how do i keep vm running after compute reboot without controller accessability

asked 2015-06-16 10:35:41 -0600

Uzi
41 ●8 ●9 ●10

updated 2015-06-16 10:37:53 -0600

Hi All i am using openstack Juno with Ubuntu controller/compute

there use cases when controller wont be accessible to compute ,

how do i keep the current compute nova/neutron state when such happen after i reboot the compute even if controller not accessible ? (that is keep the VMs active , keep the current neutron/ml2/ovs flows as they been )

post reboot if controller is accessible resume the VMs to active is via this nova.conf attribute

resume_guests_state_on_host_boot=true

thanks

edit retag flag offensive close merge delete

Comments

The best practice is not to reboot compute nodes during controller connection outage window. You can restart a VM by using virsh if using qemu/kvm. The VM however might not be able to get its IP unless DHCP agent is also running on that compute node.

alfredcs@yahoo.com ( 2015-06-17 01:53:44 -0600 )edit

1 answer

Sort by » oldest newest most voted

0

answered 2015-06-17 01:57:56 -0600

673 ●6 ●16 ●27 http://uttamhoode.blog...

updated 2015-06-17 01:59:03 -0600

Sometimes compute node will have difficulty in reconnecting to control node when it is up. So it better to configure your controller node in HA.

edit flag offensive delete link

Comments

thanks ! , VM is at active state i encounter with yet another problem

the OVS bridges cleared from flows when controller accessible flows re populated with the correct flows derived from VMs attached networks

can i configure the ovs/neutron to keep last flows configuration after reboot ?

Uzi ( 2015-06-17 03:10:04 -0600 )edit

Get to know Ask OpenStack

Resources for moderators

Question Tools

1 follower

Stats

Asked: 2015-06-16 10:35:41 -0600

Seen: 710 times

Last updated: Jun 17 '15

Related questions