Problem on running "nova boot myserver"

[Derek Higgins]

On 08/02/2012 04:30 PM, Stephen Liu wrote:
> 
> 
> - snip -
> 
>> Ok, requests to 10.0.0.2 are going somewhere, you may need to fire up
>> traceroute to see where the requests are going, 10.0.0.2 might clash
>> with the IP range on your local network or across a VPN. If that the
>> case then you should use another range in the call to
> 
> $ traceroute 10.0.0.2
> traceroute to 10.0.0.2 (10.0.0.2), 30 hops max, 60 byte packets
>  1  * * *
>  2  * * *
>  3  * * *
>  4  * * *
>  5  * * *
>  6  * * *
>  7  * * *
>  8  * * *
>  9  * * *
> 10  * * *
> 11  * * *
> 12  * * *
> 13  * * *
> 14  * * *
> 15  * * *
> 16  * * *
> .......

All of the packets are timing out, do you still get a ssh response to
this ipaddress?

> 
> 
>> "Failed to schedule_run_instance: No valid host was found."
>>
>> this usually means the scheduler couldn't find a suitable compute host
>> on which to run a VM, an example of why this could happen is that your
>> compute node doesn't have enough memory to run a VM on. Can you delete
>> the VM in the error state and try nova boot with --flavor 1 which will
>> request less memory.
> 
> Pls explain in more detail on "which VM to be deleted"?  All other VMs were not running when I tested Fedora 17

Sorry, what I meant was, can you delete the openstack instance that that
failed to start. Run
> nova list
> nova delete <uuid of instance in Error state>

then start a new instance with flavor = 1

> 
>>>  nova delete <uuid>
>>>  nova boot ....
>>
>> If memory isn't the problem it may also have something todo with the
>> error logged in cert.log above
> 
> memory assigned to Fedora 17 is 1536M.  I can assign more memory to it.  I have 8G memory onboard.
> 
> $ tail /var/log/nova/cert.log 

my previous email mentioned checking permission of /var/lib/nova/CA to
make sure everything is owned by nova, can you check that also please
and if it isn't change the owner and restart nova-cert

thanks,
Derek.