Ask Your Question
0

vm instance stuck in migrating status

asked 2016-08-16 01:44:48 -0500

yee379 gravatar image

I have live block migrations working in my RDO-Liberty deployment. I do not have shared storage and use the local disks for /var/lib/nova.

As i have some rather large (disk) instances, i set live_migration_progress_timeout to 0.

unfortuantely, this seems to have been a bad idea, as i was manually live migrating a couple of instances off a hypervisor at the same time... for other instances, if i did it one at a time, it took about 5 minutes; for these... well... one of them just gave up after about ~20 mins, and the other one is stuck in MIGRATING status (as per nova) - and has been for about 6 hours now.

is there a way i can just cancel the migration?

edit retag flag offensive close merge delete

1 answer

Sort by ยป oldest newest most voted
0

answered 2016-08-16 18:57:50 -0500

fifi gravatar image

updated 2016-08-16 19:01:11 -0500

I assume you are using QEMU and libvirt. To cancel the migration you run the following commands on your compute node which hosts the migrating instance:

 service nova-compute restart


service libvirt-bin restart

The reason your instance gets stuck in migration or it takes a long time can be:

1- wrong configuration

2- network bandwidth limit

3-RabbitMQ problem

If you post more detail, I can give you more accurate answer

edit flag offensive delete link more

Comments

does restarting nova-compute and libvirt-bin on hypervisor will cause any side effects to other VM's running on it

Deepa gravatar imageDeepa ( 2017-10-11 04:04:24 -0500 )edit

Your Answer

Please start posting anonymously - your entry will be published after you log in or create a new account.

Add Answer

Get to know Ask OpenStack

Resources for moderators

Question Tools

1 follower

Stats

Asked: 2016-08-16 01:44:48 -0500

Seen: 1,858 times

Last updated: Aug 16 '16