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.