOTE: See the <<corosync-conf-glossary,glossary>> if you do not know what each parameter means.
642 So if you have two networks, one on the 10.10.10.1/24 and the other on the
643 10.10.20.1/24 subnet you would execute:
647 pvecm create CLUSTERNAME -bindnet0_addr 10.10.10.1 -ring0_addr 10.10.10.1 \
648 -bindnet1_addr 10.10.20.1 -ring1_addr 10.10.20.1
651 RRP On Existing Clusters
652 ~~~~~~~~~~~~~~~~~~~~~~~~
654 You will take similar steps as described in
655 <<separate-cluster-net-after-creation,separating the cluster network>> to
656 enable RRP on an already running cluster. The single difference is, that you
657 will add `ring1` and use it instead of `ring0`.
659 First add a new `interface` subsection in the `totem` section, set its
660 `ringnumber` property to `1`. Set the interfaces `bindnetaddr` property to an
661 address of the subnet you have configured for your new ring.
662 Further set the `rrp_mode` to `passive`, this is the only stable mode.
664 Then add to each node entry in the `nodelist` section its new `ring1_addr`
665 property with the nodes additional ring address.
667 So if you have two networks, one on the 10.10.10.1/24 and the other on the
668 10.10.20.1/24 subnet, the final configuration file should look like:
679 bindnetaddr: 10.10.10.1
683 bindnetaddr: 10.10.20.1
693 ring0_addr: 10.10.10.1
694 ring1_addr: 10.10.20.1
701 ring0_addr: 10.10.10.2
702 ring1_addr: 10.10.20.2
705 [...] # other cluster nodes here
708 [...] # other remaining config sections here
712 Bring it in effect like described in the
713 <<edit-corosync-conf,edit the corosync.conf file>> section.
715 This is a change which cannot take live in effect and needs at least a restart
716 of corosync. Recommended is a restart of the whole cluster.
718 If you cannot reboot the whole cluster ensure no High Availability services are
719 configured and the stop the corosync service on all nodes. After corosync is
720 stopped on all nodes start it one after the other again.
722 Corosync Configuration
723 ----------------------
725 The `/etc/pve/corosync.conf` file plays a central role in {pve} cluster. It
726 controls the cluster member ship and its network.
727 For reading more about it check the corosync.conf man page:
733 For node membership you should always use the `pvecm` tool provided by {pve}.
734 You may have to edit the configuration file manually for other changes.
735 Here are a few best practice tips for doing this.
737 [[edit-corosync-conf]]
741 Editing the corosync.conf file can be not always straight forward. There are
742 two on each cluster, one in `/etc/pve/corosync.conf` and the other in
743 `/etc/corosync/corosync.conf`. Editing the one in our cluster file system will
744 propagate the changes to the local one, but not vice versa.
746 The configuration will get updated automatically as soon as the file changes.
747 This means changes which can be integrated in a running corosync will take
748 instantly effect. So you should always make a copy and edit that instead, to
749 avoid triggering some unwanted changes by an in between safe.
753 cp /etc/pve/corosync.conf /etc/pve/corosync.conf.new
756 Then open the Config file with your favorite editor, `nano` and `vim.tiny` are
757 preinstalled on {pve} for example.
759 NOTE: Always increment the 'config_version' number on configuration changes,
760 omitting this can lead to problems.
762 After making the necessary changes create another copy of the current working
763 configuration file. This serves as a backup if the new configuration fails to
764 apply or makes problems in other ways.
768 cp /etc/pve/corosync.conf /etc/pve/corosync.conf.bak
771 Then move the new configuration file over the old one:
774 mv /etc/pve/corosync.conf.new /etc/pve/corosync.conf
777 You may check with the commands
780 systemctl status corosync
781 journalctl -b -u corosync
784 If the change could applied automatically. If not you may have to restart the
785 corosync service via:
788 systemctl restart corosync
791 On errors check the troubleshooting section below.
796 Issue: 'quorum.expected_votes must be configured'
797 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
799 When corosync starts to fail and you get the following message in the system log:
803 corosync[1647]: [QUORUM] Quorum provider: corosync_votequorum failed to initialize.
804 corosync[1647]: [SERV ] Service engine 'corosync_quorum' failed to load for reason
805 'configuration error: nodelist or quorum.expected_votes must be configured!'
809 It means that the hostname you set for corosync 'ringX_addr' in the
810 configuration could not be resolved.
813 Write Configuration When Not Quorate
814 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
816 If you need to change '/etc/pve/corosync.conf' on an node with no quorum, and you
817 know what you do, use:
823 This sets the expected vote count to 1 and makes the cluster quorate. You can
824 now fix your configuration, or revert it back to the last working backup.
826 This is not enough if corosync cannot start anymore. Here its best to edit the
827 local copy of the corosync configuration in '/etc/corosync/corosync.conf' so
828 that corosync can start again. Ensure that on all nodes this configuration has
829 the same content to avoid split brains. If you are not sure what went wrong
830 it's best to ask the Proxmox Community to help you.
833 [[corosync-conf-glossary]]
834 Corosync Configuration Glossary
835 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
838 This names the different ring addresses for the corosync totem rings used for
839 the cluster communication.
842 Defines to which interface the ring should bind to. It may be any address of
843 the subnet configured on the interface we want to use. In general its the
844 recommended to just use an address a node uses on this interface.
847 Specifies the mode of the redundant ring protocol and may be passive, active or
848 none. Note that use of active is highly experimental and not official
849 supported. Passive is the preferred mode, it may double the cluster
850 communication throughput and increases availability.
856 It is obvious that a cluster is not quorate when all nodes are
857 offline. This is a common case after a power failure.
859 NOTE: It is always a good idea to use an uninterruptible power supply
860 (``UPS'', also called ``battery backup'') to avoid this state, especially if
863 On node startup, the `pve-guests` service is started and waits for
864 quorum. Once quorate, it starts all guests which have the `onboot`
867 When you turn on nodes, or when power comes back after power failure,
868 it is likely that some nodes boots faster than others. Please keep in
869 mind that guest startup is delayed until you reach quorum.
875 Migrating virtual guests to other nodes is a useful feature in a
876 cluster. There are settings to control the behavior of such
877 migrations. This can be done via the configuration file
878 `datacenter.cfg` or for a specific migration via API or command line
881 It makes a difference if a Guest is online or offline, or if it has
882 local resources (like a local disk).
884 For Details about Virtual Machine Migration see the
885 xref:qm_migration[QEMU/KVM Migration Chapter]
887 For Details about Container Migration see the
888 xref:pct_migration[Container Migration Chapter]
893 The migration type defines if the migration data should be sent over an
894 encrypted (`secure`) channel or an unencrypted (`insecure`) one.
895 Setting the migration type to insecure means that the RAM content of a
896 virtual guest gets also transferred unencrypted, which can lead to
897 information disclosure of critical data from inside the guest (for
898 example passwords or encryption keys).
900 Therefore, we strongly recommend using the secure channel if you do
901 not have full control over the network and can not guarantee that no
902 one is eavesdropping to it.
904 NOTE: Storage migration does not follow this setting. Currently, it
905 always sends the storage content over a secure channel.
907 Encryption requires a lot of computing power, so this setting is often
908 changed to "unsafe" to achieve better performance. The impact on
909 modern systems is lower because they implement AES encryption in
910 hardware. The performance impact is particularly evident in fast
911 networks where you can transfer 10 Gbps or more.
917 By default, {pve} uses the network in which cluster communication
918 takes place to send the migration traffic. This is not optimal because
919 sensitive cluster traffic can be disrupted and this network may not
920 have the best bandwidth available on the node.
922 Setting the migration network parameter allows the use of a dedicated
923 network for the entire migration traffic. In addition to the memory,
924 this also affects the storage traffic for offline migrations.
926 The migration network is set as a network in the CIDR notation. This
927 has the advantage that you do not have to set individual IP addresses
928 for each node. {pve} can determine the real address on the
929 destination node from the network specified in the CIDR form. To
930 enable this, the network must be specified so that each node has one,
931 but only one IP in the respective network.
937 We assume that we have a three-node setup with three separate
938 networks. One for public communication with the Internet, one for
939 cluster communication and a very fast one, which we want to use as a
940 dedicated network for migration.
942 A network configuration for such a setup might look as follows:
945 iface eno1 inet manual
949 iface vmbr0 inet static
951 netmask 255.255.250.0
959 iface eno2 inet static
961 netmask 255.255.255.0
965 iface eno3 inet static
967 netmask 255.255.255.0
970 Here, we will use the network 10.1.2.0/24 as a migration network. For
971 a single migration, you can do this using the `migration_network`
972 parameter of the command line tool:
975 # qm migrate 106 tre --online --migration_network 10.1.2.0/24
978 To configure this as the default network for all migrations in the
979 cluster, set the `migration` property of the `/etc/pve/datacenter.cfg`
983 # use dedicated migration network
984 migration: secure,network=10.1.2.0/24
987 NOTE: The migration type must always be set when the migration network
988 gets set in `/etc/pve/datacenter.cfg`.
992 include::pve-copyright.adoc[]