vm going to error state while migration
I have created a multi node OpenStack setup with two Compute nodes, Controller node and a Network node. I am trying to migrate (Static migrate) an instance from one compute host to another by running the command “nova migrate” but it results in the vm going in error state.
The error is as follow:
Command: ssh 10.10.10.31 mkdir -p /var/lib/nova/instances/9edcde17-38db-4f80-94e7-fb60bf5c55a0 u'ssh: connect to host 10.10.10.31 port 22: Connection refused "message": "Unexpected error while running command.
Please provide information on what could be the reason for this as all the services are up and running.
root@controller:/home# nova service-list +----+------------------+---------------------+----------+---------+-------+----------------------------+-----------------+ | Id | Binary | Host | Zone | Status | State | Updated_at | Disabled Reason | +----+------------------+---------------------+----------+---------+-------+----------------------------+-----------------+ | 1 | nova-consoleauth | controller | internal | enabled | up | 2015-05-06T12:16:09.344861 | - | | 2 | nova-conductor | controller | internal | enabled | up | 2015-05-06T12:16:05.275906 | - | | 3 | nova-cert | controller | internal | enabled | up | 2015-05-06T12:16:06.121017 | - | | 4 | nova-scheduler | controller | internal | enabled | up | 2015-05-06T12:16:05.063736 | - | | 5 | nova-compute | compute | nova | enabled | up | 2015-05-06T12:16:10.137009 | - | | 7 | nova-compute | compute1-VirtualBox | nova | enabled | up | 2015-05-06T12:16:02.091715 |
root@controller:/home# neutron agent-list +--------------------------------------+--------------------+---------+-------+----------------+---------------------------+ | id | agent_type | host | alive | admin_state_up | binary | +--------------------------------------+--------------------+---------+-------+----------------+---------------------------+ | 3002d753-5900-4111-944d-88c6248206e9 | DHCP agent | network | :-) | True | neutron-dhcp-agent | | 3b3720dc-b149-47fb-b920-69a1eafd7629 | Open vSwitch agent | network | :-) | True | neutron-openvswitch-agent | | 627ef370-4a68-48e7-b541-51d9ca34d991 | Open vSwitch agent | compute | :-) | True | neutron-openvswitch-agent | | 660cbe66-b73e-4845-bf4e-23e5fc3e5158 | L3 agent | network | :-) | True | neutron-l3-agent | | 6c9af5dc-3eb2-412c-aa59-1172eba0473b | Metadata agent | network | :-) | True | neutron-metadata-agent | +--------------------------------------+--------------------+---------+-------+----------------
are these Volume backed VM's(persistent instances) or just normal instances (ephemeral instances)? Did you see anything interesting in the nova or libvirt logs? Did you use 'nova reset-state' to bring the instance back to an active state?
yes we are just booting normal instances and while migrating the state goes go to resize and then end up with an error. after using 'nova reset-state' the instance is going to an active state on the same host.
Error: connect to host 10.10.10.31 (ip of new compute-host)port 22: Connection refused
Are you using a shared filesystem? I would look at the systems log files where you are trying to move the vm to during the process (tail -f )..look at both really during the process(the errors are straight forward).
I am running nova migrate vm1 command. I didn't created any storage node (cinder or swift ) ... Do I need to create a storage node for migration ?