Note | |
---|---|
This feature is for cloud administrators only. |
Migration allows an administrator to move a virtual machine instance from one compute host to another. This feature is useful when a compute host requires maintenance. Migration can also be useful to redistribute the load when many VM instances are running on a specific physical machine.
There are two types of migration:
Migration (or non-live migration): In this case the instance will be shut down (and the instance will know that it has been rebooted) for a period of time in order to be moved to another hypervisor.
Live migration (or true live migration): Almost no instance downtime, it is useful when the instances must be kept running during the migration.
There are two types of live migration:
Shared storage based live migration: In this case both hypervisors have access to a shared storage.
Block live migration: for this type of migration, no shared storage is required.
The following sections describe how to configure your hosts and compute nodes for migrations using the KVM and XenServer hypervisors.
Prerequisites
Hypervisor: KVM with libvirt
Shared storage:
(egNOVA-INST-DIR
/instances//var/lib/nova/instances
) has to be mounted by shared storage. This guide uses NFS but other options, including the OpenStack Gluster Connector are available.Instances: Instance can be migrated with iSCSI based volumes
Note | |
---|---|
Migrations done by the Compute service do not use libvirt's live migration functionality by default. Because of this, guests are suspended before migration and may therefore experience several minutes of downtime. See True Migration for KVM and Libvirt for more details. |
Note | |
---|---|
This guide assumes the default value for |
Note | |
---|---|
You must specify |
Example Nova Installation Environment
Prepare 3 servers at least; for example,
HostA
,HostB
andHostC
HostA
is the "Cloud Controller", and should be running:nova-api
,nova-scheduler
,nova-network
,cinder-volume
,nova-objectstore
.HostB
andHostC
are the "compute nodes", runningnova-compute
.Ensure that,
(set withNOVA-INST-DIR
state_path
innova.conf
) is same on all hosts.In this example,
HostA
will be the NFSv4 server which exports
, andNOVA-INST-DIR
/instancesHostB
andHostC
mount it.
System configuration
Configure your DNS or
/etc/hosts
and ensure it is consistent across all hosts. Make sure that the three hosts can perform name resolution with each other. As a test, use the ping command to ping each host from one another.$ ping HostA $ ping HostB $ ping HostC
Ensure that the UID and GID of your nova and libvirt users are identical between each of your servers. This ensures that the permissions on the NFS mount will work correctly.
Follow the instructions at the Ubuntu NFS HowTo to setup an NFS server on
HostA
, and NFS Clients onHostB
andHostC
.Our aim is to export
fromNOVA-INST-DIR
/instancesHostA
, and have it readable and writable by the nova user onHostB
andHostC
.Using your knowledge from the Ubuntu documentation, configure the NFS server at
HostA
by adding a line to/etc/exports
NOVA-INST-DIR
/instances HostA/255.255.0.0(rw,sync,fsid=0,no_root_squash)Change the subnet mask (
255.255.0.0
) to the appropriate value to include the IP addresses ofHostB
andHostC
. Then restart the NFS server.$ /etc/init.d/nfs-kernel-server restart $ /etc/init.d/idmapd restart
Set the 'execute/search' bit on your shared directory
On both compute nodes, make sure to enable the 'execute/search' bit to allow qemu to be able to use the images within the directories. On all hosts, execute the following command:
$ chmod o+x
NOVA-INST-DIR
/instancesConfigure NFS at HostB and HostC by adding below to
/etc/fstab
.HostA:/ /
NOVA-INST-DIR
/instances nfs4 defaults 0 0Then ensure that the exported directory can be mounted.
$ mount -a -v
Check that "
" directory can be seen at HostANOVA-INST-DIR
/instances/$ ls -ld
NOVA-INST-DIR
/instances/drwxr-xr-x 2 nova nova 4096 2012-05-19 14:34 nova-install-dir/instances/
Perform the same check at HostB and HostC - paying special attention to the permissions (nova should be able to write)
$ ls -ld
NOVA-INST-DIR
/instances/drwxr-xr-x 2 nova nova 4096 2012-05-07 14:34 nova-install-dir/instances/
$ df -k
Filesystem 1K-blocks Used Available Use% Mounted on /dev/sda1 921514972 4180880 870523828 1% / none 16498340 1228 16497112 1% /dev none 16502856 0 16502856 0% /dev/shm none 16502856 368 16502488 1% /var/run none 16502856 0 16502856 0% /var/lock none 16502856 0 16502856 0% /lib/init/rw HostA: 921515008 101921792 772783104 12% /var/lib/nova/instances ( <--- this line is important.)
Update the libvirt configurations so that the calls can be made securely. These methods enable remote access over TCP and are not documented here, please consult your network administrator for assistance in deciding how to configure access.
SSH tunnel to libvirtd's UNIX socket
libvirtd TCP socket, with GSSAPI/Kerberos for auth+data encryption
libvirtd TCP socket, with TLS for encryption and x509 client certs for authentication
libvirtd TCP socket, with TLS for encryption and Kerberos for authentication
Restart libvirt. After you run the command, ensure that libvirt is successfully restarted:
$ stop libvirt-bin && start libvirt-bin $ ps -ef | grep libvirt
root 1145 1 0 Nov27 ? 00:00:03 /usr/sbin/libvirtd -d -l
Configure your firewall to allow libvirt to communicate between nodes.
Information about ports used with libvirt can be found at the libvirt documentation By default, libvirt listens on TCP port 16509 and an ephemeral TCP range from 49152 to 49261 is used for the KVM communications. As this guide has disabled libvirt auth, you should take good care that these ports are only open to hosts within your installation.
You can now configure options for live migration. In most cases, you do not need to configure any options. The following chart is for advanced usage only.
Configuration option=Default value | (Type) Description |
live_migration_bandwidth=0 | (IntOpt)Maximum bandwidth to be used during migration, in Mbps |
live_migration_flag=VIR_MIGRATE_UNDEFINE_SOURCE, VIR_MIGRATE_PEER2PEER | (StrOpt)Migration flags to be set for live migration |
live_migration_retry_count=30 | (IntOpt)Number of 1 second retries needed in live_migration |
live_migration_uri=qemu+tcp://%s/system | (StrOpt)Migration target URI (any included "%s" is replaced with the migration target hostname) |
By default, the Compute service does not use libvirt's live migration functionality. To
enable this functionality, add the following line to nova.conf
:
live_migration_flag=VIR_MIGRATE_UNDEFINE_SOURCE,VIR_MIGRATE_PEER2PEER,VIR_MIGRATE_LIVE
The Compute service does not use libvirt's live miration by default because there is a risk that the migration process will never terminate. This can happen if the guest operating system dirties blocks on the disk faster than they can migrated.
Prerequisites
Compatible XenServer hypervisors. For more information, please refer to the Requirements for Creating Resource Pools section of the XenServer Administrator's Guide.
Shared storage: an NFS export, visible to all XenServer hosts.
Note Please check the NFS VHD section of the XenServer Administrator's Guide for the supported NFS versions.
In order to use shared storage live migration with XenServer hypervisors, the hosts must be joined to a XenServer pool. In order to create that pool, a host aggregate must be created with special metadata. This metadata will be used by the XAPI plugins to establish the pool.
Add an NFS VHD storage to your master XenServer, and set it as default SR. For more information, please refer to the NFS VHD section of the XenServer Administrator's Guide.
Configure all the compute nodes to use the default sr for pool operations, by including:
sr_matching_filter=default-sr:true
in your
nova.conf
configuration files across your compute nodes.Create a host aggregate
$ nova aggregate-create <name-for-pool> <availability-zone>
The command will display a table which contains the id of the newly created aggregate. Now add special metadata to the aggregate, to mark it as a hypervisor pool
$ nova aggregate-set-metadata <aggregate-id> hypervisor_pool=true $ nova aggregate-set-metadata <aggregate-id> operational_state=created
Make the first compute node part of that aggregate
$ nova aggregate-add-host <aggregate-id> <name-of-master-compute>
At this point, the host is part of a XenServer pool.
Add additional hosts to the pool:
$ nova aggregate-add-host <aggregate-id> <compute-host-name>
Note At this point the added compute node and the host will be shut down, in order to join the host to the XenServer pool. The operation will fail, if any server other than the compute node is running/suspended on your host.
Prerequisites
Compatible XenServer hypervisors. The hypervisors must support the Storage XenMotion feature. Please refer to the manual of your XenServer to make sure your edition has this feature.
Note | |
---|---|
Please note, that you need to use an extra option |
Note | |
---|---|
Please note, that block migration works only with EXT local storage SRs, and the server should not have any volumes attached. |