openstack team mailing list archive

[Sylvain Bauza <sylvain.bauza@xxxxxxxxxxxx>]

Le 20/02/2013 14:04, Chathura M. Sarathchandra Magurawalage a écrit :
> There are apparently two instances running in the compute node but 
> nova just see only one. Probably when I have deleted an instance 
> earlier it had not deleted the instance properly.
>
> root@controller:~# nova list
> +--------------------------------------+--------+--------+-------------------+
> | ID                                   | Name   | Status | Networks   
>        |
> +--------------------------------------+--------+--------+-------------------+
> | 42e18cd5-de6f-4181-b238-320fe37ef6f1 | master | ACTIVE | 
> demo-net=10.5.5.3 |
> +--------------------------------------+--------+--------+-------------------+
>
>
> virsh -c qemu+ssh://root@computenode/system list
> root@computenode's password:
>  Id    Name                           State
> ----------------------------------------------------
>  14    instance-0000002c              running
>  18    instance-0000001e              running

You should have seen at 'sudo virsh list --all', plus looking at 
/etc/libvirt/qemu/*.xml to check how many instances were defined.
I do suspect also that for some reason (probably nova-compute down), a 
clean-up of 00002c probably didn't work. Anyway, this is fixed as you 
mention.


> Then I have deleted all instances and created a new one. But still 
> cant ping or ssh the new VM.
>
>     <interface type='bridge'>
>       <mac address='fa:16:3e:a2:6e:02'/>
>       <source bridge='qbrff8933bf-ba'/>
>       <model type='virtio'/>
>       <filterref filter='nova-instance-instance-00000035-fa163ea26e02'>
>         <parameter name='DHCPSERVER' value='10.5.5.2'/>
>         <parameter name='IP' value='10.5.5.3'/>
>         <parameter name='PROJMASK' value='255.255.255.0'/>
>         <parameter name='PROJNET' value='10.5.5.0'/>
>       </filterref>
>       <address type='pci' domain='0x0000' bus='0x00' slot='0x03' 
> function='0x0'\
> />
>     </interface>
>
> Starting network...
> udhcpc (v1.18.5) started
> Sending discover...
> Sending select for 10.5.5.3...
> Lease of 10.5.5.3 obtained, lease time 120
> deleting routers
> route: SIOCDELRT: No such process
> adding dns 8.8.8.8


The DHCP reply is correctly received by the instance from the network 
node to the compute node. This is not a network issue (at least for IP 
assignation).
I would be more concerned about the SIOCDELRT error above. Do you try to 
manually remove a network route at bootup ? Seems like the 'route del' 
is failing because the route is not already existing.


As already said, you absolutely need VNC support for investigating. 
Could you please fix your VNC setup which is incorrect ?

>     <graphics type='vnc' port='-1' autoport='yes' listen='0.0.0.0' 
> keymap='en-u\
> s'>
>       <listen type='address' address='0.0.0.0'/>
>     </graphics>

Try in nova-compute.conf :
vncserver_proxyclient_address=<compute node mgmt IP>
vncserver_listen=<compute node mgmt IP>
and in nova.conf :
novncproxy_base_url=http://<controler node mgmt IP>:6080/vnc_auto.html

and restart nova-compute.



> On 20 February 2013 11:57, Sylvain Bauza <sylvain.bauza@xxxxxxxxxxxx 
> <mailto:sylvain.bauza@xxxxxxxxxxxx>> wrote:
>
>     Could you please paste :
>      - /etc/libvirt/qemu/<your_instance_id>.xml
>      - ip a show vnet0
>      - brctl show
>
>     Sounds like your virtual device is not created. Could you please
>     launch a new VM and paste /var/log/nova/nova-compute.log ?
>
>     Thanks,
>     -Sylvain