[one-users] Problem submitting virtual machines: pend status

[Javier Fontan]

Hello,

On Jun 30, 2008, at 11:25 AM, Alvaro Canales wrote:
>
> Could not execute remote script in localhost: /tmp/ne_im_scripts/ 
> ne_im-edc4cd6c5b8e0c6ceacc7da2cd0228ea
>
> The thing is that if I execute it locally, then I can make it work!  
> (and oned.log says that the host is succesfully monitored...):
> [...]
> My local permissions are:
> ls -l /tmp/ne_im_scripts/ne_im-edc4cd6c5b8e0c6ceacc7da2cd0228ea
> -rwxr-xr-- 1 oneadmin xen 1629 Jun 30 11:18 /tmp/ne_im_scripts/ne_im- 
> edc4cd6c5b8e0c6ceacc7da2cd0228ea

Permissions seem to be ok. Please try to do the same test using ssh,  
something like this:

--8<------
$ ssh -n localhost /tmp/ne_im_scripts/ne_im- 
edc4cd6c5b8e0c6ceacc7da2cd0228ea
------>8--

In some distributions sudoers file come with a directive to require a  
tty to use sudo. If this is the case you will get an error telling so.  
To fix this you have to comment the line in sudoers, it should have  
something like "requiretty". I don't have a machine with that  
configuration to tell you how exactly it is.

Also if te error was the requiretty thing then it is also causing one  
being unable to deploy machines (it uses the same sudo mechanism).

> If I try to execute bin/one_im_ssh alone, I get this:
>  bin/one_im_ssh /tmp/ne_im_scripts/ne_im- 
> edc4cd6c5b8e0c6ceacc7da2cd0228ea
> /OVS/opennebula/libexec/madcommon.sh: line 52: /tmp/one_fifo. 
> 27687.err: No such file or directory
> And get a new line in the one_im_ssh.log file (even if I create the  
> fifo manually...):
> Malformed line in configuration file: require "pp"

This is caused because this mad gets a configuration file as  
parameter, not an script file.

> I still don't know why when I submit my images, I get them in "pend"  
> state.
> If I execute "sudo /usr/sbin/xm create /OVS/opennebula/var/VM_ID/ 
> deployment.0 -c", it works and I can log in to my machine (do no net  
> or DHCP).

ONE puts new VM's in pend state. They will be kept in this state until  
they are deployed. That is, doing it by hand (using onevm deploy) or a  
deploy issued by the scheduler. If your machines go to boot state  
without doing anything is because scheduler found a suitable host  
where tor un the VM and sent deploy command.

> I'm very lost, as I don't know what else I can do...
> Any help would be very much appreciated, thank you very much!! :)

I hope you can advance with this information a bit more.

> PS: Campeooones!!! Vivo en Alemania y... jejeje ;p


No veas como estaba ayer Madrid.


Bye