Cirros bootable but no ubuntu

asked 2014-10-01 06:28:22 -0600

n7rxDE gravatar image

updated 2014-10-02 04:43:13 -0600

Hey there,

i have a nearly working 3 node icehouse setup with ubuntu server.

my 3 nodes are setup as virtualbox VMs.

I can start the recommended cirros image withe both tiny and small flavor.

But if i try to get an ubuntu image up it returns any time no valid host found.

tried ubuntu 32 bit and 64 bit both in 14.04 version.

my ubuntu images does come from:

any thoughts on this topic would be nice!

UPDATE: nova-compute.log

2014-10-02 11:38:56.144 2184 INFO [req-c2f02e1e-71cf-4370-b28f-b08b4583267b 2b315f1eddbe4d8ca6fe8d51ea5e5e3a f1d56b01ff284d77b2e07e404e28b1ad] [instance: cc2a49e0-73de-4b93-b4ac-52b66405da57] Unable to reset device ID for port None
2014-10-02 11:38:56.283 2184 ERROR nova.compute.manager [req-c2f02e1e-71cf-4370-b28f-b08b4583267b 2b315f1eddbe4d8ca6fe8d51ea5e5e3a f1d56b01ff284d77b2e07e404e28b1ad] [instance: cc2a49e0-73de-4b93-b4ac-52b66405da57] Error: [Errno 32] Corrupted image. Checksum was 90767da1889cf9c5958b082fc317bb47 expected af77e9467d7b71d85ccd064b5b848c60


edit retag flag offensive close merge delete

3 answers

Sort by ยป oldest newest most voted

answered 2014-10-02 04:51:22 -0600

SGPJ gravatar image

Try Ubuntu Cloud Images from with small flav:

edit flag offensive delete link more


which img file i have to use?

n7rxDE gravatar imagen7rxDE ( 2014-10-02 05:04:30 -0600 )edit

new image same problem... on spawning it says on compute log that checksums does not match...

n7rxDE gravatar imagen7rxDE ( 2014-10-02 05:32:53 -0600 )edit

try to use 12.04 it is working for me.

SGPJ gravatar imageSGPJ ( 2014-10-03 11:49:10 -0600 )edit

answered 2014-10-01 10:31:05 -0600

mpetason gravatar image

It may be due to the flavor you are selecting for the instance. Do you currently have enough resources on a compute node for the new instance? Most of the time you are either going to be using a Flavor that is too small for the image, or a flavor that has too many resources for the compute node.

On your compute node you can find more details about the instance by looking in /var/log/nova/compute.log.

On the controller you can run these commands to figure out what the rest of the error message was and where the instance is running (in case you add more compute nodes.)

source openrc
nova list --all-tenants

Find the instance and copy the ID - it's a UUID.

nova show {uuid}

Now you can find the Hypervisor and you can even see the traceback that shows what happened when launching.

If there isn't anything about the instance then it could also be on the controller node in the scheduling logs. Use the UUID again and look for any relevant error messages:

grep -r {uuid} /var/log/nova/*
edit flag offensive delete link more


thx first of all, i had used 2gb ram 1vcpu and 20gb flavor for both... works only for cirros, i had thought it should be good enough for ubuntu? i will have a look to your thoughts and hope i find someting

n7rxDE gravatar imagen7rxDE ( 2014-10-01 16:17:39 -0600 )edit

answered 2014-10-02 06:52:35 -0600

n7rxDE gravatar image

For now it works, but i dont know why.

System chrashes and i had to restart at all, after that i can start up ubuntu and cirros instance both are working well...

any thoughts where this belongs to ?

edit flag offensive delete link more

Get to know Ask OpenStack

Resources for moderators

Question Tools

1 follower


Asked: 2014-10-01 06:28:22 -0600

Seen: 596 times

Last updated: Oct 02 '14