[one-users] How does opennebula monitor the vm's state?

[Valentin Bud]

Hello Dylan,

On Tue, Nov 13, 2012 at 01:26:14PM +0800, cmcc.dylan wrote:
> Hi, everyone!
>     As we know, opennebula defines vm's life-cycle state, for example: RUNNING,UNKNOWN. I have some questions about vm's life-cycle state.
> 1. I find a problem that when vm's state becomes unknown , after that, even though the vm becomes right, the vm's state will keep unknown all the time.
> 2. opennebula uses what way to judge the vm's state when it executes monitoring action. for example, which funtions used in opennebula source code.

The Information Manager [1], is in charge of monitoring. Start your way
from `/etc/one/oned.conf`, the IM_MAD section. I use KVM, your mileage
may vary.

```
IM_MAD = [
      name       = "im_kvm",
      executable = "one_im_ssh",
      arguments  = "-r 0 -t 15 kvm" ]
```

The executable can be found in `/usr/lib/one/mads/one_im_ssh`. You can
track from there which programs get executed and see the functions you
are interested in.

I have followed the scripts path a couple of days ago. I can tell you that
there are some shell scripts involved and quite a lot of Ruby code. Goodluck.

> 
> 
> look forward to some helps!

Cheers and Goodwill,
v