[one-users] Help with VM launch

[Robert Schweikert]

On 06/09/2011 12:40 PM, Héctor Sanjuán wrote:
> Hi,
>
> I think the deployment is failing because the host is en error state.
>
> It seems that when ONE is trying to monitor it, the 192.168.1.5 host
> closes the connection.
>
> You should make sure that you are running OpenNebula commands with the
> 'oneadmin' user and that this user has passwordless ssh access to the
> host. You can assign a password to the oneadmin unix user with 'passwd
> oneadmin' as root. Then you should be able to login with it and test if
> 'ssh 192.168.1.5' is working well.

OK, I created a password for the oneadmin user and was able to use ssh 
interactively to get to the 192.168.1.5 host without using a password. 
It was a bit slow but it worked.

I also played around with the onehost command, deleting the previously 
added host and using

# onehost create 192.168.1.5 im_kvm vmm_kvm tm_nfs

used this when logged in as oneadmin and with "sudo -u oneadmin" prefix 
when logged in as root. With both approaches I get the same result, 
initially when I create the host the state shown by onehost list is 
"on". But after a while it changes to "err"

Monitoring the log I found the following message:

Thu Jun  9 14:18:05 2011 [InM][I]: Command execution fail: 'if [ -x 
"/var/tmp/one/im/run_probes" ]; then /var/tmp/one/im/run_probes kvm 
192.168.1.5; else                              exit 42; fi'
Thu Jun  9 14:18:05 2011 [InM][I]: STDERR follows.
Thu Jun  9 14:18:05 2011 [InM][I]: error: authentication failed
Thu Jun  9 14:18:05 2011 [InM][I]: error: failed to connect to the 
hypervisor
Thu Jun  9 14:18:05 2011 [InM][I]: Error executing kvm.rb
Thu Jun  9 14:18:05 2011 [InM][I]: ExitCode: 255
Thu Jun  9 14:18:05 2011 [InM][E]: Error monitoring host 4 : MONITOR 
FAILURE 4 Could not monitor host 192.168.1.5.

And on the cloud node in the system log I get:

Jun 10 22:18:03 node1 sshd[4387]: Accepted publickey for oneadmin from 
192.168.1.2 port 60835 ssh2
Jun 10 22:18:03 node1 sshd[4389]: Received disconnect from 192.168.1.2: 
11: disconnected by user
Jun 10 22:18:13 node1 sshd[4413]: Accepted publickey for oneadmin from 
192.168.1.2 port 60836 ssh2
Jun 10 22:18:13 node1 libvirtd: 22:18:13.366: error : 
remoteDispatchAuthPolkit:3797 : Policy kit denied action 
org.libvirt.unix.manage from pid 4453, uid 1000, result: 512


Thus I'd say my issue is on the cloud node side. The passwordless ssh 
connection is clearly working, based on the "Accepted" message, but then 
it looks like policykit get in the way.

The oneadmin user is part of the kvm and the root group on the cloud 
node (libvirtd runs as root on SUSE).

Any ideas what I need to do to beat policy kit into submission?

Thanks,
Robert


-- 
Robert Schweikert                           MAY THE SOURCE BE WITH YOU
SUSE-IBM Software Integration Center                   LINUX
Tech Lead
rschweikert at novell.com
rschweikert at ca.ibm.com
781-464-8147