OpenStack launching instance error

0
votes

I'm trying to launch an instance of Ubuntu in OpenStack, but I'm getting this error:

"Error: Failed to perform requested operation on instance , the instance has an error status: Please try again later [Error: Build of instance aborted: Block Device Mapping is Invalid.]".

I've tried this: https://ask.openstack.org/en/question/62636/instance-creation-fails-block-device-mapping/ but none of the answers seem to work in my case.

What could I do?

1
cloudopenstack
Which version of OpenStack you're using? Also let me know the Ubuntu version. - Ashish Srivastava
@AshishSrivastava I'm using the latest version of Packstack OpenStack and Ubuntu Server 16.04 LTS (cloud image). - Rubén
So you have freshly installed OpenStack-Newton or just updated? - Ashish Srivastava

1 Answers

1
votes

A simple trick that I did recently.

  1. SSH into the openstack controller or cinder node
  2. cd into /var/log/cinder
  3. tail volume.log //I used yad --tail though;)

Read the last lines of the log file (I assume you dint try spinning any VM’s since this error was received.)

The last line will look something like this:

ERROR *****: Image *****: Error message

Hope this helps you to find the actual error and solve it.

UPDATE: Error: No valid host was found. There are not enough hosts available

I would suggest to tail logs before powering on the Instances. Something like

( tail -f /var/log/nova/* | grep -i error

Probably the problem is with space on your server. Open the /etc/nova/nova.conf file and replace (also try to change the value of cpu_allocation_ratio/disk_allocation_ratio):

ram_allocation_ratio=1.5 (probably 0.0 in some cases) to ram_allocation_ratio=3

Than restart the services:

sudo service openstack-nova-scheduler restart

sudo service openstack-nova-conductor restart

Also have a look here.

If you are running everything on a virtual machine, set virt_type=qemu in /etc/nova/nova.conf ([libvirt] section) as this post says.